Description
Title of Invention: NODE APPARATUS AND COMMUNICATION METHOD
Technical Field
[0001] The present invention relates to communication in a
network including a plurality of node apparatuses.
Background Art
[0002] With regard to communication in a network, various
techniques have been developed from various viewpoints, such
as control in case of occurrence of a fault, redundancy for fault
tolerance, and routing control.
For instance, in order to facilitate reduction in network
cost and quickly determine a fault, the following communication
system has been disclosed. That is, a transmitter-side router
A continuously transmits, to a receiver-side router B, packets
including user packets at time intervals shorter than a fault
determination time that is agreed to by the router B in advance.
The router B includes a timer expiring in the fault determination
time.
[0003] In a case where a network between the routers A and
B is normal, the router B receives a packet before expiration
of the timer. By using this, the router B judges that the network
is normal and resets the timer if it receives a packet before
expiration of the timer. If the timer expires before reception
of a packet, the router B can judge that the network is not normal
and therefore determines that a fault occurs. Such application
of the determination by the timer to transmission and reception
of packets including user packets enables quick determination
of a fault in the network, while avoiding network duplication
and while facilitating reduction in cost.
[0004] The following telephony system has also been disclosed
in order to solve a problem that a conventional Internet telephony
system is incapable of distinguishing whether voice packets are

not received due to silence compression or the voice packets
cannot be received owing to occurrence of abnormality in the
Internet.
[0005] That is, after a call state has been establishedbetween
telephones, a calling-sidemedia gateway periodically transmits
a monitoring packet to a called-side media gateway via the
Internet. When receiving the monitoring packet, the
called-side media gateway transmits a response packet to the
calling-side media gateway via the Internet. When the number
of times at which response or monitoring packets to be received
fail to be periodically inputted reaches a prescribed threshold,
the two media gateways issue a non-packet-receipt notification
to a call agent. When receiving the non-packet-receipt
notification, the call agent controls the two media gateways
to execute a disconnection process.
[0006] The following communication path switching control
system has also been disclosed in order to solve a problem that,
even if a communication path is switched by a router, apparatuses
other than the router cannot recognize that the communication
path has been switched.
[0007] The communication path switching control system
includes a first IPGW (Internet Protocol Gateway) for
accommodating and connecting a first PBX (Private Branch
exchange), and a first router responsible for a communication
interface between a network and the first IPGW. When detecting
a line fault in a communication path currently in use among a
plurality of communication paths in the network, the first router
searches for a communication path for bypassing the line fault
from among the plurality of communication paths, and switches
its connection to the searched-for communication path. The
first router includes a switching information notification
section that notifies, upon switching of the connection to the
searched-for communication path, the first IPGWof path switching
information pertaining to the communication path to which the
connection has been switched. The first IPGW includes a line

control section that performs, when detecting the path switching
information from the switching information notification section,
a line control on the first PBX according to the path switching
information.
[0008] Incidentally, in recent years, ad hoc networks have
become a focus of attention. As to optimal routing in an ad
hoc wireless communication network, the following system and
method have been disclosed for calculating an optimal route at
a node.
[0009] That is, routing metric used in the system and the
method, if carefully chosen, can provide stability to a network
and also provide features like self-healing and load balancing.
A routing metric is calculated as a scalar number based upon
a number of factors, such as the number of hops, a data rate,
link quality, and a device type. Each factor can be determined
by evaluation of hallo messages, or other routing messages as
required.
Not only wireless ad hoc networks but also wired ad hoc
networks have been researched, and application thereof to a
sensor network has also been attempted.
[0010] For instance, in a wired sensor network referred to
as "S-wire", each node apparatus is connected to a plurality
of node apparatuses in a wired manner, and data communication
and power supply are performed in a wired manner. Advantages
of the wired system include a feature that the sensors can be
embedded in earth, water, a structure and the like and a feature
that a breakage and the like can be detected.
Citation List
Patent Literature
[0011] Patent Literature 1: Japanese Laid-Open Patent
Publication No. 2003-273964
Patent Literature 2: Japanese Laid-Open Patent Publication
No. 2002-271399
Patent Literature 3: Japanese Laid-Open Patent Publication

No. 2006-340165
Patent Literature 4: Japanese Laid-Open Patent Publication
No. 2006-526937
Non Patent Literature
[0012] Non Patent Literature 1: Tadashige Iwao, Ken j i Yamada,
Koji Nomura, and Takeshi Hosokawa, "Multipurpose Practical
Sensor Network: S-wire" journal FUJITSU, May 2006 (Vol. 57, No.
3), pp. 285-290
Summary of Invention
Technical Problem
[0013] As exemplified above, as to communication in a network
including a plurality of node apparatuses, techniques have been
developed from various viewpoints, such as control in case of
occurrence of a fault, redundancy for fault tolerance, and
routing control. However, in a network system including a
plurality of node apparatuses, there is room for improving
self-contained natures of individual node apparatuses since it
is typically practiced that the node apparatuses exchange
inf ormationpertainingto a route andthereby realize distributed
coordination.
[0014] Thus, it is an object of the present invention toprovide
a technique according to which, in a network including aplurality
of node apparatuses, the individual node apparatuses operate
according to information obtainable in a self-contained manner
and thereby realize an appropriate distributed coordination in
the entire network.
Solution to Problem
[0015] An aspect of the present invention provides a first
node apparatus in a network including a plurality of node
apparatuses including the first node apparatus and a second node
apparatus that are connected in a wired manner.
The first node apparatus includes: a plurality of ports;
loop detection information storage means for storing loop

detection information; routing information storage means for
storing routing information; receiving means; routing
information updating means; and transmitting means.
[0016] Each of the plurality of ports is a port to connect,
in the wiredmanner, the first node apparatus to another different
node apparatus other than the first node apparatus among the
plurality of node apparatuses.
The loop detection information associates, with first
identification information for uniquely identifying a first
frame, destination port distinguishing information for
distinguishing, from among the plurality of ports, a first port
that is a destination port for a case where the first node apparatus
has transmitted the first frame.
[0017] The routing information associates, with each of the
plurality of node apparatuses, state information that is
information for indicating whether frame transmission from each
of the plurality of ports is feasible or not.
The receivingmeans receives, from the secondnode apparatus,
a second frame that includes second identification information
for uniquely identifying the second frame.
[0018] When the second identification information is
identical with the first identification information, the routing
information updating means updates the state information
associated by the routing information with a destination node
apparatus that is one of the plurality of node apparatuses and
that is a destination of the second frame so as to indicate that
frame transmission from the first port is not feasible.
[0019] The transmittingmeans selects a second port f romwhich
transmission of the second frame is feasible from among the
plurality of ports, accordingto the stateinf ormation associated
by the routing information with the destination node apparatus.
The transmitting means then transmits the second frame from the
second port.
[0020] Another aspect of the present invention provides a
communication method executed by the first node apparatus.

Advantageous Effects of Invention
[0021] If the first frame loops in the network and is received
again as the second frame by the first node apparatus, the second
identification information is identical with the first
identification information. Accordingly, update is performed
by the routing information updating means. As a result, the
first node apparatus is enabled to transmit the second frame
while avoiding a port on a route causing the loop.
[0022] In addition, the loop detection information and the
routing information, which are used by the first node apparatus,
are pieces of information obtainable by the first node apparatus
itself in a self-contained manner. That is, the first node
apparatus is capable of obtaining the loop detection information
and the routing information, without need to exchange, in
addition to the first and second frames, information pertaining
to a route with another node apparatus.
[0023] Accordingly, in a network including a plurality of
node apparatuses each of which is eguivalent to the first node
apparatus, the individual node apparatuses operate according
to information obtainable in a self-contained manner, thereby
leading the entire network so that a frame is relayed via an
appropriate route that does not loop.
Brief Description of Drawings
[0024] Figure 1 is a configuration diagram which explains
an overview of a node apparatus;
Figure 2A is a diagram which illustrates an example of a
network to which the node apparatus of this embodiment is applied;
Figure 2B is a conceptual diagramwhich explains alternative
route finding in case of occurrence of a fault;
Figure 3 is a diagram which explains a frame format;
Figure 4 is a diagram which explains that a broadcast frame
is prevented from congestion;
Figure 5 is a hardware configuration diagram of the node

apparatus;
Figure 6 is a functional configuration diagram of the node
apparatus;
Figure 7 is a diagram which illustrates an example of a
pause state table;
Figure 8 is a diagram which illustrates an example of a
port link state table;
Figure 9 is a diagram which illustrates an example of a
MAC table;
Figure 10 is a diagram which illustrates an example of a
pause state management data;
Figure 11A is a flowchart (No. 1) which explains a routing
process;
Figure 11B is a flowchart (No. 2) which explains the routing
process;
Figure 11C is a flowchart (No. 3) which explains the routing
process;
Figure 11D is a flowchart (No. 4) which explains the routing
process;
Figure 12 is a flowchart which explains aging of a routing
table;
Figure 13 is a flowchart which explains setting of a timer
of an entry in the routing table;
Figure 14 is a flowchart which explains aging of a loop
detection table;
Figure 15 is a flowchart which explains setting of a timer
of an entry in the loop detection table;
Figure 16 is a flowchart of a pause state canceling process;
Figure 17 is a flowchart which explains a pause controlling
process executed by the node apparatus on the node apparatus
itself;
Figure 18 is a flowchart of a pause starting process;
Figure 19 is a flowchart of a pause canceling process; and
Figure 20 is a flowchart of a port monitoring process.

Description of Embodiments
[0025] Embodiments will hereinafter be described in detail
with reference to drawings. The order of description is as
follows.
First, several terms are defined and then an overview of
a configuration of a node apparatus of this embodiment is
described with reference to Figure 1. Next, an overview of
distributed coordination operation in a network including a
plurality of node apparatuses each including components as in
Figure 1 is described with reference to Figures 2A, 2B and 4 .
Along therewith, an example of a frame format is also described
with reference to Figure 3.
[0026] Subsequently, the configuration of the node apparatus
is described in detail with reference to Figures 5 and 6. Further,
specific examples of data omitted in Figure 1 among data used
by the node apparatus are described with reference to Figures
7 to 10.
[0027] Moreover, the operation of an individual node
apparatus is described in detail with reference to flowcharts
of Figures 11A to 20. In the description of the operation of
the individual node apparatus, a relationship between the
operation of the individual node apparatus and the distributed
coordination operation realized in the entire network is also
described, with reference to Figure 2A or 4, as appropriate.
Subsequently, the node apparatus of this embodiment is
recapitulated. Further, advantages of this embodiment are
described in comparison with some techniques. Finally,
modifications of this embodiment are also described.
[0028] Here, some terms used in this embodiment are defined.
In this embodiment, a plurality of node apparatuses are
included in a network. Hereinafter, first and second node
apparatuses are referred to as "adjacent" if the first and second
node apparatuses are directly connectedby a cable. With respect
to the first node apparatus, the second node apparatus may be
referred to as an "adjacent node apparatus". Likewise, with

respect to the second node apparatus, the first node apparatus
is its "adjacent node apparatus".
[0029] In this embodiment, various frames are processed. The
frames are classified from several viewpoints, as described
below.
A frame is classifiable according to whether it includes
an "ad hoc header" illustrated in Figure 3 or not. Hereinafter,
a frame including the ad hoc header is referred to as an "ad
hoc frame". On the other hand, examples of frames without an
ad hoc header include, for instance, an Ethernet frame used in
well-known Ethernet (registered trademark).
[0030] The ad hoc frame may be further classified into a normal
frame including the Ethernet frame subsequent to the ad hoc header
and a control frame including control data in another specific
format subsequent to the ad hoc header. There may be a plurality
of types of the control frames. An example of the control frame
is a "pause frame" described later.
[0031] The frames may also be classified into a broadcast
frame that is broadcast in a network, and a unicast frame that
is unicast. Although detailed description will be made later,
the pause frame is broadcast in an exceptional manner limited
to adjacent node apparatuses.
[0032] In the following description, a classification of a
frame from a viewpoint as described above is explicitly described,
as necessary. In a passage where the classification is apparent
from the context or a passage irrelevant to the classification,
the simple notation "f rame"may be used for the sake of simplicity.
[0033] Figure 1 is a configuration diagram which explains
an overview of a node apparatus.
The node apparatus 100 illustrated in Figure 1 includes a
plurality of ports 101-1 to 101-x (1 < x) for wired connection
with adjacent node apparatuses (not illustrated in Figure 1),
and a routing engine 102 that routes frames.
[0034] The ports 101-1 to 101-x in Figure 1 are those for
a wired ad hoc network of this embodiment. Note that, in Figure

5 described later, the ports 101-1 to 101-x in Figure 1 are
represented as "wired ad hoc network ports", which are
distinguished from a general LAN (Local Area Network) port
conforming to general Ethernet specifications. Hereinafter,
the simple notation "port" means a port for the wired ad hoc
network according to this embodiment.
[0035] The node apparatus 100 also includes tables, namely,
a routing table 103, a broadcast management table 104 and a loop
detection table 105. The routing engine 102 performs routing
with reference to these tables. The routing engine 102 also
updates these tables.
[0036] As illustrated in Figure 1, the routing table 103 of
this embodiment is a table that stores a "destination" and "x"
number of "port states", which respectively represent states
of the ports 101-1 to 101-x, in association with each other.
[0037] Figure 1 exemplifies the routing table 103 including
"n" number of entries . As will be described later with reference
to Figure 11C, each entry is one that has been added to the routing
table 103 on the basis of a frame received by the node apparatus
100. As will be described later with reference to Figures 12
and 13, each entry is deleted from the routing table 103 if not
accessed for a prescribed time.
[0038] For instance, in the i-th (1 < i < n) entry in the
routing table 103, states RPli to RPxi of the respective ports
101-1 to 101-x are associated with a destination GDi.
[0039] Here, the destination GDi is identification
information that identifies the node apparatus that is to be
the destination of a frame in a wired ad hoc network including
a plurality of node apparatuses each equivalent to the node
apparatus 100. In this embodiment, each node apparatus is
assigned a three-byte ID (IDentification) that is unique in the
wired ad hoc network. Hereinafter, the ID assigned to the node
apparatus is referred to as a "node ID". More specifically,
the destination GDi is a node ID.
[0040] When the node apparatus 100 receives a frame addressed

to the destination GDi at any of the ports 101-1 to 101-x and
tries to relay the received frame, the routing engine 102 refers
to the routing table 103. Then, the routing engine 102 selects
a port for transmitting (i.e., outputting) the received frame
on the basis of the states RPli to RPxi associated with the
destination GDi in the routing table 103.
[0041] In the i-th entry in the routing table 103, a state
RPji of the j-th (1 ≤ j ≤ x) port 101-j is any of five states,
namely, "U" (Used), ME" (Empty), "L" (Loop), "P" (Pause), and
"D" (Down).
[0042] "U" indicates an in-use state. More specif ically, the
state RPji is "U" if the port 101-j is actually used as a destination
port in the node apparatus 100 for relaying a frame addressed
to another node apparatus identified by a node ID being the
destination GDi. That is, the "U" state is an example of a state
indicating "transmission is feasible".
[0043] "E" indicates an unused state. More specif ically, the
state RPji is "E" if the port 101-j has not been used as a destination
port in the node apparatus 100 for relaying a frame addressed
to another node apparatus identified by a node ID being the
destination GDi. That is, the "E" state is an example of a state
indicating "transmission is feasible".
[0044] "L" indicates a loop state. More specifically, the
state RPji is updated to "L" if both of the following two conditions
hold.
• The port 101-j has been used as a destination port in the
node apparatus 100 for relaying a frame addressed to another
node apparatus identified by a node ID being the destination
GDi.
[0045] • A frame having been transmitted from the port 101-j
toward another node apparatus identified by a node ID being the
destination GDi loops in the network and is returned to the node
apparatus 100 itself (i.e., is received by the node apparatus
100 again).
[0046] Note that, in a case where the state RPji is "L", the

state RPji means that a frame addressed to the destination GDi
should not be transmitted from the port 101-j because the frame
addressed to the destination GDi will loop if transmitted from
the port 101-j . That is, the "L" state is an example of a state
indicating "transmission is not feasible".
[0047] "P" indicates a pause state. More specifically, the
state RPji turns into "P" if a "pause request" is received from
the port 101-j (i.e., from an adjacent node apparatus which is
not illustrated and which is connected via the port 101-j ) . Note
that the "pause request" is a request by which a node apparatus
whose processing load increases owing to concentration of frames
and the like requests its adjacent node apparatus to "pause"
transmission of frames to this node apparatus itself for a certain
time period. The detailed description on the pause request will
be made later with reference to Figures 16 to 19. The overview
thereof is as follows.
[0048] For instance, assume that the node apparatus 100 in
Figure 1 is connected with a second node apparatus, which is
not illustrated, via the port 101-j . Hereinafter, a state in
which the processing load exceeds a prescribed criterion is
referred to as a "busy state".
[0049] If the second node apparatus falls into the busy state,
it transmits the pause request to each of a plurality of its
adjacent node apparatuses. More specifically, the pause
request is represented by a "pause frame", namely, an ad hoc
frame that has a specific format and that includes a value
designating the length of the above-mentioned certain time
period.
[0050] Accordingly, the node apparatus 100 in Figure 1
receives the pause frame from the port 101-j, and recognizes
that the second node apparatus is currently in the busy state.
That is, the node apparatus 100 recognizes that frames should
not be transmitted to the port 101-j during the above-mentioned
certain time period, which is designated in the pause frame,
because the second node apparatus is unable to process the frames

due to an overload even if the frames are transmitted to the
port 101-j.
[0051] On the basis of the above recognition, the routing
engine 102 of the node apparatus 100 then sets the state RPji,
which corresponds to the port 101-j in the routing table 103,
to "P". As apparent from the above, the "P" state is an example
of a state indicating "transmission is not feasible".
[0052] "D" indicates a link-down state. The state RPj± turns
into "D" in cases where a signal carried on a cable is not
electrically detectable at the port 101-j, including a case such
as that where a cable connected to the port 101-j is physically
broken. The "D" state is also an example of a state indicating
"transmission is not feasible".
[0053] Note that the "P" and "D" states are concepts
independent of the destination of a frame. However, according
to this embodiment, there is a case where, in the routing table
103, part of the states RPJX to RPjn of a certain port (e.g.,
the port 101-j) is/are being set to "P" and the other part of
them is/are being set to a state other than "P". Likewise,
according to this embodiment, there is a case where part of the
states RPji to RPjn of a certain port (e.g. , the port 101-j) is/are
being set to "D", and the other part of them is/are being set
to a state other than "D". This is because, in this embodiment,
an entry of the routing table 103 is updated at an opportunity
of reception of a frame, as will be described later, and thus
there are time differences in the timing of the updating among
the entries.
[0054] Subsequently, the other tables illustrated in Figure
1 are described.
The broadcast management table 104 is a table that stores
a "MAC-SA" (Media Access Control Source Address) and a "time"
in association with each other. Figure 1 exemplifies the
broadcast management table 104 including entries, the number
of which is "a".
[0055] As will be described later with reference to Figure

11B, each entry is one that has been added to the broadcast
management table 104 on the basis of a broadcast frame received
by the node apparatus 100. Although not illustrated in the
drawings, each entry is deleted from the broadcast management
table 104 if not accessed for a prescribed time period, as will
be described later.
[0056] For instance, in the i-th (1 ≤ i ≤ a) entry in the
broadcast management table 104, a MAC address MACSAi and a time
Ti are associated with each other. Here, the MAC address MACSAi
is the MAC address of a source node apparatus of a broadcast
frame received by the node apparatus 100. The time Ti is a time
at which the routing engine 102 processed this broadcast frame.
[0057] When the node apparatus 100 receives a broadcast frame
from any of ports 101-1 to 101-x, the routing engine 102 refers
to the broadcast management table 104. The routing engine 102
then determines whether to discard the received broadcast frame
and thereby prevent congestion or to relay the received broadcast
frame (i.e., to transmit it to a port other than the port from
which this broadcast frame has been received). As will be
described later in detail, some embodiments make it possible
to prevent congestion of a broadcast frame only with the loop
detection table 105.
[0058] The loop detection table 105 is a table to associate
a "receiving port number" and "x" number of "port states", which
respectively indicate states of the ports 101-1 to 101-x, with
a pair of a "source" and an "FID" (Frame IDentification) and
to store them.
[0059] Figure 1 exemplifies the loop detection table 105
including "m" number of entries. As will be described later
with reference to Figures 11B and 11C, each entry is added to
the loop detection table 105 on the basis of a frame received
by the node apparatus 100. In addition, as will be described
later with reference to Figures 14 and 15, each entry is deleted
from the loop detection table 105 after a prescribed time period
has elapsed.

[0060] For instance, in the i-th (1 ≤ i ≤ m) entry in the
loop detection table 105, a receiving port number RCVPNi and
states LPii to LPxi of the respective ports 101-1 to 101-x are
associated with a pair of a source GSi and FIDi.
[0061] Here, the source GSi is identification information for
identifying a node apparatus being the source of a frame in a
wired ad hoc network including a plurality of node apparatuses
each equivalent to the node apparatus 100, and more specifically
is a node ID. The FIDi is identification information that the
source node apparatus (i.e., a node apparatus whose node ID is
GSi) has assigned to a frame to be transmitted and that uniquely
identifies the frame. The FIDi may be, for instance, a sequence
number.
[0062] When the node apparatus 100 receives a frame at any
of the ports 101-1 to 101-x and tries to relay the received frame,
the routing engine 102 refers not only to the routing table 103
but also to the loop detection table 105. The routing engine
102 then selects the port for transmitting the received frame
also on the basis of the loop detection table 105.
[0063] The receiving port number RCVPNi of the i-th entry in
the loop detection table 105 indicates a port from which a frame
corresponding to the i-th entry has been received by the node
apparatus 100. More specifically, in a case where a frame to
which an FID being FIDi is assigned by a source node apparatus
(i.e., a node apparatus identified by a node ID being GSi) is
received at the port 101-r (1 ≤ r ≤ x) of the node apparatus
100, the receiving port number RCVPNi is "r".
[0064] In the i-th entry in the loop detection table 105,
the state LPji of the j-th (1 ≤ j ≤ x) port 101-j is any of three
states, namely, "U", "E", and "L".
[0065] As described on the routing table 103, "U" indicates
an in-use state. More specifically, in the loop detection table
105, the state LPji is "U" if the following two conditions hold.
• The node apparatus 100 has relayed a frame to which an
FID being FIDi is assigned by a source node apparatus identified

by a node ID being the source GSi.
[0066] • In the relaying, the port 101-j has been used as
a destination port in the node apparatus 100.
As described on the routing table 103, "E" indicates an unused
state. In the loop detection table 105, more specifically, the
meaning is as follows. That is, the state LPji is "E" if the
destination port is not the port 101-j when the node apparatus
100 has relayed a frame to which an FID being FIDi is assigned
by a source node apparatus identified by a node ID being the
source GSi.
[0067] As described on the routing table 103, "L" indicates
a loop state. In the loop detection table 105, more specifically,
the state LPji turns into "L" if the following two conditions
hold.
[0068] • The node apparatus 100 has relayed a frame to which
an FIDbeing FIDi is assignedby a source node apparatus identified
by a node ID being the source GSi, while using the port 101-j
as the destination port.
[0069] • This frame is received by the node apparatus 100
again after the relaying.
In the routing table 103, the "U" and "E" states are common
in that they mean "transmission is feasible", namely, "selectable
as a destination port". However, in the loop detection table
105, the "U" and "E" states are sharply discriminated from each
other as follows. More specifically, the "U" state in the loop
detection table 105 indicates a target to be transitioned to
the "L" state, which means "transmission is not feasible", if
a loop is detected hereafter. On the other hand, the "E" state
in the loop detection table 105 indicates that, even if a loop
is detected hereafter, it is still selectable as a destination
port at the time of detection of the loop.
[0070] Next, an overview of a distributed coordination
operation in a network including a plurality of node apparatuses
each including components as with Figure 1 is described with
reference to Figures 2A to 4 along with an example of a frame

format.
[0071] Figure 2A is a diagram which illustrates an example
of a network to which the node apparatus of this embodiment is
applied.
A wired ad hoc network 200 in Figure 2A includes a plurality
of node apparatuses 100a to lOOi. In the wired ad hoc network
200, the node apparatuses 100a to lOOi are physically connected
by cables (e.g., metal wire cables, such as copper cables, or
optical fiber cables) in a mesh configuration (i.e., in a grid
configuration).
[0072] It is a matter of course that any physical connection
topology in the wired ad hoc network may be adopted according
to embodiments, and the mesh configuration is not necessarily
adopted.
Each of the node apparatuses 100a to lOOi includes components
as with Figure 1. Note that the node apparatus 100 includes
the "x" number of ports 101-1 to 101-x in Figure 1 and that Figure
2A specifically illustrates an example where x=4.
[0073] For instance, the node apparatus 100a includes four
ports 101a-l to 101a-4. This is applicable to the other node
apparatuses 100b to lOOi. Components analogous to each other
may be assigned reference signs, such as "101-1", "101a-l", and
"101b-l", that are identical with each other except for suffixes
such as "a" and "b", and the detailed description thereof may
be omitted.
[0074] The physical topology in the mesh configuration
illustrated in Figure 2A is specifically realized by the
following cable wiring.
• The node apparatuses 100a and lOOd are connected by a link
215 between the ports 101a-l and 101d-l.
• The node apparatuses 100a and 100b are connected by a link
216 between the ports 101a-4 and 101b-l.
• The node apparatuses 100b and lOOe are connected by a link
217 between the ports 101b-2 and 101e-2.
• The node apparatuses 100b and 100c are connected by a link

218 between the ports 101b-4 and 101c-l.
• The node apparatuses 100c and lOOf are connected by a link
219 between the ports 101c-3 and 101f-3.
• The node apparatuses lOOd and lOOg are connected by a link
221 between the ports 101d-2 and 101g-2.
• The node apparatuses lOOd and lOOe are connected by a link
222 between the ports 101d-4 and 101e-l.
• The node apparatuses lOOe and lOOh are connected by a link
223 between the ports 101e-3 and 101h-3.
• The node apparatuses lOOe and lOOf are connected by a link
224 between the ports 101e-4 and 101f-l.
• The node apparatuses lOOf and lOOi are connected by a link
225 between the ports 101f-4 and 101i-4.
• The node apparatuses lOOg and lOOh are connected by a link
226 between the ports 101g-4 and 101h-l.
• The node apparatuses lOOh and lOOi are connected by a link
227 between the ports 101h-4 and 101i-l.
It is a matter of course that a mesh topology equivalent
to that in Figure 2A is also realizable in some embodiments by
connecting pairs of ports, which are other than the pairs
exemplified in Figure 2A, using cables.
[0075] Incidentally, in the example of Figure 2A, the wired
ad hoc network 200 is not an isolated network, but is connected
to another network (hereinafter also referred to as an "external
network") such as a LAN and a WAN (Wide Area Network).
[0076] More specifically, the node apparatuses 100a to lOOi
in this embodiment include not only the components illustrated
in Figure 1 but also respective general LAN ports 106a to 106i
as connection interfaces between the wired ad hoc network 200
and the external network. In Figure 2A, the general LAN ports
106a to 106i are hatched, which distinguishes them from the wired
ad hoc network ports 101a-l to 101i-4. The general LAN ports
106a to 106i in this embodiment are wired LAN port. However,
in some embodiments, wireless LAN interfaces may be adopted
instead.

[0077] For instance, in the example of Figure 2A, the wired
ad hoc network 200 is connected to the external network as follows .
That is, an L2SW (Layer 2 Switch) 202 connected to a PC (Personal
Computer) 201 via a link 211 is connected to the respective general
LAN ports 106a and 106b of the node apparatuses 100a and 100b
via respective links 212 and 213. In addition, PCs 203, 205,
and 206 are connected to the respective general LAN ports 106c,
106g, and 106h of the node apparatuses 100c, lOOg and lOOh via
respective links 214, 228 and 229.
[0078] The L2SW 202 may further be connected to a router and/or
another PC(s), which are not illustrated in the drawings. The
PCs 203, 205 and 206may also be connected to another external
network, which is not illustrated in the drawings.
[0079] For convenience of description, Figure 2A exemplifies
the wired ad hoc network 200 including the nine node apparatuses
100a to 1001. However, in some embodiments, a wired ad hoc
network may include a large number of node apparatuses, such
as several thousands to several hundred thousands of them.
[0080] For instance, the wired ad hoc network of this
embodiment may be applied to a sensor network, which is a network
for collecting various pieces of information from a large number
of sensors disposed here and there. In this case, the wired
ad hoc network may include a large number of node apparatuses
in the order of several thousands to several hundred thousands
corresponding to the large number of sensors. In the sensor
network, arbitrary types of sensors, such as image sensors,
temperature sensors, humidity sensors, pressure sensors, and
acceleration sensors, are used.
[0081] Figure 2A also illustrates one example in which the
wired ad hoc network of this embodiment is applied to a sensor
network. More specifically, in Figure 2A, sensors 204 and 207
each equipped with a LAN interface are connectedto the respective
general LAN ports 106e and 106i of the node apparatuses lOOe
and lOOi via respective links 220 and 230. The sensors 204 and
207 output data representing sensed results as Ethernet frames

via the LAN interfaces. The types of the sensors 204 and 207
are arbitrary.
[0082] Use of the wired ad hoc network of this embodiment
as illustrated in Figure 2A enables construction of a sensor
network even in a harsh environment because it is often the case
that wired communication is feasible even in an environment in
which wireless communication is difficult.
[0083] For instance, sensors and node apparatuses connected
to the sensors via general LAN ports may be embedded in earth
including fields and escarpments, in water including paddy fields,
rivers and sea, and/or in structures including walls and pillars
of buildings. Even in such cases, the node apparatuses are
capable of surely communicating with other node apparatuses via
wired connection. Accordingly, adoption of the wired ad hoc
network of this embodiment enables construction of a sensor
network even in environments, such as in earth, water and
structures, where wireless communication is difficult.
[0084] The network configuration in Figure 2A has been
described above. Next, an overview of the distributed
coordination operation in the wired ad hoc network 200 in Figure
2A will be described. Detailed operations of the individual
node apparatuses 100a to lOOi will be described later with
reference to flowcharts of Figures 11A to 20.
[0085] Hereinafter, switching of a frame relay route in the
wired ad hoc network 200 in case of occurrence of a fault will
be described using an example in which the PC 203 transmits a
frame to the PC 205.
[0086] When the PC 203 transmits an Ethernet frame via the
link 214, the node apparatus 100c receives the Ethernet frame
and adds an ad hoc header to the received Ethernet frame.
[0087] According to autonomous distributed coordination,
which is realized by each node apparatus operating according
to the flowcharts of Figures 11A to 20, the ad hoc frame is relayed
from the node apparatus 100c to the node apparatus lOOg in the
wired ad hoc network 200. The node apparatus lOOg then removes

the ad hoc header from the ad hoc frame. The Ethernet frame
acquired as a result thereof is outputted from the node apparatus
lOOg to the PC 205 via the link 228.
[0088] Assume that, for instance, a route formed by the links
218, 216, 215, and 221 is selected as the relay route from the
node apparatus 100c to the node apparatus lOOg in a case where
each link illustrated in Figure 2A is normal. This route is
selected by the following distributed coordination.
[0089] The node apparatus 100c, having received the frame
from the PC 203, selects the port 101c-l connected to the node
apparatus 100b via the link 218 as the destination port of the
frame. Then, the node apparatus 100b, having received the frame
from the node apparatus 100c, selects the port 101b-l connected
to the node apparatus 100a via the link 216 as the destination
pert of the frame.
[0090] The node apparatus 100a, having received the frame
from the node apparatus 100b, then selects the port 101a-l
connected to the node apparatus lOOd via the link 215 as the
destination port of the frame. Subsequently, the node apparatus
lOOd, having received the frame from the node apparatus 100a,
selects the port 101d-2 connected to the node apparatus lOOg
via the link 221 as the destination port of the frame.
[0091] In the meantime, assume that a physical fault occurs
in the link 215 at a certain point in time. If the PC 203 tries
to transmit a frame to the PC 205 again after the occurrence
of the fault in the link 215, the relay route in the wired ad
hoc network 200 is switched according to the following
distributed coordination control.
[0092] The node apparatus 100c selects the port 101c-l as
the destination port of the frame, as with the normal case. The
node apparatus 100b selects the port 101b-l as the destination
port of the frame, also as with the normal case. The node
apparatus 100a then receives the frame.
[0093] However, as illustrated in Figure 2A, the node
apparatus 100a is only connected to the two node apparatuses,

namely, the node apparatuses 100b and lOOd in the wired ad hoc
network 200, and the fault is occurring in the link 215 with
the node apparatus lOOd. Accordingly, the node apparatus 100a
is unable to transmit the frame received from the node apparatus
100b via the link 216 to a node apparatus in the wired ad hoc
network 200 other than the node apparatus 100b via a normal link.
[0094] Thus, the node apparatus 100a returns the frame, via
the link 216, to the node apparatus 100b in order to notify it
that the node apparatus 100a itself is unable to forward the
frame. The node apparatus 100b then receives the frame that
has been transmitted by the node apparatus 100b itself.
[0095] Accordingly, the node apparatus 100b recognizes that
the port 101b-l is in the "L" state (i.e., the loop state) in
the port state information that is associated with the node
apparatus lOOg in the routing table 103. More specifically,
the node apparatus 100b recognizes "in a case of relaying a frame
whose final destination in the wired ad hoc network 200 is the
node apparatus lOOg, the frame should not be transmitted to the
port 101b-l, which is connected to the node apparatus 100a".
[0096] According to the above-mentioned recognition, the node
apparatus 100b searches for a normal port that is other than
the port 101b-l and that is usable as the destination port of
the frame. In this case, the port 101b-2, which is connected
to the node apparatus lOOe via the normal link 217, is found.
[0097] Accordingly, the node apparatus 100b transmits the
frame, which has been returned from the node apparatus 100a,
to the port 101b-2 at this time.
The node apparatus lOOe then receives the frame. In the
example of Figure 2A, each link connected to the node apparatus
lOOe is normal. Accordingly, the node apparatus lOOe is enabled
to select, for instance, the port 101e-3, which is connected
to the node apparatus lOOh via the link 223, as the destination
port of the frame.
[0098] As a result of the selection, the frame is transmitted
from the port lOle-3, and received by the node apparatus lOOh

via the link 223. The node apparatus lOOh then selects the port
101h-l, which is connected to the node apparatus lOOg via the
normal link 226, and transmits the frame from the port 101h-l.
[0099] As a result, the node apparatus lOOg is enabled to
receive the frame. As described above, even in a case where
the fault occurs on the link 215 on the route having been used
so far, the distributed coordination of the node apparatuses
in the wired ad hoc network 200 automatically switches the route .
[0100] Since the routing table 103 includes the entries on
a destination-by-destination basis as illustrated in Figure 1,
there may be a case where the node apparatus 100b selects the
port 101b-l as the destination port depending on the destination
of the frame. In other words, the node apparatus 100b is still
enabled to transmit a frame, which is addressed to the node
apparatus 100a, from the port 101b-l even1in a case where the
node apparatus 100b recognizes the port 101b-l as the "L" state
corresponding to the node apparatus lOOg as described above.
Thus, according to this embodiment, feasibility of transmitting
a frame from a certain port is managed depending on the destination
of the frame.
[0101] Next, referring to Figure 2B, the switching of the
route described above will be described from another viewpoint.
Figure 2B is a conceptual diagram which explains alternative
route finding in case of occurrence of a fault. Figure 2B
illustrates a search tree 300 whose root node is the node apparatus
100c being the source in the wired ad hoc network 200 in the
example above, and whose each leaf node is the node apparatus
lOOg being the destination in the wired ad hoc network 200.
[0102] For the sake of convenience, let a certain node in
the search tree 300 be referred to as a "first node", and let
a node apparatus which corresponds to the first node and which
is in the wired ad hoc network 200 be referred to as a "first
node apparatus". Also assume that, in an entry which is in the
routing table 103 in the first node apparatus and which
corresponds to the node apparatus lOOg being the destination,

the state of another node apparatus (for the sake of convenience,
referred to as a "second node apparatus") is the "U" or "E" state.
In this case, in the search tree 300, the first node has a child
node corresponding to the second node apparatus.
[0103] The relay route, which has been used before the
occurrenceof the fault in the link215 and which has been described
pertaining to Figure 2A, is a route having been found
corresponding to the search path 301 in Figure 2B as a result
of the distributed coordination in the wired ad hoc network 200.
The search path 301 is a path from the node apparatus 100c, which
is the root node in the search tree 300, to the node apparatus
lOOg, which is the leaf node in the search tree 300, via the
node apparatuses 100b, 100a, and lOOd.
[0104] The route switching (i.e., the alternative route
finding), which is made in case of occurrence of the fault in
the link 215 and which has been described pertaining to Figure
2A, corresponds to a search path 302 in Figure 2B. As illustrated
in Figure 2B, the search path 302 includes backtracking 303.
[0105] That is, the route switching described pertaining to
Figure 2A may be regarded as corresponding to that a depth-first
search including the backtracking 303 in the search tree 300
is performed in the entire wired ad hoc network 200. If the
search path 302 including the backtracking 303 reaches the node
apparatus lOOg being the leaf node, the search succeeds and a
relay route of the frame in the wired ad hoc network 200 is
established.
[0106] The occurrence of the backtracking 303 at the node
apparatus 100a in the search tree 300 corresponds to that the
node apparatus 100a returns the frame to the node apparatus 100b
in the description of Figure 2A. The path, which is from the
root node to the leaf node in the search tree 300 and which has
finally been found according to the search path 302 in the search
tree 300, corresponds to a new frame relay route from the node
apparatus 100c to the node apparatus lOOg in the wired ad hoc
network 200. That is, as illustrated in Figure 2B, the route

passing through the node apparatuses 100c, 100b, 100e, 100h,
and 100g is found as the new frame relay route.
[0107] As understood from the above description, as long as
at least one route from the node apparatus 100c being the source
to the node apparatus 100g being the destination exists in the
wired ad hoc network 200, the route is surely found as a result
of the depth-first search.
[0108] Next, difference in frame format between the inside
and outside of the wired ad hoc network 200 in Figure 2A will
be described with reference to Figure 3. Figure 3 is a diagram
which explains a frame format. Note that the "ad hoc frame"
in this embodiment is a frame transmitted and received between
the node apparatuses in the wired ad hoc network.
[0109] As illustrated in Figure 3, the ad hoc frame 400 is
a frame in a format in which an "ad hoc header", which is a header
in a specific format defined in this embodiment, is prepended
to a general Ethernet frame 420. In other words, the ad hoc
header 410 is information that functions as an interface between
the Ethernet being an external network and the wired adhoc network
of this embodiment.
[0110] The ad hoc header 410 of this embodiment is 14 bytes
long as illustrated in Figure 3, and includes respective fields
of a GD 411, a GS 412, a type 413, an FID 414, a TTL 415, a length
416, and an FCS 417 . Although detailed description will be made
later, the ad hoc header 410 is given by a node apparatus in
the wired ad hoc network. Hereinafter, for simplification of
notation, for instance, the "GD field" may simply be denoted
by a "GD".
[0111] Designated in the three-byte GD (Global Destination)
411 is identification information (e.g., a node ID) for
identifying a node apparatus, which is the source of the frame
in the wired ad hoc network, uniquely in the wired ad hoc network.
Designated in the three-byte GS (Global Source) 412 is
identification information (e.g., a node ID) for identifying
a node apparatus, which is the destination of the frame in the

wired ad hoc network, uniquely in the wired ad hoc network.
[0112] For instance, in the example in Figure 2A, in a case
where a frame from the PC 203 addressed to the PC 205 is transmitted
via the wired ad hoc network 200, the node apparatus being the
source in the wired ad hoc network 200 is the node apparatus
100c, which is connected to the PC 203 . The node apparatus being
the destination in the wired ad hoc network 200 is the node
apparatus 100g, which is connected to the PC 205.
[0113] Accordingly, in this case, in the ad hoc frame 400
transmitted within the wired ad hoc network 200, the node ID
of the node apparatus 100c is designated inthe GD 411. Meanwhile,
the node ID of the node apparatus 100g is designated in the GS
412.
[0114] The values designated in the GD 411 and the GS 412
are, for instance, node IDs preliminarily assigned to the
respective node apparatuses as described above. The node IDs
may be identification numbers. Alternatively, the
least-significant three bytes of the MAC address of the node
apparatus may also be used as identification information for
the GD 411 and the GS 412 if each node apparatus in the wired
ad hoc network is a product of the same vendor. This is because
the most-significant three bytes of the MAC address indicate
an OUI (Organizationally Unique Identifier) and the
least-significant three bytes are unique in the vender.
[0115] The implication of the term "global" in the names of
"GD" and "GS" will be described according to the example in Figure
2A, as follows.
When the ad hoc frame 400 is transmitted from the node
apparatus 100c toward the node apparatus 100g in the wired ad
hoc network 200, the immediate destination from the node
apparatus 100c is a node apparatus (e.g., the node apparatus
100b) adjacent to the node apparatus 100c. Thus, when explicitly
discriminating the node apparatus 100b being the immediate
destination from the node apparatus 100g being the final
destination, the node apparatus 100g, which is the final

destination within the wired ad hoc network 200, is referred
to as a "global destination". Meanwhile, the node apparatus
100b, which is directly connected by the cable to the node
apparatus 100c and which is adjacent to the node apparatus 100c,
is referred to as a "local destination".
[0116] Accordingly, with respect to the node apparatus 100b,
the node apparatus 100c is the global source and is also the
local source. With respect to the node apparatus 100g, the node
apparatus 100c is the global source, but is not the local source
(because the node apparatuses 100g and 100c are not directly
connected to each other). As also apparent in Figure 2A, in
the wired ad hoc network, a local destination corresponds to
a destination port of a frame in a one-to-one manner, and a local
source corresponds to a port from which the frame is received
in a one-to-one manner.
[0117] Returning to the description in Figure 3, a value
identifying the type (category) of the ad hoc frame 400 is
designated in the four-bit type 413. For instance, according
to embodiments, a value for distinguishing some types, which
are exemplified below, from each other is designated in the type
413.
[0118] • a normal frame (a type of the ad hoc frame 400 including
the Ethernet frame 420 subsequent to the ad hoc header 410)
• a pause frame (a type of the ad hoc frame including data,
which is not the Ethernet frame 420 and which is in a prescribed
format, subsequent to the ad hoc header 410)
• other some types of control frames (types of the ad hoc
frames each including data, which is not the Ethernet frame 420
and which is in a prescribed format, subsequent to the ad hoc
header 410)
• a broadcast frame received by the node apparatus via the
general LAN port (a type of the ad hoc frame 400 including the
Ethernet frame 420 subsequent to the ad hoc header 410)
• a broadcast frame originated in the node apparatus itself
(a type of the ad hoc frame 400 including the Ethernet frame

420 subsequent to the ad hoc header 410)
Meanwhile, designated in the 12-bit FID (Frame
Identification) 414 is identification information assigned to
the ad hoc frame 400 by the node apparatus being the source
identified by the value of the GS 412. The value designated
in the FID 414 may be, for instance, a sequence number generated
by the node apparatus using a built-in counter circuit, which
is not illustrated in the drawings.
[0119] Meanwhile, the valid duration of the ad hoc frame 400
is designated in the two-byte TTL (Time To Live) 415 in terms
of the number of hops in the wired ad hoc network. The value
of the TTL 415 is counted down by a node apparatus every time
the ad hoc frame 400 hops in the wired ad hoc network.
[0120] In this embodiment, the initial value of the TTL 415
(i.e., a value assigned by the node apparatus being the source
identified by the value in the GS 412) is a value in which all
bits in the two bytes are set to "1" (i.e., 65535) . Since such
a large value is set as the initial value of the TTL 415, this
embodiment is preferably applicable to a wired ad hoc network
including several hundred thousands of the node apparatuses.
It is a matter of course that the initial value of the TTL 415
is arbitrary according to embodiments.
[0121] Designated in the two-byte length 416 is a value
counting the length of a part other than the ad hoc header 410
(e.g. , in the case of the ad hoc frame 400 exemplified in Figure
3, the length of the Ethernet frame 420) in units of bytes.
[0122] Designated in the two-byte FCS (Frame Check Sequence)
417 is an error detecting code, such as a CRC (Cyclic Redundancy
Check) code, acquired from the fields from the GD 411 to the
length 416 in the ad hoc header 410.
[0123] The Ethernet frame 420 includes the following fields.
• a six-byte MAC-DA (Media Access Control Destination
Address) 421
• a six-byte MAC-SA (Media Access Control Source Address)
422

• a two-byte type/length 423
• an L3 (Layer 3) packet 424 of a variable length from 46
to 1500 bytes
• a four-byte FCS 425
Since the format of the Ethernet frame 420 is well-known,
detailed description thereof is omitted. However, described
below is a specific example, for instance, in which a frame is
transmitted from the PC 203 to the PC 205 via the wired ad hoc
network 200 in the example in Figure 2A.
[0124] That is, the PC 203 generates an Ethernet frame 420,
in which the MAC address of the PC 205 is designated in the MAC-DA
421 and in which the MAC address of the PC 203 itself is designated
in the MAC-SA 422, and transmits the Ethernet frame 420 to the
node apparatus 100c via the link 214. The node apparatus 100c
then prepends the ad hoc header 410 as described above to the
Ethernet frame 420, thereby creating an ad hoc frame 400.
[0125] The ad hoc frame 400 is transmitted to the node apparatus
100g in the wired ad hoc network 200. The node apparatus 100g
outputs the Ethernet frame 420, which is other than the ad hoc
header 410, to the PC 205 via the link 228.
[0126] In a case where a value indicating the pause frame
or another control frame is designated in the type 413 in the
ad hoc header 410, the ad hoc frame includes a payload, which
is other than the Ethernet frame 420 and which is in a prescribed
format, subsequent to the ad hoc header 410.
[0127] Next described with reference to Figure 4 is the
behavior of the entire wired ad hoc network of this embodiment
in a case where a broadcast frame is transmitted. Figure 4 is
a diagram which explains that the broadcast frame is prevented
from congestion.
[0128] A wired ad hoc network 250 in Figure 4 includes a
plurality of node apparatuses 100p to 100t each including
components analogous to those of the node apparatus 100 in Figure
1. The node apparatuses in the wired ad hoc network 250 are
physically connected as follows.

[0129] • The node apparatus 100p is connected to the node
apparatus 100q.
• The node apparatus 100q is connected not only to the node
apparatus 100p but also to the node apparatuses 100r and 100s.
[0130] • The node apparatus 100r is connected not only to
the node apparatus 100q but also to the node apparatuses 100s
and 100t.
• The node apparatus 100s is connected not only to the node
apparatuses 100q and 100r but also to the node apparatus 100t.
[0131] • The node apparatus 100t is connected to the node
apparatuses 100r and 100s, as described above.
For instance, assume that a broadcast frame is transmitted
from the node apparatus 100p in the wired ad hoc network 250
with the physical topology as described above. Note that the
"broadcast frame" is an Ethernet frame or an ad hoc frame in
which the broadcast address (i.e., an address, all of whose bits
have a value of 1) is designated in the MAC-DA 421.
[0132] For instance, the node apparatus 100p is connected
to an external device (such as a PC or a sensor) , which is not
illustrated in Figure 4, via a general LAN port. This external
device transmits an Ethernet frame 420, in which the broadcast
address is designated in the MAC-DA 421, to the node apparatus
100p.
[0133] The node apparatus 100p then prepends an ad hoc header
410 to the Ethernet frame 420, thereby creating an ad hoc frame
400. The ad hoc frame 400 created here is a broadcast frame.
The node apparatus 100p transmits the created ad hoc frame 400
to the node apparatus 100q in step S101 in order to broadcast
it in the wired ad hoc network 250.
[0134] The node apparatus 100q then transmits the broadcast
frame to the node apparatus 100r in step S102, and also transmits
the broadcast frame to the node apparatus 100s in step S103.
[0135] The node apparatus 100r, having received the broadcast
frame from the node apparatus 100q, in turn transmits the
broadcast frame to the node apparatus 100s in step S104 . However,

since the node apparatus 100s has already received the broadcast
frame f romthe node apparatus 100q in step S103, the node apparatus
100s discards the broadcast frame received from the node
apparatus 100r in step S104.
[0136] More specifically, the node apparatus 100s stores the
value of the MAC-SA 422 in the broadcast frame received in step
S103 in the broadcast management table 104 in Figure 1 in
association with the time at which the reception process in step
S103 was performed. Steps S103 and S104 are performed
substantially simultaneously.
[0137] Accordingly, upon receipt of the broadcast frame in
step S104, the node apparatus 100s searches the broadcast
management table 104 using the value of the MAC-SA 422 in the
received broadcast frame as a key and thereby acquires the
associated time. In a case where the difference between the
current time at which step S104 is performed and the acquired
time is shorter than a prescribed time period, the node apparatus
100s discards the broadcast frame received in step S104.
[0138] As illustrated in step S105, the node apparatus 100s
also transmits the broadcast frame received from the node
apparatus 100q in step S103 to the node apparatus 100r. However,
the node apparatus 100r has already received the broadcast frame
from the node apparatus 100q in step S102, and has already
transmitted the broadcast frame in step S104 . Accordingly, the
node apparatus 100r discards the broadcast frame received from
the node apparatus 100s in step S105. The discarding by the
node apparatus 100r is specifically realized as follows.
[0139] The node apparatus 100r has already transmitted the
broadcast frame to the node apparatus 100s in step S104.
Accordingly, the node apparatus 100r has an entry including the
values of the GS 412 and the FID 414 in the broadcast frame,
in the 100p detection table 105 in Figure 1.
[0140] In step S105, the node apparatus 100r searches the
100p detection table 105 using the values of the GS 412 and the
FID 414 of the broadcast frame received from the node apparatus

100s as search keys. The node apparatus 100r then verifies that
the entry matching with the search keys exists. That is, the
node apparatus 100r recognizes that the broadcast frame
transmitted by the node apparatus 100r itself in step S104 100ps
in the wired ad hoc network 250 and returns to the node apparatus
100r itself. As a result of this recognition, in step S105,
the node apparatus 100r discards the broadcast frame received
from the node apparatus 100s.
[0141] In some embodiments, the transmission in step S105
may be performed before the reception in step S104. In this
case, the node apparatus 100s discards the broadcast frame on
the basis of the 100p detection table 105.
[0142] Meanwhile, as illustrated as step S106, the node
apparatus 100r transmits the broadcast frame received from the
node apparatus 100q in step S102 also to the node apparatus 100t.
[0143] In addition, as illustrated as step S107, the node
apparatus 100s transmits the broadcast frame received from the
node apparatus 100q in step S103 also to the node apparatus 100t.
However, the node apparatus 100t has already received the
broadcast frame in step S106. Accordingly, in step S107, the
node apparatus 100t refers to the broadcast management table
104 and discards the broadcast frame received from the node
apparatus 100s, as with the node apparatus 100s in step S104.
[0144] Meanwhile, as illustrated as step S108, the node
apparatus 100t transmits the broadcast frame received from the
node apparatus 100r in step S106 to the node apparatus 100s.
However, the node apparatus 100s has already transmitted the
broadcast frame in steps S105 and S107. Accordingly, in step
S108, the node apparatus 100s refers to the 100p detection table
105 and discards the broadcast frame received from the node
apparatus 100t, as with the node apparatus 100r in step S105.
[0145] In some embodiments, the reception in the step S108
may be performed before the transmission in the step S107.
However, also in this case, since the node apparatus 100s has
already transmitted the broadcast frame in step S105, the node

apparatus 100s is capable of appropriately discarding the
broadcast frame from the node apparatus 100t on the basis of
the 100p detection table 105.
[0146] As described above, according to this embodiment, each
individual node apparatus appropriately refers to the broadcast
management table 104 or the 100p detection table 105, and
determines whether to discard the received broadcast frame or
not. In addition, each individual node apparatus transmits the
broadcast frame to its adjacent node apparatuses only when not
discarding the broadcast frame.
[0147] As a result, the broadcast frame is delivered to all
the node apparatuses in the wired ad hoc network 250. Further,
a situation in which the broadcast frame continues to 100p
endlessly in the wired ad hoc network 250 (i.e., congestion of
the broadcast frame) is prevented. The prevention of the
congestion is a result of the distributed coordination by the
node apparatuses in the wired ad hoc network 250.
[0148] The overview of the distributed coordination in the
wired ad hoc network including the plurality of node apparatuses
each including components in Figure 1 has been described as above
with reference to Figures 2A to 4. Hereinafter, a specific
configuration and a specific operation of the individual node
apparatus to realize the distributed coordination as described
above will be described in detail.
[0149] Figure 5 is a hardware configuration diagram of the
node apparatus. Figure 5 illustrates a specific example of a
hardware configuration for realizing the node apparatus 100 in
Figure 1.
The node apparatus 100 in Figure 5 includes a plurality of
wired ad hoc network ports 101-1 to 101-x (which are simply
referred to as "ports" in Figure 1) for wired connection to
adjacent node apparatuses, which are not illustrated in Figure
5. The node apparatus 100 in Figure 5 also includes a general
LAN port 106 that is analogous to the general LAN ports 106a
to 106i exemplified in Figure 2A and that is for wired connection

to an external device (such as a PC, an L2SW, or a sensor) , which
is not illustrated in Figure 5.
[0150] Further, the node apparatus 100 in Figure 5 includes
a plurality of PHY (PHYsical layer) chips 111-1 to 111-x connected
to the respective wired ad hoc network ports 101-1 to 101-x.
The node apparatus 100 also includes an L2SW (Layer 2 Switch)
unit 112 connected to the general LAN port 106.
[0151] In addition, the node apparatus 100 includes an FPGA
(Field Programmable Gate Array) 113 that realizes the routing
engine 102 in Figure 1, and further includes an MPU
(Microprocessing Unit) 114 and various types of storage devices.
[0152] More specifically, the node apparatus 100 includes
a CAM (Content Addressable Memory) 115, an SRAM (Static Random
Access Memory) 116, and an SDRAM (Synchronous Dynamic Random
Access Memory) 117, which are accessible from the FPGA 113. In
this embodiment, an address acquired as a result of searching
the CAM 115 is an address of the SRAM 116. The routing table
103, the broadcast management table 104, and the 100p detection
table 105 in Figure 1 are realized by the CAM 115 and the SRAM
116 in this embodiment.
[0153] The node apparatus 100 further includes a DDR (Double
Data Rate) 2 SDRAM 118 anda f lashmemory 119, which are accessible
from the MPU 114. Stored in the flash memory 119, which is a
type of nonvolatile memories, are, for instance, the node ID
of the node apparatus 100 itself, the MAC address of the node
apparatus 100 itself, a firmware program(s) executed by the MPU
114, and the like. The DDR2 SDRAM 118 is used as a working area
when the MPU 114 executes the firmware program(s).
[0154] The node apparatus 100 further includes a PHY chip
120 connected to the L2SW unit 112.
The respective PHY chips 111-1 to 111-x are connected to
the FPGA 113 by respective Mils (Media Independent Interfaces)
121-1 to 121-x, which are interfaces between the physical layer
and the MAC sublayer. The L2SW unit 112 and the FPGA 113 are
also connected by an Mil 122.

[0155] The FPGA 113 and the MPU 114 are connected by a PCI
(Peripheral Component Interconnect) bus 123 and also by an Mil
124. The PHY chip 120 connected to the L2SW unit 112 is also
connected to the MPU 114 by an Mil 125.
[0156] In this embodiment, an internal connection port 114a
of the MPU 114 connected to the FPGA 113 by the Mil 124 and an
internal connection port 114b of the MPU 114 connected to the
PHY chip 120 by the Mil 125 are assigned respective MAC addresses.
[0157] The frame forwarding paths within the node apparatus
100 are various, as described as follows.
(1) A frame received from the wired ad hoc network port 101-i
(1 ≤ i ≤ x) is inputted into the FPGA 113 via the PHY chip 111-i
and the Mil 121-i. The frame is then outputted from the FPGA
113 to the PHY chip 111-j (1 ≤ j ≤ x) via the Mil 121-j, and
transmitted from the wired ad hoc network port 101-j.
[0158] (2) A frame received from the wired ad hoc network
port 101-i (1 ≤ i ≤ x) is inputted into the FPGA 113 via the
PHY chip 111-i and the Mil 121-i. The frame is then outputted
to the L2SW unit 112 via the Mil 122, and transmitted from the
general LAN port 106.
[0159] (3) A frame received from the wired ad hoc network
port 101-i (1 ≤ i ≤ x) is inputted into the FPGA 113 via the
PHY chip 111-i and the Mil 121-i. The frame is then outputted
to the MPU 114 via the Mil 124, and processed in the MPU 114.
[0160] (4) A frame received from the general LAN port 106
is inputted into the FPGA 113 via the L2SW unit 112 and the Mil
122. The frame is then outputted from the FPGA 113 to the PHY
chip 111-j (1 ≤ j ≤ x) via the Mil 121-j, and transmitted from
the wired ad hoc network port 101-j.
[0161] (5) A frame received from the general LAN port 106
is inputted into the MPU 114 via the L2SW unit 112, the PHY chip
120, and the Mil 125, and processed in the MPU 114.
[0162] (6) The MPU 114 generates a frame and outputs it to
the FPGA 113 via the Mil 124. The frame is then outputted from
the FPGA 113 to the PHY chip 111-j (1 ≤ j ≤ x) via the Mil 121-j,

and transmitted from the wired ad hoc network port 101-j.
[0163] (7) The MPU 114 generates a frame and outputs it to
the PHY chip 120 via the Mil 125. The frame is then outputted
from the PHY chip 120 to the L2SW unit 112 and further transmitted
from the general LAN port 106.
[0164] Note that it is possible for the L2SW unit 112 to learn
the MAC address of the internal connection port 114b of the MPU
114 from the frame received through the forwarding path (7).
Accordingly, if the value of the MAC-DA 421 in a frame received
from the general LAN port 106 is identical with the MAC address
of the internal connection port 114b, the L2SW unit 112 outputs
the frame to the PHY chip 120 through the above described
forwarding path (5). On the other hand, if the value of the
MAC-DA 421 of a frame received from the general LAN port 106
is different from the MAC address of the internal connection
port 114b, the L2SW unit 112 outputs the frame to the FPGA 113
through the above described forwarding path (4).
[0165] Next, functions of the node apparatus 100 including
various pieces of hardware in Figure 5 will be described.
Figure 6 is a functional configuration diagram of the node
apparatus. In Figure 6, relationships between functional
blocks and the hardware in Figure 5 are indicated by broken lines,
and parts pertaining to the general LAN port 106 are omitted.
[0166] As illustrated in Figure 6, the node apparatus 100
includes a receiving unit 131 to receive the ad hoc frame 400.
More specifically, the receiving unit 131 includes a plurality
of receiving ports 132-1 to 132-x to respectively receive the
ad hoc frames 400 from a plurality of node apparatuses (not
illustrated in Figure 6) adjacent to the node apparatus 100.
[0167] The i-th receiving port 132-i (1 ≤ i ≤ x) is realized
by the wired ad hoc network port 101-i and the PHY chip 111-i.
In Figure 6, for convenience of illustration, the wired ad hoc
network ports 101-1 to 101-x are collectively represented as
"ports 101", and the PHY chips 111-1 to 111-x are collectively
represented as VVPHY chips 111".

[0168] The node apparatus 100 also includes a received frame
controller 133 that classifies the ad hoc frame 400 received
by the receiving unit 131 according to information such as a
type and/or a destination.
The node apparatus 100 further includes an upper layer
processor 134 that processes the ad hoc frame 400 addressed to
the node apparatus 100 itself. The "upper layer" means a layer
higher than the MAC sublayer, on which the ad hoc frame 4 00 is
defined.
[0169] If the value of GD 411 in the ad hoc frame 400 received
by the receiving unit 131 is identical with the node ID of the
node apparatus 100 itself, the received frame controller 133
outputs the ad hoc frame 400 to the upper layer processor 134.
The upper layer processor 134 then extracts the Ethernet frame
420 from the ad hoc frame 400 and processes the Ethernet frame
420. The upper layer processor 134 is realized by the MPU 114,
the DDR2 SDRAM 118, and the flash memory 119.
[0170] The node apparatus 100 also includes a table storage
unit 135 realized by the CAM 115 and the SRAM 116. Stored in
the table storage unit 135 are the routing table 103, the broadcast
management table 104, and the 100p detection table 105 in Figure
1. Tables that will be described later with reference to Figures
7 to 9 are also stored in the table storage unit 135.
[0171] The node apparatus 100 further includes a table
controller 136 that controls the contents of the various tables
by creating, updating, and deleting entries in the various tables
stored in the table storage unit 135. More specifically, the
table controller 136 is capable of searching the CAM 115 and
accessing the address of the SRAM 116 acquired as the search
result.
[0172] The node apparatus 100 further includes a port monitor
137, a port selector 138, a timer 139, and a transmission frame
controller 140 (functions of these elements will be described
later).
[0173] The table controller 136 controls the contents of the

various tables according to control information provided from
the received frame controller 133, the port monitor 137, the
port selector 138, the timer 139, and the transmission frame
controller 140. Specific control will be described later with
reference to flowcharts of Figures 11A to 20.
[0174] In this embodiment, the received frame controller 133,
the table controller 136, the port monitor 137, the port selector
138, the timer 139, and the transmission frame controller 140
are realized by the FPGA 113.
[01*75] The node apparatus 100 further includes a transmitting
unit 141 to transmit the ad hoc frame 400.
More specifically, the transmitting unit 141 includes a
buffer unit 142 to buffer the ad hoc frame 400. For instance,
the buffer unit 142 may be realized by the SDRAM 117. In the
buffer unit 142, there may be areas segmented corresponding to
respective transmitting ports 143-1 to 143-x.
[0176] In some embodiments, the receiving unit 131 may further
include another buffer unit for reception.
The transmitting unit 141 also includes a plurality of
transmitting ports 143-1 to 143-x to each transmit the ad hoc
frame 400 to a node apparatus (not illustrated in Figure 6)
adj acent to the node apparatus 100 . The transmitting ports 143-1
to 143-x each read out the ad hoc frame 400 buffered in the buffer
unit 142 and transmit the read-out ad hoc frame 4 00 to the adjacent
node apparatus, which is not illustrated in Figure 6.
[0177] The i-th transmitting port 143-i (l≤i≤x) is realized
by the wired ad hoc network port 101-i and the PHY chip 111-i.
That is, common pieces of hardware realize the receiving port
132-i and the transmitting port 143-i.
[0178] Described next are functions of the port monitor 137,
the port selector 138, the timer 139, and the transmission frame
controller 140, description of which has been omitted above.
The port monitor 137 monitors the respective states of the
receiving ports 132-1 to 132-x in the receiving unit 131, the
buffer unit 142 in the transmitting unit 141, and the transmitting

ports 143-1 to 143-x in the transmitting unit 141.
[0179] For instance, the PHY chip 111-i may detect whether
the link connected to the wired ad hoc network port 101-i is
down or not, and may output the detection result to the FPGA
113 (or to a prescribed register accessible from the FPGA 113,
for instance). The port monitor 137 realized by the FPGA 113
monitors the detection result outputted from the PHY chip 111-i
and thereby monitors whether the receiving port 132-i and the
transmitting port 143-i are in the link-down state or not.
[0180] The port monitor 137 also monitors whether the frames
are excessively concentrated in the node apparatus 100 or not,
i.e., whether the node apparatus 100 is in the busy state or
not, by monitoring the usage percentage of the buffer unit 142.
[0181] The port selector 138 refers to the tables stored in
the table storage unit 135, and selects the destination port
of the frame outputted from the received frame controller 133.
The port selector 138 then gives the transmission frame
controller 140 an instruction on the selected port, and outputs
the frame to the transmission frame controller 140.
[0182] The port selector 138 is capable of acquiring an address
of the SRAM 116 by causing the table controller 136 to search
the CAM 115, and is capable of directly accessing the SRAM 116
using the acquired address to refer to or update data.
[0183] The timer 139 notifies the table controller 136 of
time information that is used for aging each entry in the various
tables.
The transmission frame controller 140 outputs the frame to
the buffer unit 142 . The transmission frame controller 140 also
performs control for transmitting the frame stored in the buffer
unit 142 from the transmitting port (one of the transmitting
ports 143-1 to 143-x) selected by the port selector 138.
[0184] The hardware configuration and functional block
configuration of the node apparatus 100 have been described above
with reference to Figures 5 and 6. Next, other data not
illustrated in Figure 1 will be described with reference to

Figures 7 to 10.
[0185] Figure 7 is a diagram which illustrates an example
of a pause state table. The pause state table 151 in Figure
7 is stored in, for instance, the table storage unit 135 in Figure
6. The detailed processing pertaining to the pause state table
151 will be described later with reference to Figures 11A and
16.
[0186] The pause state table 151 includes "x" number of entries
corresponding to the "x" number of ports 101-1 to 101-x,
respectively. For instance, in the entry corresponding to the
i-th (1 ≤ i ≤ x) port 101-i, the following three pieces of
information are associated with each other.
[0187] • the port number PNiof the port 101-i as identification
information for identifying the port 101-i
• a pause state PSi indicating whether or not the pause request
has been received from the adjacent node apparatus that is
directly connected via the port 101-i
• a counter County
The value of the pause state PSi is "P" (Pause) in a case
where the pause request has been received from the adjacent node
apparatus that is directly connected via the port 101-i and the
time period designated in the pause request has not elapsed yet.
[0188] On the other hand, the value of the pause state PSi
is "N" (Normal) in a case where the pause request has not been
received from the adjacent node apparatus that is directly
connected via the port 101-i. In addition, the value of the
pause state PSi turns into "N" when a notification of canceling
the pause request is received, and the value of the pause state
PSi also turns into "N" after the time period designated in the
pause request has elapsed. Since the pause state PSi is binary,
it can be represented in one bit.
[0189] When the pause request is received from the adjacent
node apparatus that is directly connected via the port 101-i,
the counter Counti is set to a value designated in the pause
frame and then counted down thereafter.

[0190] Figure 8 is a diagram which illustrates an example
of a port link state table. The port link state table 152 in
Figure 8 is stored in, for instance, the table storage unit 135
in Figure 6. The detailed processing pertaining to the port
link state table 152 will be described later with reference to
Figure 20.
[0191] The port link state table 152 includes "x" number of
entries corresponding to the "x" number of ports 101-1 to 101-x,
respectively. For instance, in the entry corresponding to the
i-th (1 ≤ i ≤ x) port 101-i, the port number PNi of the port
101-i and a link state Li, which indicates whether the link
connected to the port 101-i is in a communicable state or not,
are associated with each other.
[0192] If the link connected to the port 101-i is in the
communicable state, the value of the link state Li is "C"
(Connected) . If the link connected to the port 101-i is down
and not in the communicable state, the value of the link state
Li is "D" (Down) . Since the link state Li is binary, it can be
represented in one bit.
[0193] Figure 9 is a diagram which illustrates an example
of a MAC table. The MAC table 153 in Figure 9 is stored in,
for instance, the table storage unit 135 in Figure 6.
One entry in the MAC table 153 corresponds to one node
apparatus in the wired ad hoc network. Figure 9 exemplifies
the MAC table 153 including "b" number of entries. For instance,
in the i-th (1 ≤ i ≤ b) entry in the MAC table 153, the following
two pieces of information are associated with each other.
[0194] • a MAC address MACi of an external device (e.g., a
PC, a sensor, etc.) connected to a certain node apparatus in
the wired ad hoc network.
• an NIDi being a node ID that is unique in the wired ad hoc
network and that is assigned to the above-mentioned node
apparatus connected with the above-mentioned external device
with the MAC address MACi.
When receiving the ad hoc frame 4 00, the node apparatus 100

learns the pair of the MAC-SA 422 and the GS 412.
[0195] More specifically, the table controller 136 judges
whether an entry matching with the values of the MAC-SA 422 and
the GS 412 in the received ad hoc frame 400 exists in the MAC
table 153 or not. If the entry matching with the values of the
MAC-SA 422 and the GS 412 does not exist, the table controller
136 creates a new entry associating the value of the MAC-SA 422
and the value of the GS 412 with each other, and adds the new
entry to the MAC table 153.
[0196] In addition, when receiving the Ethernet frame 420
from a first external device connected via the general LAN port
106, the node apparatus 100 searches the MAC table 153 using
the value of the MAC-DA 421 as a search key.
[0197] If an entry is found as a result of the search, the
node apparatus 100 creates the ad hoc header 410 in which the
value of the node ID included in the found entry is set in the
GD 411, and prepends the ad hoc header 410 to the Ethernet frame
420. More specifically, the node apparatus 100 sets, in the
GD 411, the node ID of another node apparatus that is in the
wired ad hoc network and that is connected to a second external
device identified by the value of the MAC-DA 421.
[0198] For instance, when a frame is transmitted from the
PC 203 to the PC 205 in the example in Figure 2A, the node apparatus
100b on the transmission route in the wired ad hoc network 200
learns the MAC table 153 as follows.
[0199] In the above example, in the ad hoc frame 400 received
by the node apparatus 100b from the node apparatus 100c, the
MAC address of the PC 203 is designated in the MAC-SA 422 and
the node ID of the node apparatus 100c is designated in the GS
412. Accordingly, the node apparatus 100b learns the
association between the MAC address of the PC 203 and the node
ID of the node apparatus 100c, and adds an entry to the MAC table
153 as necessary.
[0200] Meanwhile, assume that the node apparatus 100b
receives a new Ethernet frame 420 addressed to the PC 203, for

instance from the PC 201 via the L2SW 202, after the above learning.
The node apparatus 100b then searches the MAC table 153 using
the value of the MAC-DA 421 in the new Ethernet frame 420 as
a search key.
[0201] That is, the node apparatus 100b searches the MAC table
153 using the MAC address of the PC 203 as the search key. As
a result, the node apparatus 100b--acquires the node ID of the
node apparatus 100c, which is in the wired ad hoc network 200
and which is connected to the PC 203. Accordingly, the node
apparatus 100b generates an ad hoc header 410 in which the node
ID of the node apparatus 100c is designated in the GD 411, and
prepends the ad hoc header 410 to the new Ethernet frame 420,
thereby generating a new ad hoc frame 400. The node apparatus
100b then transmits the new ad hoc frame 400.
[0202] The node apparatus 100c may operate for instance as
follows, if it receives the Ethernet frame 420 addressed to the
PC 205 from the PC 203 before learning the association between
the MAC address of the PC 205 and the node ID of the node apparatus
100g.
[0203] That is, the node apparatus 100c may broadcast, into
the wired ad hoc network 200, an ad hoc frame that is used for
inquiring a node ID and that includes an ad hoc header 410 in
which a prescribed value to indicate broadcasting is designated
in the GD 411. The payload of the ad hoc frame used for inquiring
the node ID includes the MAC address of the PC 205.
[0204] Due to the distributed coordination mechanism
exemplified in Figure 4, the ad hoc frame used for inquiring
the node ID is not congested in the wired ad hoc network 200.
The node apparatus 100g, which receives the inquiry by the
frame having been broadcast in the wired ad hoc network 200,
replies to the node apparatus 100c. That is, the node apparatus
100g returns a special ad hoc frame for control use and for
notification of the association between the MAC address of the
PC 205 and the node ID of the node apparatus 100g. As a result,
the node apparatus 100c learns the association between the MAC

address of the PC 205 and the node ID of the node apparatus 100g
from the returned ad hoc frame, and adds an entry to the MAC
table 153.
[0205] Figure 10 is a diagram which illustrates an example
of pause state management data. The pause state management data
154 in Figure 10 is managed by the port monitor 137 in Figure
6, and used for controlling pause requests and notifications
of canceling the pause requests that are from the node apparatus
100 itself to its adjacent node apparatuses. The detailed
processing pertaining to the pause state management data 154
will be described later with reference to Figures 17 to 19.
[0206] As illustrated in Figure 10, the pause state management
data 154 includes a pause state SS, a counter SCount, and a pause
start time StartTime.
The value of the pause state SS is "P" (Pause) in a case
where the node apparatus 100 itself is in the busy state, and
"N" (Normal) in a case where the node apparatus 100 itself is
not in the busy state. For instance, the port monitor 137
monitors the usage percentage of the buffer unit 142. If the
usage percentage of the buffer unit 142 exceeds a prescribed
threshold, the port monitor 137 may judges that the node apparatus
100 itself is in the busy state, and may set the value of the
pause state SS to "P".
[0207] A value that is to be designated in the pause frames
when the node apparatus 100 itself transmits the pause requests
to other node apparatuses is set in the counter SCount. That
is, the length of a time period during which the node apparatus
100 itself requests its adjacent node apparatuses to pause
transmission of frames is set in the counter SCount.
[0208] The port monitor 137 may arbitrarily set the value
of the counter SCount. Depending on the embodiments, the port
monitor 137 may set a predetermined constant value in the counter
SCount, or may set a value of a parameter, which is variable
depending on the usage percentage of the buffer unit 142, in
the counter SCount.

[0209] A time at which the node apparatus 100 itself transmits
the pause requests to the adjacent node apparatuses is set as
the pause start time StartTime.
The configuration of the node apparatus 100 and various data
structures have been described as above. Hereinafter, detailed
operation of the node apparatus 100 will be described with
reference to flowcharts.
[0210] Figures 11A to 11D are flowcharts which explain a
routing process. As represented as an infinite 100p in the
flowcharts in Figures llAto 11D, the FPGA113 continues to execute
processing in Figures 11A to 11D while the power of the node
apparatus 100 is being turned on.
[0211] When the power of the node apparatus 100 is turned
on, initialization is performed in step S201. For instance,
the MPU 114 may read out the node ID of the node apparatus 100
itself from the flash memory 119, then may store it in the DDR2
SDRAM 118, and may output it to the FPGA 113. In addition, the
table controller 136 may initialize the various tables in step
S201.
[0212] For instance, since nothing is learned at the time
of step S201, the routing table 103, the broadcast management
table 104, the 100p detection table 105, and the MAC table 153
are initialized into a state with no entry. The table controller
136 may also set the pause state PSi to "N" and may set the counter
County to zero in each entry (1 ≤ i ≤ x) in the pause state table
151. Further, the table controller 136 may set the link state
Li to "C" in each entry (1 ≤ i ≤ x) in the port link state table
152.
[0213] In addition, the port monitor 137 may initialize the
pause state management data 154 in step S201. For instance,
the port monitor 137 may set the pause state SS to "N" and may
set the counter SCount to zero.
[0214] Subsequently, in step S202, the node apparatus 100
waits until a frame is received from any of the ports 101-1 to
101-x.

In step S202, if the frame is received from, for instance,
the port 101-r (1 ≤r ≤x) , the frame is output ted from the receiving
port 132-r to the received frame controller 133 as illustrated
in Figure 6. Further, the received frame controller 133
instructs the table controller 136 to perform learning of the
MAC table 153, which is described pertaining to Figure 9, and
the table controller 136 performs the learning of the MAC table
153. Then, the processing proceeds to step S203.
[0215] In step S203, the received frame controller 133 judges
whether the frame received in step S202 is a pause frame or not
on the basis of the type 413 in the received frame. If the frame
received in step S202 is the pause frame, the processing proceeds
to step S204. If the frame received in step S202 is not the
pause frame, the processing proceeds to step S207.
[0216] In step S204, the received frame controller 133 judges
whether the value of the pause counter designated in the received
pause frame is zero or not. If the value of the pause counter
is zero, the processing proceeds to step S205. If the value
of the pause counter is not zero, the processing proceeds to
step S206.
[0217] Step S205 is executed in a case where the value of
the pause counter is zero, that is, in a case where cancellation
of a pause request is notified. More specif ically, instepS205,
the received frame controller 133 instructs the table controller
136 to cancel the pause state corresponding to the port from
which the pause frame has been received in step S202 and to clear
the counter. The table controller 136 then operates according
to the instruction.
[0218] For instance, in a case where the pause frame is received
from the wired ad hoc network port 101-r (1 ≤ r ≤ x) in step
S202, the table controller 136 operates, in step S205, according
to the instruction of the received frame controller 133 as
follows.
[0219] That is, the table controller 136 sets the pause state
PSr and the counter Countr, which are associated with the port

number PNr of the wired ad hoc network port 101-r, to "N" and
zero, respectively, in the pause state table 151 in Figure 7.
Then, the processing returns to step S202.
[0220] On the other hand, step S206 is executed in a case
where the value of the pause counter is not zero, that is, in
a case where a new pause request or a pause extension request
is received.
Note that the pause extension request is a type of pause
requests. In a case where a node apparatus having previously
issued a first pause request continues to be in the busy state
even after the time period designated in a pause frame indicating
the first pause request has elapsed, this node apparatus may
issue a second pause request again instead of notifying
cancellation of the first pause request. The second pause
request is the pause extension request.
[0221] More specifically, in step S206, the received frame
controller 133 instructs the table controller 136 to set the
pause state corresponding to the port from which the pause frame
has been received in step S202 to "P" and to set the counter
according to the pause frame. The table controller 136 then
operates according to the instruction.
[0222] Forinstance, inacase where the pause frame is received
from the wired ad hoc network port 101-r (1 ≤ r ≤ x) in step
S202, the table controller 136 operates, in step S2QS, as follows
according to the instruction of the received frame controller
133.
[0223] That is, the table controller 136 sets the pause state
PSr, which is associated with the port number PNr of the wired
ad hoc network port 101-r in the pause state table 151 in Figure
7, to "P". The table controller 136 also sets the counterCountr,
which is associated with the port number PNr in the pause state
table 151 in Figure 7, to the value designated in the pause frame
received in step S202. Then, the processing returns to step
S202.
[0224] Step S207 is executed in a case where a frame whose

type is other than the pause frame is received in step S202.
In step S207, the received frame controller 133 judges whether
the frame received in step S202 is a broadcast frame or not.
[0225] More specifically, if the MAC-DA 421 is the broadcast
address (i.e., the address in which every bit is "1") , the received
frame controller 133 judges that the received frame is the
broadcast frame, and then the processing proceeds to step S208
in Figure 11B. On the other hand, if the MAC-DA 421 is not the
broadcast address, the processing proceeds to step S216 in Figure
11A.
[0226] Steps S208 to S215 illustrated in Figure 11B are
processes for a case where the broadcast frame is received in
step S202.
Instep S208, the received frame controller 133 judges whether
an entry corresponding to the frame received in step S202 exists
in the 100p detection table 105 or not. That is, the received
frame controller 133 searches, via the table controller 136,
the 100p detection table 105 using the pair of the values of
the GS 412 and the FID 414 in the frame received in step S202
as search keys.
[0227] If an entry whose source GSi is identical with the value
of the GS 412 and whose FIDi is identical with the value of the
FID 414 is found as a result of the search (1 ≤i ≤m) , the processing
proceeds to step S209. On the other hand, if the entry above
is not found as a result of the search, the processing proceeds
to step S210.
[0228] Note that no more than one entry as described above
exists, even if it exists.
Step S209 is executed in a case where it is apparent that
a broadcast frame having previously been transmitted from the
node apparatus 100 is received in the node apparatus 100 again.
Accordingly, in step S209, the received frame controller 133
discards the frame received in step S202. Then, the processing
returns to step S202. The operation of the node apparatus 100r
in step S105 and the operation of the node apparatus 100s in

step S108, as in the example in Figure 4, correspond to step
S209.
[0229] In step S210, the received frame controller 133
searches the broadcast management table 104 via the table
controller 136. According to the search, the received frame
controller 133 judges whether the frame received in step S202
is one that is received again due to broadcasting from the same
MAC address in the broadcast management table 104 within a
prescribed time period (e.g., 10 ms).
[0230] More specifically, the received frame controller 133
searches, via the table controller 136, the broadcast management
table 104 using the value of the MAC-SA 422 in the frame received
in step S202 as a search key. If an entry in which a time that
differs from the current time by the prescribed time period (10
ms in this embodiment) or less is associated with the search
key is found as a result of the search, the processing proceeds
to step S211. On the other hand, if the entry above is not found
as a result of the search, the processing proceeds to step S212.
[0231] Step S211 is executed in a case where a broadcast frame
having recently (within the time period of 10 ms in this
embodiment) been received by the node apparatus 100 is received
again by the node apparatus 100. Accordingly, in step S211,
the received frame controller 133 discards the frame received
in step S202. Then, the processing returns to step S202. The
operation of the node apparatus 100s in step SI04 and the operation
of the node apparatus 100t in step S107, as in the example in
Figure 4, correspond to step S211.
[0232] In a case where the entry is not found in the broadcast
management table 104 as a result of the search in step S210,
the received frame controller 133 judges that it is not the case
that a broadcast frame identical to one that has previously been
received is received again in step S202. Accordingly, in step
S212, an upward transfer process is performed.
[0233] That is, in step S212, the received frame controller
133 outputs the frame to the upper layer processor 134 and then

the upper layer processor 134 processes the frame. Described
in line with Figure 5, in step S212, the frame is outputted from
the FPGA 113 to the MPU 114 via the Mil 124 and the frame is
processed by the MPU 114.
[0234] Next, in step S213, the received frame controller 133
instructs the port selector 138 to perform broadcasting into
the wired ad hocnetwork, and outputs the frame to the port selector
138. Then, in step S213, the port selector 138 selects all the
ports satisfying both of the following two conditions from among
the ports 101-1 to 101-x as the destination ports.
[0235] • Not being the receiving port (i.e. , the port having
received the broadcast frame in step S202).
• The pause state in the pause state table 151 in Figure
7 is not "P" (i.e., the pause state is "N").
[0236] Instead, in some embodiments, the port selector 138
may further narrow down the destination ports according to a
condition that the link state is "C" in the port link state table
152 in Figure 8.
[0237] The port selector 138 gives the transmission frame
controller 140 an instruction on the ports selected from among
the ports 101-1 to 101-x (i.e., the transmitting ports selected
from among the transmitting ports 143-1 to 143-x), and outputs
the frame to the transmission frame controller 140. Note that
the number of the selected ports may be zero, one, or plural.
[0238] The transmission frame controller 140 generates a new
ad hoc header 410 by decrementing the value of the TTL 415 in
the received frame by one and recalculating the FCS 417, and
outputs an ad hoc frame 400 including the new ad hoc header 410
to the buffer unit 142 . Then, the transmission frame controller
140 performs control for transmitting the ad hoc frame 400, which
is buffered in the buffer unit 142, from the transmitting ports
selected from among the transmitting ports 143-1 to 143-x.
[0239] As described above, in step S213, the frame is
transmitted from the transmitting ports selected from among the
transmitting ports 143-1 to 143-x. Then, the processing

proceeds to step S214.
[0240] In step S214, the received frame controller 133
registers an entry corresponding to the broadcast frame received
in step S202 in the broadcast management table 104 via the table
controller 136. That is, the entry, in which the value of the
MAC-SA 422 in the broadcast frame received in the step S202 and
the current time (i.e., the time stamp) acquired from the timer
139 are associated with each other, is added to the broadcast
management table 104 in step S214.
[0241] Then in step S215, the port selector 138 registers
a new entry in the 100p detection tablel05via the table controller
136 according to the selection result in step S213.
[0242] The new entry, which is registered in step S215, is
as follows. For instance, assume that the ad hoc frame 400 is
received from the port 101-r (1 ≤ r ≤ x) in step S202. In this
case, a new entry associating the pair of the values of the GS
412 and the FID 414 in the received ad hoc frame 400 with the
port number of the port 101-r and the states of the respective
ports 101-1 to 101-x is added to the 100p detection table 105
instepS215. Assuming that the new entry is the (m+1) -th entry
in the 100p detection table 105, the state LPi(m+D of the port
101-i (1 ≤ i ≤ x) in the new entry is set as follows.
[0243] • In a case where the port 101-i is selected as the
destination port in step S213, the value of the state LPi(m+1)
is "U".
• Otherwise, the value of the state LPi(m+i) is "E".
[0244] In some embodiments, the order of the processes of
steps S212 to S215 may be changed, or steps S212 to S215 may
be executed in parallel.
Here, in order to further clarify themeaning of the processes
in steps S210 and S211, another specific example will further
be described with reference to Figure 2A.
[0245] For instance, there is a case where the PC 206 in Figure
2A broadcasts an Ethernet frame 420 including an ARP (Address
Resolution Protocol) packet as an L3 packet 424. In this case,

the broadcast address is designated in the MAC-DA 421 in the
Ethernet frame 420.
[0246] In the node apparatus lOOh connected to the PC 206
being the source via the link 229, the Ethernet frame 420 that
is the broadcast frame is provided with an ad hoc header 410
and thereby turns into an ad hoc frame 400. Here, the node ID
of the node apparatus lOOh is designated in the GS 412 in the
ad hoc header 410, and a first FID generatedby the node apparatus
lOOh is designated in the FID 414.
[0247] The ad hoc frame 400 being the broadcast frame is then
broadcast from the node apparatus lOOh to the wired ad hoc network
200. More specif ically, the node apparatus lOOh transmits (i.e.,
broadcasts) the ad hoc frame 4 00 from the respective ports 101h-l,
101h-3, and 101h-4.
[0248] Subsequently, the ad hoc frame 400 may reach the node
apparatus 100b via, for instance, the links 223 and 217, and
may further reach the node apparatus 100a from the node apparatus
100b via the link 216.
[0249] In this case, the node apparatus 100b executes step
S212 . As a result, the node apparatus 100b transmits the Ethernet
frame 420, which results from removal of the ad hoc header 410
from the received ad hoc frame 400, to the L2SW 202 connected
to the general LAN port 106b.
[0250] The L2SW 202, having received the Ethernet frame 420
being the broadcast frame from the node apparatus 100b, then
broadcasts the received Ethernet frame 420. That is, the L2SW
202 transmits the Ethernet frame 420 to the PC 201 via the link
211 and also to the node apparatus 100a via the link 212.
[0251] Incidentally, as described above, the node apparatus
100a has received the ad hoc frame 400 from the port 101a-4 via
the link 216 as a result of the broadcasting in the wired ad
hoc network 200. In the node apparatus 100a, there is also a
possibility that the ad hoc frame 400 is further received from
the node apparatus lOOd via the link 215.
[0252] In a case where the node apparatus 100a redundantly

receives the ad hoc frame 400 being the broadcast frame from
both the node apparatuses 100b and lOOd in the wired ad hoc network
200, the frame received later is discarded in step S209. The
discard in step S209 is enabled by the 100p detection table 105.
Step S209 prevents congestion of broadcast frames in the wired
ad hoc network 200.
[0253] On the contrary, if the broadcast management table
104 did not exist, there would be a possibility that a broadcast
frame having once passed through the L2SW 202 external to the
wired ad hoc network 200 would return again into the wired ad
hoc network 200, thus causing congestion. Accordingly, in this
embodiment, the node apparatus 100 includes the broadcast
management table 104.
[0254] If the broadcast management table 104 did not exist,
the node apparatus 100a having received the ad hoc frame 400
from the node apparatus 100b or lOOd by the broadcasting in the
wired ad hoc network 200 would remove the ad hoc header 410 and
would transmit the Ethernet frame 420 to the L2SW 202. The L2SW
202 would then transmit the Ethernet frame 420 being the broadcast
frame to both the PC 201 and the node apparatus 100b.
[0255] Here, any Ethernet frame 420 received by the node
apparatus 100b from the L2SW 202 does not include the ad hoc
header 410. Accordingly, use of the 100p detection table 105
does not enable the node apparatus 100b to recognize that the
same Ethernet frame 420 as a broadcast frame that the node
apparatus 100b itself has previously transmitted to the L2SW
202 is received again.
[0256] On the other hand, if the node apparatus 100b did
broadcast the broadcast frame received from the L2SW 202 into
the wired ad hoc network 200, the same Ethernet frame 420 would
be broadcast again. Further, there would be a possibility that
the same broadcast frame would continue to 100p among the L2SW
202, the node apparatus 100b, and the node apparatus 100a. This
is because the broadcast frame having once exited out of the
wired ad hoc network 200 and then returned into the wired ad

hoc network 200 is assigned a new FID every time it returns into
the wired ad hoc network 200.
[0257] Accordingly, in order to prevent congestion, the node
apparatus 100b is required to utilize information other than
the 100p detection table 105 and to judge identity between the
Ethernet frames 420, which include no ad hoc header 410.
[0258] In this embodiment, use of the broadcast management
table 104 enables the node apparatus 100 to judge the identity
between the Ethernet frames 420, which include no ad hoc header
410, according to the combination of the value of the MAC-SA
422 and the time. That is, the broadcast management table 104
is an example of a mechanism that enables prevention of congestion
even in a case where a broadcast frame having once passed through
the outside of the wired ad hoc network 200 returns again into
the wired ad hoc network 200.
[0259] Note that the value "10 ms" exemplified above may
appropriately be changed according to embodiments. For
instance, an appropriate value may be determined by performing
a preliminary experiment. The appropriate value varies
depending on, for instance, the frequency at which broadcast
frames are transmitted, and time taken for a frame to pass through
a device (such as the L2SW 202) external to the wired ad hoc
network 200 and to return again into the wired ad hoc network
200.
[0260] Here, returning to the description of Figure 11A,
processing in step S216 performed in a case where a frame other
than the broadcast frame is received in step S202 will be
described.
In step S216, the received frame controller 133 judges whether
the frame received in step S202 is addressed to the node apparatus
100 itself or not. That is, the received frame controller 133
judges whether or not the node ID of the node apparatus 100 itself
preliminarily read out from the flash memory 119 in step S201
and the value of the GD 411 in the ad hoc frame 400 received
in step S202 are identical with each other.

[0261] If the node ID of the node apparatus 100 itself and
the value of the GD 411 are identical with each other, the frame
received in step S202 is one addressed to the node apparatus
100 itself, and accordingly, the processing proceeds to step
S217. In contrast, if the node ID of the node apparatus 100
itself and the value of the GD 411 are not identical with each
other, the frame received in step S202 is not one addressed to
the node apparatus 100 itself, and accordingly, the processing
proceeds to step S218 in Figure lie in order to relay the frame
to another node apparatus.
[0262] In step S217, an upward transfer process is performed.
That is, in step S217, the received frame controller 133 outputs
the frame to the upper layer processor 134 and then the upper
layer processor 134 processes the frame. Described in line with
Figure 5, in step S217, the frame is outputted from the FPGA
113 to the MPU 114 and the frame is processed by the MPU 114.
In step S217, the upper layer processor 134 processes the data
included in the payload of the received frame as the L3 packet
424, and/or performs appropriate control based on the received
frame. Then, the processing returns to step S202.
[0263] Meanwhile, if it is judged in step S216 that the frame
received in step S202 is not one addressed to the node apparatus
100 itself, the processing proceeds to step S218 in Figure 11C.
[0264] In step S218, the received frame controller 133 judges
whether an entry corresponding to the frame received in step
S202 exists in the 100p detection table 105 or not. That is,
the received frame controller 133 searches, via the table
controller 136, the 100p detect ion table 105 using the pair of
the values of the GS 412 and the FID 414 in the frame received
in step S202 as search keys.
[0265] If an entry whose source GSi is identical with the value
of the GS 412 and whose FIDi is identical with the value of the
FID 414 is foundas a result of the search (1 ≤i≤m) , the processing
proceeds to step S219 in Figure 11D. On the other hand, if the
entry above is not found as a result of the search, the processing

proceeds to step S230 in Figure 11C.
[0266] Note that no more than one entry as described above
exists, even if it exists.
Steps S219 to S229 illustrated in Figure 11D are processes
in a case where a unicast frame having previously been transmitted
by the node apparatus 100 itself 100ps in the wired ad hoc network,
returns to the node apparatus 100, and is received at the port
101-r in step S202. For instance, in Figure 2B, the process
that follows the backtracking 303 in the search path 302 and
that is made in the node apparatus 100b, which receives the frame
from the node apparatus 100a, corresponds to Figure 11D.
[0267] In step S219, the port selector 138 recognizes the
port from which the frame has been received in step S202 as the
receiving port.
For instance, when the entry is found in step S218, the
received frame controller 133 may notify the port selector 138
of the address of the SRAM 116 notified by the table controller
136 as the address of the found entry. The received frame
controller 133 may then instruct the port selector 138 to perform
port selection for a case where a 100p is detected. As a result,
the port selector 138 may executes the processing in and after
step S219.
[0268] For instance, assume that the frame is received at
the port 101-r (1 ≤ r ≤ x) in step S202, that the i-th entry
in the 100p detection table 105 is matched as a search result
in step S218, and that the 100p detection table 105 is in the
format illustrated in Figure 1. In this case, in step S219,
the port selector 138 updates the port number RCVPNi in the i-th
entry to the number of the port 101-r.
[0269] Subsequently, in step S220, the port selector 138
changes the port state in the "U" state in the entry in the 100p
detection table 105 detected in step S218 to the "L" state.
[0270] For instance, assume that the i-th entry in the 100p
detection table 105 is matched in step S218 as a result of the
search. Since step S220 is a step executed pertaining to the

frame having been unicast, there is only one port state whose
value is "U" among the port states LPU to LPxi in the i-th entry.
Hereinafter, for convenience of description, assume that the
value of the port state LPt± of the port 101-t (1 ≤ t ≤ x) is
"U".
[0271] That is, in a case where the node apparatus 100 has
previously transmitted, from the port 101-t, an ad hoc frame
400 in which the source GSi is set in the GS 412 and in which
the same value as the FIDi is set in the FID 414, the value of
the port state LPti is "U" in step S220. In this case, the value
of the port state LPU is updated from "U" to "L" in step S220.
[0272] Further, instepS221, the port selector 138 searches,
via the table controller 136, the routing table 103 for an entry
corresponding to the value of the GD 411 in the frame received
in step S202. In this embodiment, the aging time period (i.e.,
valid duration of each entry) of the routing table 103 is longer
than the aging time period of the 100p detection table 105.
Consequently, it is assured that one entry is necessarily matched
in step S221 as a result of the search. Therefore, according
to step S221, it is possible for the port selector 138 to acquire
the address of the SRAM 116 corresponding to the matched entry
from the table controller 136, and to directly access the matched
entry thereafter.
[0273] The aging time period of the routing table 103 may
be a relatively long time period, for instance, such as one that
is 225 seconds long. In contrast, it is preferable to set the
aging time period of the 100p detection table 105 to a relatively
short time period, for instance, such as one that is shorter
than one second. This is because one entry in the routing table
103 corresponds to one node apparatus in the wired ad hoc network
while one entry in the 100p detection table 105 corresponds to
one ad hoc frame 4 00.
[0274] That is, it is expected that entries are added to the
100p detection table 105 far frequently than to the routing table
103. Accordingly, in order to prevent overflow due to excessive

increase in the number of entries in the 100p detection table
105, the aging time period of the 100p detection table 105 is
set to a relatively short time period.
[0275] It is preferable to determine an appropriate value
for the length of the aging time period of the 100p detection
table 105 in consideration of time taken for a 100p in the wired
ad hoc network, for instance, by performing a preliminary
examination.
[0276] Hereinafter, for convenience of description, assume
that the j-th entry in the routing table 103 is matched in step
S221. That is, assume that the value of the GD 411 in the ad
hoc frame 400 received in step S202 is identical with the value
of the destination GDj in the routing table 103.
[0277] Thus, in step S222, the port selector 138 makes the
current port link states and the current pause states reflected
in the entry in the routing table 103 detected in step S221.
That is, in step S222, the following operations are performed
for each port 101-k (1 ≤ k ≤ x) .
[0278] • In a case where the value of the pause state PS*,
which corresponds to the port 101-k, in the pause state table
151 in Figure 7 is "P", the value of the state RPkj of the port
101-k is updated to "P" in the j-th entry matched in the routing
table 103. Instead, in a case where the current value of the
state RPkj of the port 101-k is "P" in the j-th entry in the
routing table 103 and the value of the pause state PSk in the
pause state table 151 is "N", the value of the state RPkj is
updated to "E".
[0279] • In a case where the value of the link state Lk, which
corresponds to the port 101-k, in the port link state table 152
in Figure 8 is "D", the value of the state RPkj of the port 101-k
is updated to "D" in the j-th entry matched in the routing table
103. Instead, in a case where the current value of the state
RPkj of the port 101-k is "D" in the j-th entry in the routing
table 103 and the value of the link state Lk in the port link
state table 152 is "C", the value of the state RPkj is updated

to X,E".
[0280] In this embodiment, an operation to make the content
of the port link state table 152 reflected is performed after
an operation to make the content of the pause state table 151
reflected. Accordingly, the link state L* is reflected in the
routing table 103 with a higher priority than that of the pause
statePSk. Inaddition, as understood from the above description,
the pause state table 151 and the port link state table 152 are
used as a type of mask data for the routing table 103.
[0281] Further, in step S223, the port selector 138 sets the
focused-on port 101-t to the "L" state in the entry in the routing
table 103 detected in step S221. That is, with respect to the
port 101-t, which has been changed from the "U" state to the
>NL" state in the 100p detection table 105 in step S220, the state
is set to "L" also in the routing table 103 in step S223. More
specifically, the value of the state RPtj of the port 101-t is
updated to the "L" in the j-th entry in the routing table 103.
[0282] In a case where the node apparatus 100 receives the
frame having 100ped in the wired ad hoc network and returned,
the above steps S220 to S223 prevent a frame whose destination
is identical to that of the returned frame from being thereafter
transmitted from the destination port used so far. That is,
the node apparatus 100 learns the port causing the 100p, and
thereafter restrains transmission to the port causing the 100p.
As a result, useless traffic is reduced in the entire wired ad
hoc network.
[0283] Subsequently, in step S224, the port selector 138
searches the j-th entry in the routing table 103 for a port in
the "E" state. That is, the port selector 138 searches for a
port that has not yet been tried as the destination port for
relaying an ad hoc frame 400 with the designated destination
GDj and that is currently available (i.e., that is not in the
"P" or X,D" state) .
[0284] Then in step S225, the port selector 138 judges whether
a port in the "E" state has been found as a result of the search

in step S224 or not. If not found, the processing proceeds tc
step S226. If found, the processing proceeds to step S227.
[0285] In step S226, the port selector 138 selects the port
101-r, which has been recognized as the receiving port in step
S219, as the destination port. The processing then proceeds
to step S229.
[0286] Step S226 is executed in a case where it is tried to
relay the frame received from the port 101-r in step S202 but
a port from which transmission for relaying is feasible is not
found. More specifically, the "port from which transmission
is feasible" means a port that is not in the 100p state or in
the link-down state and that has not received a pause request
from the adjacent node apparatus.
[0287] That is, step S226 is executed in a case where the
node apparatus 100 is unable to forward the frame received in
step S202 to another adjacent node apparatus other than the
adjacent node apparatus connected to the receiving port.
Accordingly, in order to return the received frame, the node
apparatus 100 sets the port 101-r, from which the frame has been
received in step S202, as the destination port.
[0288] The node apparatus 100 is able to notify, through the
return, the adjacent node apparatus connected to the port 101-r
that the relay route is dead-ended at the node apparatus 100.
As a result, the adjacent node apparatus that receives the
returned frame is enabled by its 100p detection table 105 to
recognize that a frame addressed to the destination GDj will
100p even if it is transmitted to the port connected to the node
apparatus 100.
[0289] For instance, in the example in Figure 2B, assume that,
in the node apparatus 100b having detected the 100p by receiving
the frame returned from the node apparatus 100a by the
backtracking 303, the link 217 to the node apparatus lOOe is
in the link-down state. In this case, unlike the search path
302, backtracking further occurs also in the node apparatus 100b.
That is, in step S226, the node apparatus 100b selects the port

lOlb-4, which is connected to the node apparatus 100c, as the
destination port.
[0290] As a result, the node apparatus 100c receives the frame
returned from the node apparatus 100b. Accordingly, the node
apparatus 100c recognizes that the port 101c-l connected to the
node apparatus 100b is inappropriate as the destination port
of an ad hoc frame 400 in which the node ID of the node apparatus
100g is designated in the GD 411. That is, in the routing table
103 in the node apparatus 100c, the state of the port 101c-l
associated with the node ID of the node apparatus 100g is set
to "L".
[0291] The node apparatus 100c then selects the port 101c-3,
which is different from the port 101c-l, in step S225. As
described above, there may be a case where a new relay route
is found by multiple repetitions of backtracking in the search
tree 300 (n.b., in the search path 302 in Figure 2B, the new
relay route is found by only one instance of backtracking).
[0292] Here, the description is returned to continuation of
that of step S225.
In a case where a port(s) in the "E" state is/are found in
step S224 , the port selector 138 selects anyone (e.g. , the firstly
found port) of the port(s) in the "E" state found in step S224
as the destination port in step S227, which is subsequent to
step S225. The port selector 138 then sets the state of the
concerned port selected as the destination port to "U" in the
j-th entry in the routing table 103.
[0293] For instance, assuming that the port 101-g (1 ≤ g ≤
x) is selected in step S227, the value of the state RPgj of the
port 101-g is set to "U" in the j-th entry in the routing table
103 matched in step S221.
[0294] Then in step S228, the port selector 138 further sets
the port concerned to the "U" state also in the 100p detection
table 105. That is, in the i-th entry in the 100p detection
table 105 found in step S218, the value of the state LPgi of
the port 101-g selected as the destination port in step S227

is set to "U". The processing then proceeds to step S229.
[0295] In step S229, the frame is transmitted to the port
selected as the destination port in step S226 or S227.
That is, in step S229, the port selector 138 gives the
transmission frame controller 140 an instruction on the port
selected as the destination port, and outputs the frame to the
/frame controller 140. Then, the transmission frame controller
140 generates a new ad hoc header 410 by decrementing the TTL
415 in the received frame by one and recalculating the FCS 417,
and outputs an ad hoc frame 400 including the new ad hoc header
410 to the buffer unit 142. Further, the transmission frame
controller 140 performs control for transmitting the ad hoc frame
400 buffered in the buffer unit 142 from the transmitting port
143-g (i.e., the port 101-g) selected by the port selector 138.
[0296] Thus, the frame is transmitted from the port 101-g
instepS229. Subsequently, theprocessing returns to step S202 .
As described above, in a case where a 100p occurs, another
new relay route is tried according to steps S227 to S229 or
backtracking occurs according to steps S226 and S229 . Note that
the switching to the new relay route according to steps S227
to S229 is executed by the FPGA 113 in the node apparatus 100
within several micro seconds.
[0297] Even if the backtracking occurs, the relay route
switching is completed within a significantly short time period
even in the entire wired ad hoc network as long as the number
of steps of traversing the search tree 300 upward by the
backtracking is small, for instance, as in the search path 302
in Figure 2B. Thus, this embodiment realizes switching of a
relay route of a frame at a significantly high speed in a case
where a 100p occurs due to occurrence of a fault or another reason.
[0298] Here, returning to the description of Figure 11C, the
description will be made of processes in steps S230 to S239 for
a case where the entry has not been found in the search in step
S218.
In step S230, the port selector 138 searches the routing

table 103 using the value of the GD 411 in the frame received
in step S202, via the table controller 136.
[0299] Then in step S231, the port selector 138 judges whether
or not an entry associating the value of the GD 411 with the
state of each port has been found in the routing table 103 as
a result of the search in step S230. If the entry above has
not been found, the processing proceeds to step S232. If the
entry above has been found, the processing proceeds to step S233.
Note that no more than one entry as described above exists, even
if it is found.
[0300] In step S232, the port selector 138 registers a new
entry in the routing table 103 via the table controller 136.
In the new entry, the value of the destination is set to the
value of the GD 411 in the frame received in step S202, and the
state of the every port 101-1 to 101-x is set to the "E" state.
Then, the processing proceeds to step S233.
[0301] In the following description pertaining to Figure 11C,
let a "focused-on entry" be the one entry found as the result
of the search in step S230 or the entry newly registered in step
5232, and assume that the focused-on entry is the f-th entry
in the routing table 103.
[0302] Described in line with the example in Figure 1, l≤f≤n
holds in a case where the focused-on entry is found in step S230,
and f=n+l holds in a case where the focused-on entry is registered
in step S232. The port selector 138 acquires, as a result of
step S230 or S232, the address of the focused-on entry on the
SRAM 116 from the table controller 136, and accordingly, is
capable of directly accessing the focused-on entry thereafter.
[0303] In step S233, the port selector 138 makes the current
port link states and the current pause states reflected in the
focused-on entry in the routing table 103. That is, in step
5233, the following operations analogous to those in step S222
are performed for each port 101-k (1 ≤ k ≤ x) .
[0304] • In a case where the value of the pause state PSk,
which corresponds to the port 101-k, in the pause state table

151 in Figure 7 is "P", the value of the state RPk£ of the port
101-k is updated to VVP" in the f-th entry in the routing table
103. Instead, in a case where the current value of the state
RPkf of the port 101-k is "P" in the f-th entry in the routing
table 103 and the value of the pause state PSk in the pause state
table 151 is "N", the value of the state RPkf is updated to "E".
[0305] • In a case where the value of the link state Lk, which
corresponds to the port 101-k, in the port link state table 152
in Figure 8 is WD", the value of the state,RPkf of the port 101-k
is updated to >XD" in the f-th entry in the routing table 103.
Instead, in a case where the current value of the state RPkf
of the port 101-k is "D" in the f-th entry in the routing table
103 and the value of the link state Lk in the port link state
table 152 is "C", the value of the state RPkf is updated to "E".
[0306] Further, in step S234, the port selector 138 refers
to the focused-on entry in the routing table 103. The port
selector 138 then selects, as the destination port, one of the
port (s) that is/are other than the port 101-r (1 ≤ r ≤ x), from
which the frame has been received in step S202, and that is/are
in the state other than the "L", "P", and "D" states. Note that
the state other than the "L", "P", and "D" states is, in other
words, the "E" state or the "U" state.
[0307] In this embodiment, the port selector 138 selects a
port in the "U" state more preferentially than a port in the
WE" state in step S234. This is because the port in the ,VU"
state has previously and actually been used as the destination
port and is confirmed, in step 218, not to cause a 100p, and
thereby its reliability is established.
[0308] Step S234 is executed in a case where the frame received
in step S202 is a unicast frame. Accordingly, no more than one
port in the "U" state exists in the focused-on entry, even if
it exists.
[0309] In a case where the port selector 138 selects a port
in the "E" state in step S234 and a plurality of ports in the
"E" state exist in the focused-on entry, the port selector 138

may select, for instance, the firstly found port in the "E" state.
[0310] However, there is a possibility that there is no port
that satisfies a condition that it is other than the port 101-r
and its state is set to "E" or "U" in the focused-on entry.
Thus, in step S235, the port selector 138 judges whether
there is a destination port selected in step S234 or not. If
there is a destination port selected in step S234, the processing
proceeds to step S236. If the destination port is not found
in step S234, the processing proceeds to step S238.
[0311] In step S236, the port selector 138 updates the state
of the destination port in the focused-on entry in the routing
table 103 to the "U" state. For instance, in a case where the
port 101-e (1 ≤ e ≤ x) is selected as the destination port in
step S234, the value of the state RPef is set to "U" in step
S236.
[0312] Then in step S237, the port selector 138 additionally
registers a new entry to the 100p detection table 105 via the
tablecontroller 136. Assumingthat the new entry is the (m+l)-th
entry for instance, the content of each field in the new entry
is as follows.
[0313] • The source GSm+1 is set to the value of the GS 412
in the frame received in step S202.
• The FIDm+i is set to the value of the FID 414 in the frame
received in step S202.
[0314] • The receiving port number RCVPNm+1 is set to the number
(i.e., >xr") of the port 101-r, from which the frame has been
received in step S202.
• The state LPe(m+i) corresponding to the port 101-e is set
to "U", and all the states corresponding to the other ports are
set to "E".
[0315] After the new entry as described above is added to
the 100p detection table 10 5 in step S2 37, the processing proceeds
to step S239.
On the other hand, in a case where the destination port is
not acquired in step S234, the port selector 138 selects the

port 101-r, from which the frame has been received in step S202,
as the destination port in step S238. That is, step S23g is
a step analogous to step S226 in Figure 11D. The port 101-r
is selected in step S238 as a port for returning the frame in
order to provide notification that trying to relay the frame
received from the port 101-r in step S202 results in failure
to find a port from which transmission for relaying is feasible.
The processing then proceeds to step S239.
[0316] In step S239, the frame is transmitted to the port
selected as the destination port in step S234 or S238. Since
the process in step S239 is analogous to that in step S229 in
Figure 11D, the detailed description thereof is omitted. After
transmission of the frame, the processing returns to step S202.
[0317] Although illustration is omitted in Figures 11A to
11D for the simplicity of the description, the node apparatus
100 may discard the frame in the routing process on the basis
of the value of the TTL 415, in some cases. That is, in a case
where the value of the TTL 415 in the frame received in step
S202 is one, processes are performed as follows, for instance.
[0318] • The transmission frame controller 140 discards the
frame in step S213 in Figure 11B, steps S214 and S215 are omitted,
and the processing returns to step S202.
• Steps S225 to S229 in Figure 11D are omitted.
[0319] • Steps S230 to S239 in Figure 11C are omitted.
Although Figures 11A to 11D illustrate the processing
pertaining to a frame received at any of the wired ad hoc network
ports 101-1 to 101-x in Figure 5, the node apparatus 100 also
executes a process pertaining to a frame received at the general
LAN port 106. The node apparatus 100 further performs a process
pertaining to a f rame genera ted by the MPU114 in the node apparatus
100 itself.
[0320] More specifically, in a case where an Ethernet frame
420 is received at the general LAN port 106, the Ethernet frame
420 is outputted to the MPU 114 or the FPGA 113 according to
its MAC-DA 421. An Ethernet frame 420 addressed to the node

apparatus 100 itself is processed by the MPU 114.
[0321] In contrast, an Ethernet frame 420 addressed to another
node apparatus is provided with an ad hoc header 410 by the FPGA
113, and is transmitted from any of the wired ad hoc network
ports 101-1 to 101-x. In this case, as with step S237 in Figure
11C, an entry is also added to the 100p detection table 105.
In a case where a broadcast frame is received at the general
LAN port 106, the FPGA 113 may set a specific value, which
represents broadcasting from an external device, in the type
413 in the ad hoc header 410.
[0322] In a case where the MPU 114 generates an Ethernet frame
420, any of the following processes is executed on the Ethernet
frame 420 according to its MAC-DA 421.
[0323] • The Ethernet frame 420 is transmitted from the general
LAN port 105 via the internal connection port 114b, the PHY chip
120, and the L2SW unit 112.
• The Ethernet frame 420 is outputted to the FPGA 113, is
provided with the ad hoc header 410, and is transmitted from
any of the wired ad hoc network ports 101-1 to 101-x. In this
case, as with step S237 in Figure 11C, an entry is also added
to the 100p detection table 105. In a case where the MPU 114
generates a broadcast frame, the FPGA 113 may set a specific
value, which represents broadcasting from the MPU 114, in the
type 413 in the ad hoc header 410.
[0324] Next, various processes executed independently of the
processing in Figures 11A to 11D will be described with reference
to Figures 12 to 20.
Figure 12 is a flowchart which explains aging of the routing
table. The processing in Figure 12 is executed for each entry
in the routing table 103, for instance, at predetermined regular
intervals. Hereinafter, in the description of Figure 12, a
certain entry in the routing table 103 is referred to as a
"focused-on entry" for the sake of convenience. Described is
a case where the processing in Figure 12 is executed on the
focused-on entry.

[0325] In step S301, the table controller 136 calculates a
difference acquired by subtracting the timer value of the
focused-on entry from the current value of the timer 139 in the
node apparatus 100.
In this embodiment, timers corresponding to the respective
entries in the routing table 103 are implemented. For instance,
hardware timers corresponding to the respective entries may be
implemented on the FPGA 113. In some embodiments, each entry
in the routing table 103 may include a field representing an
entry creation time instead. In this case, in step S301, a
difference is calculated by subtracting the value of the field
representing the entry creation time of the focused-on entry
from the current value of the timer 139 in the node apparatus
100.
[0326] Next, in step S302, the table controller 136 judges
whether the calculated result acquired in step S301 is equal
to or longer than a predetermined aging time period of the routing
table 103 or not. If the difference acquired in step S301 is
shorter than the above aging time period, the processing in Figure
12 terminates.
[0327] On the other hand, if the difference acquired in step
S301 is equal to or longer than the above aging time period,
the table controller 136 deletes the focused-on entry from the
routing table 103 in step S303, and terminates the processing
in Figure 12.
[0328] Figure 13 is a flowchart which explains setting of
the timer of an entry in the routing table. The processing in
Figure 13 is executed in the following cases.
• In a case where an access to update the content of a certain
entry in the routing table 103 is performed (e.g., in step S233
or S236 in Figure 11C, or in step S222, S223, or S227 in Figure
11D) , the processing in Figure 13 is performed on the accessed
entry.
[0329] • In a case where a new entry is added to the routing
table 103 (e.g., in step S232 in Figure 11C), the processing

in Figure 13 is performed on the new entry.
• In a case where the routing table 103 is searched and a
matched entry is found as a result of the search (e.g. , in step
S230 in Figure 11C, or in step S221 in Figure 11D) , the processing
in Figure 13 is performed on the matched entry.
[0330] Hereinafter, generalizing the above three cases for
the sake of convenience, an entry as a target on which the
processing in Figure 13 is performed is referred to as a
"focused-on entry".
In step S401, the table controller 136 sets the timer (e.g.,
the hardware timer exemplified pertaining to Figure 12) of the
focused-on entry to the current value of the timer 139 in the
node apparatus 100. The processing in Figure 13 then terminates .
In embodiments where the routing table 103 includes the field
representing the entry creation time, the field representing
the entry creation time is set to the current value of the timer
139 in step S401.
[0331] The above processing in Figures 12 and 13 realizes
aging of each entry in the routing table 103. Although not
illustrated in the drawings, the table controller 136 and the
timer 139 cooperatively perform the aging process also on the
broadcast management table 104, through the processing analogous
to that in Figures 12 and 13.
[0332] Figure 14 is a flowchart which explains aging of the
100p detection table. The processing in Figure 14 is executed
for each entry in the 100p detection table 105, for instance,
at predetermined regular intervals. Hereinafter, in the
description of Figure 14, a certain entry in the 100p detection
table 105 is referred to as a "focused-on entry" for the sake
of convenience. Described is a case where the processing in
Figure 14 is executed on the focused-on entry.
[0333] In step S501, the table controller 136 calculates a
difference acquired by subtracting the timer value of the
focused-on entry from the current value of the timer 139 in the
node apparatus 100. In this embodiment, timers corresponding

to the respective entries are implemented also with respect to
the 100p detection table 105 as with the routing table 103. It
is a matter of course that, in some embodiments, each entry in
the 100p detection table 105 may include a field representing
an entry creation time.
[0334] Next, in step S502, the table controller 136 judges
whether the calculated result acquired in step S501 is equal
to or longer than a predetermined aging time period of the 100p
detection table 105 or not. If the difference acquired in step
S501 is shorter than the above aging time period, the processing
in Figure 14 terminates.
[0335] On the other hand, if the difference acquired in step
S501 is equal to or longer than the above aging time period,
the table controller 136 deletes the focused-on entry from the
100p detection table 105 in step S503, and terminates the
processing in Figure 14.
[0336] Figure 15 is a flowchart which explains setting of
the timer of an entry in the 100p detection table. Unlike in
the routing table 103, each entry in the 100p detection table
105 is deleted when a prescribed time period has elapsed after
its creation irrespective of being accessed or not. Accordingly,
the processing in Figure 15 is executed on a new entry when this
new entry is added to the 100p detection table 105 in step S215
in Figure 11B or in step S237 in Figure 11C. In other words,
the processing in Figure 15 is executed together with
transmission of a frame (more specifically, immediately after
the frame is transmitted in step S213 in Figure 11B or immediately
before the frame is transmitted in step S239 in Figure 11C),
by taking the opportunity of the process of transmitting the
frame.
[0337] More specifically, in step S601, the table controller
136 sets the timer (e.g., a hardware timer) of the new entry
added to the 100p detection table 105 to the current value of
the timer 139 in the node apparatus 100 . The processing in Figure
15 then terminates . In the embodiments where the 100p detection

table 105 includes the field representing the entry creation
time, the field representing the entry creation time is set to
the current value of the timer 139 in step S601.
[0338] The above processing in Figures 14 and 15 realizes
aging of each entry in the 100p detection table 105.
Figure 16 is a flowchart of a pause state canceling process.
The processing in Figure 16 is the processing for canceling the
"P" state set in step S206 in Figure 11A. The processing in
Figure 16 is executed for each port, for instance, at
predetermined regular intervals. Hereinafter, in the
description of Figure 16, a certain port 101-i (1 ≤ i ≤ x) is
referred to as a "focused-on port"; and described is a case where
the processing in Figure 16 is executed on the focused-on port
101-i.
[0339] In step S701, the table controller 136 refers to the
pause state table 151 in Figure 7, and judges whether the value
of the pause state PSi corresponding to the focused-on port 101-i
is WP" or not. If the value of the pause state PSi corresponding
to the focused-on port 101-i is not "P" but "N", the processing
in Figure 16 terminates.
[0340] On the other hand, if the value of the pause state
PSi corresponding to the focused-on port 101-i is VVP", the table
controller 136 counts down, in step S702, the counter Count;,,
in the entry corresponding to the focused-on port 101-i in the
pause state table 151.
[0341] Next, in step S703, the table controller 136 judges
whether or not the value of the counter Counti has become zero
as a result of the countdown in step S702. If the value of the
counter County has not become zero, the processing in Figure
16 terminates.
[0342] On the other hand, if the value of the counter Counti
has become zero, the table controller 136 cancels, in step S704,
the "P" state in the entry corresponding to the focused-on port
101-i in the pause state table 151. That is, the table controller
136 sets the pause state PSi of the entry corresponding to the

focused-on port 101-i to "N". The processing in Figure 16 then
terminates.
[0343] After the pause state is canceled in step S704, the
content of the pause state table 151 is made reflected in the
routing table 103 in step S233 in Figure 11C or in step S222
in Figure 11D.
[0344] As exemplified in Figure 7, the pause state table 151
may include the "counter" field. Instead, a hardware timer may
realize the countdown of each entry in the pause state table
151.
[0345] Instead, the pause state table 151 may include an
"expected returning time" field representing an expected
returning time from the "P" state in lieu of the "counter" field.
In step S206 in Figure 11A, a time acquired by adding a time
period designated in the pause frame to the time at which step
S206 is executed may be calculated by the table controller 136,
and may be recorded in the "expected returning time" field. In
addition, the processing in Figure 16 may be modified, instead
of the judgment in step S703, such that the step S704 is executed
in a case where the current time of the timer 139 is time after
the value of the expected returning time field.
[0346] Figure 17 is a flowchart which explains a pause
controlling process executed by the node apparatus on the node
apparatus itself. For instance, the processing in Figure 17
may be executed at predetermined regular intervals.
[0347] In step S801, the port monitor 137 refers to the value
of the pause state SS in the pause state management data 154
in Figure 10, and judges whether the node apparatus 100 itself
is in the pause state or not. If the value of the pause state
SS is "P", the processing proceeds to step S804. If the value
of the pause state SS is "N", the processing proceeds to step
S802.
[0348] In step S802, the port monitor 137 judges whether the
node apparatus 100 itself is in the busy state or not. For
instance, if the usage percentage of the buffer unit 142 exceeds

a prescribed threshold, the port monitor 137 judges that the
node apparatus 100 itself is in the busy state, and the processing
proceeds to step S803. In contrast, if the usage percentage
of the buffer unit 142 is less than or equal to the threshold,
the port monitor 137 judges that the node apparatus 100 itself
is not in the busy state, and terminates the processing in Figure
17.
[0349] In step S803, a pause frame transmission process for
starting apause is executed. Although the detaileddescription
will be made later with reference to Figure 18, the process in
step S803 is a process for outputting pause requests to the
adjacent node apparatuses. After the execution of step S803,
the processing in Figure 17 terminates.
[0350] Meanwhile, in a case where the value of the pause state
SS is "P" in step S801, the port monitor 137 judges, in step
S804, whether the node apparatus 100 itself is in the busy state
or not according to a method analogous to that of step S802.
If the port monitor 137 judges that the node apparatus 100 itself
is in the busy state, the processing proceeds to step S806. In
contrast, if the port monitor 137 judges that the node apparatus
100 itself is not in the busy state, the processing proceeds
to step S805.
[0351] In step S805, a pause frame transmission process for
canceling the pause is executed. Although the detailed
description will be made later with reference to Figure 19, the
process in step S805 is a process for issuing, to the adjacent
node apparatuses, notifications for canceling the pause requests
having previously been issued to the adjacent node apparatuses.
After the execution of step S805, the processing in Figure 17
terminates.
[0352] In step S806, the port monitor 137 compares the
difference between the current value of the timer 139 and the
pause start time StartTime in the pause state management data
154 with the counter SCount in the pause state management data
154.

[0353] Then in step S807, the port monitor 137 judges whether
the above-mentioned difference has already reached the value
of the counter SCount in the pause state management data 154
or not.
If the above-mentioned difference has already reached the
value of the counter SCount, it means that the node apparatus
100 itself is still unable to return from the busy state though
the time period requested by the node apparatus 100 to the adjacent
node apparatuses in the previous pause requests has already
elapsed. Accordingly, in this case, the processing proceeds
to step S808.
[0354] In contrast, if the above-mentioned difference is
smaller than the value of the counter SCount, it means that the
time period requested by the node apparatus 100 to the adjacent
node apparatuses in the previous pause requests has not elapsed
yet. Accordingly, the processing in Figure 17 terminates.
[0355] In step S808, a pause frame transmission process for
starting a pause is executed. Although the detailed description
will be made later with reference to Figure 18, the process in
step S808 is a process for outputting pause extension requests
to the adjacent node apparatuses. After the execution of step
S808, the processing in Figure 17 terminates.
[0356] Figure 18 is a flowchart of a pause starting process.
The process of Figure 18 is called from step S803 or S808 in
Figure 17.
In step S901, the port monitor 137 sets the current value
of the timer 139 (i.e., the current time at the time when step
S901 is executed) in the pause start time StartTime of the pause
state management data 154 in Figure 10.
[0357] Then in step S902, the port monitor 137 generates a
pause frame. As described above, the pause frame is a type of
the ad hoc frames with the ad hoc header 410, and includes, instead
of the Ethernet frame 420, one or more fields including at least
a pause counter subsequent to the ad hoc header 410.
[0358] In this embodiment, for instance, in the ad hoc header

410, the port monitor 137 sets the GD 411 to a special value
representing broadcasting, sets the GS 412 to the node ID of
the node apparatus 100 itself, and sets the type 413 to a value
representing the pause frame. In addition, the port monitor
137 sets the FID 414 to a newly generated FID, and sets the TTL
415 to a prescribed value. Since the pause frame is addressed
to the adjacent node apparatuses, the value set in the TTL 415
may be one.
[0359] In this embodiment, the pause frame has a format in
which fixed-length data follows the ad hoc header 410.
Accordingly, the length 416 is set to the value of the fixed
length that is predetermined. The port monitor 137 then
calculates the FCS 417 according to the values from the GD 411
to the length 416.
[0360] Then in step S903, the port monitor 137 determines
a pause counter value according to the busy situation of the
node apparatus 100 itself. For instance, the port monitor 137
may determine the pause counter value according to the usage
percentage of the buffer unit 142.
[0361] Here, the "pause counter value" is a value representing
the length of the time period to be requested by the node apparatus
100 to the adjacent node apparatuses to stop transmitting frames .
In other words, the "pause counter value" is a value representing
the length of the time period during which any node apparatus
adjacent to the node apparatus 100 maintains the pause state
(i.e., the "P" state) of its port that is connected to the node
apparatus 100. In some embodiments, the port monitor 137 may
use a predetermined constant value independent of the busy
situation of the node apparatus 100 itself, as the pause counter
value.
[0362] In step S903, the port monitor 137 further sets the
counter SCount of the pause state management data 154 to the
pause counter value determined as described above.
[0363] Then in step S904, the port monitor 137 sets the pause
counter of the pause frame to the pause counter value, which

has been determined in step S903.
Finally, in step S905, the port monitor 137 causes the port
selector 138 to select a port (s) from which transmission of the
pause frame is feasible, and outputs the pause frame to the
transmission frame controller 140. Meanwhile, the port
selector 138 notifies the transmission frame controller 140 of
the selection result. As a result, the transmission frame
controller 140 transmits the pause frame from each port selected
by the port selector 138.
[0364] In this embodiment, more specifically, the port
selector 138 refers to the port link state table 152 in Figure
8, and selects a port(s) in the connected state (i.e., the X,C"
state) as the "port (s) from which transmission of the pause frame
is feasible". In some embodiments, the port selector 138 may
select a port(s) that is/are in the "C" state in the port link
state table 152 in Figure 8 and that is/are in the normal state
(i.e., the "N" state) , which is not the pause state, in the pause
state table 151 in Figure 7.
[0365] As illustrated in Figure 6, in this embodiment, the
port monitor 137 is realized by the FPGA 113. Accordingly, the
pause frame is generated by the FPGA 113 in step S902 in Figure
18. Therefore, in comparison with a case where the MPU 114
generates the pause frame according to software, this embodiment
makes it possible to generate the pause frame in a shorter time
period.
[036S] Reduction in time period taken to generate the pause
frame leads to reduction in time period from the time when the
node apparatus 100 falls into the busy state to the time when
the states of the ports, which are connected to the node apparatus
100, are set to "P" in the adjacent node apparatuses and thereby
the relay route(s) of a frame(s) is/are switched. That is,
according to this embodiment, the time period taken to switch
the relay route in the entire wired ad hoc network is reduced
because the FPGA 113 generates the pause frame.
[0367] An event that the node apparatus 100 falls into the

busy state and issues pause requests is also regarded as
occurrence of a failure in a broad sense from a viewpoint of
the entire wired ad hoc network. However, as described above,
this embodiment allows quick switching to an alternative route.
Accordingly, the wired ad hoc network of this embodiment is a
fault-tolerant network preferably applicable to a
mission-critical system in which quick switching of the relay
route is desired when a failure occurs.
[0368] Figure 19 is a flowchart of thepause cancelingprocess .
The processing in Figure 19 is called from the step S805 in Figure
17.
In step S1001, the port monitor 137 clears the pause start
time StartTime in the pause state management data 154 in Figure
10. For instance, the port monitor 137 may set the pause start
time StartTime to a special value that is invalid as a time,
thereby clearing the pause start time StartTime.
[0369] Then in step S1002, the port monitor 137 generates
a pause frame in a manner similar to that in step S902 in Figure
18.
Subsequently, in step S1003, the port monitor 137 clears
the counter SCount in the pause state management data 154 . For
instance, the port monitor 137 may set the counter SCount to
zero, thereby clearing the counter SCount.
[0370] Then in step S1004, the port monitor 137 sets the pause
counter of the pause frame to zero. The pause frame whose pause
counter is set to zero is for issuing a notification to cancel
the pause state.
[0371] Then in step S1005, a process analogous to that in
step S905 in Figure 18 is executed.
More specifically, the port monitor 137 causes the port
selector 138 to select a port (s) from which transmission of the
pause frame is feasible, and outputs the pause frame to the
transmission frame controller 140. Meanwhile, the port
selector 138 notifies the transmission frame controller 140 of,
for instance, a port (s) in the "C" state as the selection result.

Instead, the port selector 138 may select a port(s) that is/are
in the NVC" state and that is/are also in the "N" state.
[0372] As a result, the transmission frame controller 140
transmits the pause frame from each port selected by the port
selector 138.
Figure 20 is a flowchart of the port monitoring process.
The port monitoring process is executed for each of the ports
101-1 to 101-x. The description will hereinafter be made using
an example of a case where the port monitoring process is executed
on the port 101-i (1 ≤ i ≤ x) .
[0373] In step S1101, the port monitor 137 judges whether
the port 101-i is in the connected state or not. For instance,
the port monitor 137 makes the judgment in step S1101 on the
basis of an output from the PHY chip 111-i indicating whether
a cable connected to the port 101-i is in a communicable state
or in a link-down state.
[0374] If the cable connected to the port 101-i is in the
communicable state and the port 101-i is in the state of being
effectively connected to the adjacent node apparatus via the
cable, the processing proceeds to step S1102 . On the other hand,
if the cable connected to the port 101-i is in the link-down
state, the processing proceeds to step S1103.
[0375] In step S1102, the port monitor 137 sets the link state
Li of the port 101-i to the "C" state in the port link state
table 152 in Figure 8 via the table controller 136. The
processing then returns to step S1101.
[0376] Meanwhile, in step S1103, the port monitor 137 sets
the link state Li of the port 101-i to the "D" state in the port
link state table 152 in Figure 8 via the table controller 136.
The processing then returns to step S1101.
[0377] As described above, the port monitoring process in
Figure 20 is repeatedly executed in a prescribed cycle.
Meanwhile, this embodiment having thus been described with
reference to Figures 1 to 20 is summarized as follows.
[0378] The node apparatus 100 of the above embodiment is used

as a first node apparatus in a network including a plurality
of node apparatuses including the first node apparatus and a
second node apparatus that are connected in a wired manner. A
specific example of the above network is the wired ad hoc network
200 in Figure 2A.
[0379] As illustrated in Figures 1 and 5, the first node
apparatus includes the plurality of ports 101-1 to 101-x. Each
port is a port to connect, in the wired manner, the first node
apparatus to another different node apparatus other than the
first node apparatus among the plurality of node apparatuses.
[0380] The node apparatus 100 as the first node apparatus
also includes the 100p detection table 105 as a specific example
of 100p detection information storage means for storing 100p
detection information. Hereinafter, for convenience of
description, a frame that has previously been transmitted by
the node apparatus 100 and that corresponds to the i-th (1 ≤
i ≤ m) entry is referred to as a "first frame".
[0381] The port states LPU to LPXi provides an example of
destination port distinguishing information for distinguishing,
from among the plurality of ports, a first port that is the
destination port for a case where the node apparatus 100 has
transmitted the first frame. That is, the states LPu to LPxi
of the respective ports 101-1 to 101-x, as a whole, serve as
information for distinguishing which port is the port in the
WrJ" state. More specifically, the port in the "U" state in the
i-th entry is the above-described first port.
[0382] The pair of the source GSi and the FIDj. is an example
of first identification information for uniquely identifying
the first frame. Thus, the data of the i-th entry in the 100p
detection table 105 is an example of the 100p detection
information that associates the destination port distinguishing
information and the,first identification information with each
other.
[0383] The first identification information in this
embodiment is a pair of the source GSi# which is

node-apparatus-identifying information (more specifically,the
node ID), and the FIDi being an FID, which is an example of
frame-identifying information for uniquely identifying each of
a plurality of frames transmitted by a source node apparatus
serving as the source. However, since the 100p detection table
105 is for detecting a frame that has 100ped, it is apparent
that any piece of identification information capable of uniquely
identifying the frame is usable instead of the pair of the node
ID and the FID.
[0384] Meanwhile, the routing table 103 is an example of
routing information storage means for storing routing
information. The routing information associates, with each of
the plurality of node apparatuses, state information that is
information for indicating whether frame transmission from each
of the plurality of ports 101-1 to 101-x of the node apparatus
100 as the first^apparatus is feasible or not.
[0385] For instance, focusing on the j-th (1 ≤ j ≤ n) entry
in the routing table 103, the states RPij to RPXj serving as a
specific example of the state information is associated with
a certain node apparatus identified by the node ID being the
destination GDj, among the plurality of node apparatuses . Here,
the states RPij to RPXj represent feasibilities of frame
transmission from the respective ports 101-1 to 101-x. More
specifically, the "U" and "E" states represent "transmittable",
and the "L", "D", and "P" states represent "non-transmittable".
[0386] The node apparatus 100 as the above-described first
node apparatus includes the receiving unit 131 in Figure 6 that
functions as receiving means for receiving a second frame from
the above-described second node apparatus, which is an adjacent
node apparatus. A specific example of second identification
information for uniquely identifying the second frame is a pair
of the values of the GS 412 and the FID 414 included in the second
frame.
[0387] Meanwhile, the routing engine 102 included in the node
apparatus 100 also functions as routing information updating

means for executing the following processing. In other words,
the FPGA 113 serving as the table controller 136 and the port
selector 138 is a specific example of the routing information
updating means.
[0388] The FPGA 113 as the routing information updating means
updates the state information when the second identification
information is identical with the first identification
information (more specifically, when an entry is matched in the
100p detection table 105).
[0389] That is, the state information associated by the
routing information with a destination node apparatus that is
the destination of the second frame is updated so as to indicate
that frame transmission from the first port is not feasible.
In this embodiment, more specifically, in step S223 in Figure
11D, the state corresponding to the first port is updated to
"L" in an entry that is in the routing table 103 and that
corresponds to the destination node apparatus, and thereby it
is indicated that frame transmission from the first port is not
feasible.
[0390] In addition, the node apparatus 100 includes the port
selector 138, the transmission frame controller 140, and the
transmitting unit 141 that function as transmitting means.
These elements functioning as the transmitting means execute
a process of selecting a second port from which transmission
of the second frame is feasible, according to the state
information associated by the routing information with the
destination node apparatus, and of transmitting the second frame
from the second port. In this embodiment, the second port is
selected as follows.
[0391] • In a case where the second identification information
is not identical with the first identification information (i.e.,
in a case where no 100p is detected) , a port in the "U" or "E"
state in the entry that is in the routing table 103 and that
corresponds to the destination node apparatus is selected as
the second port. As with this embodiment, in consideration with

the rate at which a route in the entire network converges, a
port in the "U" state may preferentially be selected in step
S234.
[0392] • In contrast, in a case where the second identification
information is identical with the first identification
information (i.e., in a case where a 100p is detected) , a port
in the WE" state in the entry that is matched in the routing
table 103 and that corresponds to the destination node apparatus
is selected as the second port.
[0393] Meanwhile, the port monitor 137 included in the node
apparatus 100 functions as link-down monitoring means for
monitoring whether the plurality of ports are in the link-down
state or not, and also functions as load monitoring means for
monitoring a load on the first node apparatus. The load is
measured by, for instance, the usage percentage of the buffer
unit 142.
[0394] When the receiving unit 131 as the receiving means
receives the second frame, the port selector 138 as the routing
information updating means updates the state information as
follows. That is, the port selector 138 updates the state
information associated by the routing information with the
destination node apparatus so as to indicate that frame
transmission from a port judged to be in the link-down state
is not feasible. In this embodiment, as illustrated in steps
S222 and S233, the "D" state indicates that frame transmission
is not feasible.
[0395] Meanwhile, if the load monitored by the port monitor
137 exceeds a predetermined criterion, a first pause frame is
generated by the port monitor 137 and is transmitted to the
adjacent node apparatus (es) by the transmission frame controller
14 0 and the transmitting unit 141, which serve as the transmitting
means, as illustrated in step S803.
[0396] There is also a case where the receiving unit 131 as
the receiving means receives the above-described second frame
from the above-described second node apparatus (i.e., one of

the adjacent node apparatuses) after receiving a second pause
frame from one of the adjacent node apparatuses.
[0397] In this case, the port selector 138 as the routing
information updating means updates the state information as
exemplified in steps S222 and S233. That is, the port selector
138 updates the state information associated by the routing
information with the destination node apparatus so as to indicate
that frame transmission from a third port which is connected
to a pause requesting node apparatus being the source of the
second pause frame is not feasible. In this embodiment, more
specifically, the "P" state indicates that frame transmission
is not feasible.
[0398] There is also a case where the port selector 138 as
the transmitting means judges that there is no port from which
transmission of the second frame is feasible, according to the
state information associated by the routing information with
the destination node apparatus. In this case, the second frame
is returned to the second node apparatus by the port selector
138, the transmission frame controller 140, and the transmitting
unit 141, which serve as the transmitting means.
[0399] Meanwhile, the broadcast management table 104 is a
specific example of broadcast management information storage
means. The broadcast management table 104 stores, in
association with a MAC address being first broadcast source
information for identifying the source of a first broadcast frame,
time information that indicates a time at which the first
broadcast frame is processed.
[0400] When the receiving unit 131 as the receiving means
receives a second broadcast frame, the second broadcast frame
is discarded as in step S211 if the following two conditions
hold.
[0401] • Second broadcast source information for identifying
the source of the second broadcast frame is identical with the
first broadcast source information.
• The difference between the above-described time information

and the current time is within a prescribed time period.
[0402] In this embodiment, the received frame controller 133
performs the discarding in step S211, while step S211 is a
preprocessing of the transmit ting processing in terms of relating
to judgment of whether to transmit the received frame, for
relaying it, from the node apparatus 100 itself or not. That
is, the received frame controller 133 also functions as a part
responsible for the preprocessing in the transmitting means.
[0403] Incidentally, in this embodiment, each branching step
in the processing in Figures 11A to 11D is a simple judgment
of whether the search result is a hit or a mishit, or of whether
a prescribed field is set to a prescribed value or not. If the
aging time period is 10 ms long, which is the same as the
above-mentioned "prescribed time period", step S210 is also
resolved into a simple judgment of whether the search result
is a hit or a mishit. Accordingly, a switching device (i.e.,
a logical operation circuit, more specifically, the FPGA 113)
is able to realize the branches in Figures 11A to 11D.
[0404] Due to independence of histories of various pieces
of information, the process that the port selector 138 as the
transmitting means selects the second port can be realized by
a combinational logic circuit, more specifically, by the FPGA
113, which is an example of a programmable logic device. In
general, an operation of an FPGA is programmable using a lookup
table whose input is a first bit pattern and whose output is
a second bit pattern. Accordingly, the FPGA 113 as the
above-described transmitting means is programmable using a
lookup table whose input and output are the following bit
patterns.
[0405] • The first bit pattern indicates which of a prescribed
number of states constituted by one or more prescribed
transmittable states and one or more prescribed
non-transmittable states each of the plurality of ports is in.
For instance, according to the processing logic in a case where
no 100p is detected, the "transmittable states" include the "E"

and "U" states in the routing table 103, and the
"non-transmittable states" include the VL", "D", and "P" states
in the routing table 103.
[0406] • The second bit pattern indicates port identification
information for identifying each of the plurality of ports.
Likewise, as to the processing logic in a case where a 100p
is detected, it is sufficient to adopt, as the input, a bit pattern
that indicates for each port whether it is in the transmittable
state (the "E" state) or in the non-transmittable state (the
X,L", XXU", WP", or "D" state) in the routing table 103.
[0407] Incidentally, this embodiment described with
reference to Figures 1 to 20 has various advantages, which are
thus described as follows.
A first advantage of this embodiment is that the time period
required to switch a route in case of occurrence of a fault is
short. Network faults in a broad sense include not only a
physical break in a link but also a case where a certain node
apparatus falls into the busy state and becomes substantially
incapable of receiving frames.
[0408] Comparison with techniques other than this embodiment
makes it clearer that the time period required to switch a route
in case of occurrence of a fault is short in this embodiment.
Accordingly, this embodiment will hereinafter be compared with
some techniques.
[0409] When configuring a network, redundant lines and the
like may be adopted so that the entire network system do not
go down even if a fault, such as a failure in a communication
node apparatus or a break in a link, occurs. For instance, STP
(Spanning Tree Protocol), which is a protocol for preventing
a 100p and which operates on Layer 2 (i.e., the data link layer)
of OSI (Open Systems Interconnection) , is applied to a redundant
network. OSPF (Open Shortest Path First), which is a routing
protocol operating on Layer 3 (i.e., the network layer), is also
applied to a redundant network. S-wire in Non Patent Literature
1 is also applicable to a redundant network in a mesh

from all the routers to know the current network topology. The
collected information is stored in a database in each router.
Each router refers to the database, creates an SPF (Shortest
Path First) tree by using SPF algorithm, and creates a routing
table from the SPFtree. According to OSPF, a routing 100p hardly
occurs because the routing table is created after the SPF tree
is created.
[0416] Meanwhile, in the route creation algorithm according
to S-wire (hereinafter, referred to as "S-wire algorithm") , each
node apparatus includes a table for managing a weight for each
port. The weight is based on the number of node apparatuses
(i.e., the number of hops) through which data passes until it
reaches a gateway as a goal of the frame forwarding. When
forwarding a frame, each node apparatus determines the
destination port according to the weight.
[0417] When communication fails because of a certain reason,
such as a fault in another node apparatus or a break in a line,
the node apparatus updates a weight corresponding to a port from
which a frame has been transmitted for this failed communication,
and retransmits the frame from another port. As a result, as
a whole of the network, route finding and route learning are
realizedby distributed coordination among the node apparatuses.
[0418] Here, making comparison and discussion in terms of
the time period required to switch a route in case of occurrence
of a fault, it is indicated that this embodiment is superior
to any of STP, OSPF, and S-wire.
According to STP, a bridge transitions among five states,
which are "Disabled", "Blocking", "Listening", "Learning", and
"Forwarding" states, when creating an alternative path (i.e.,
switching the path) in response to a fault. According to a
default timer setting, the maximum time period spent in the
"Blocking" state is 20 seconds long, each of the forward delays
for the "Listening" and "Learning" states is 15 seconds long.
Accordingly, in the default timer setting, a communication break
for 50 (=20+15+15) seconds at the maximum occurs.

[0419] Although the detailed description is omitted, also
in RSTP (Rapid Spanning Tree Protocol), which is an improved
STP, path switching requires about one second.
Meanwhile, according to OSPF, there may be a case where a
time period of tens of seconds is required from occurrence of
a fault to recovery because it takes a certain time period to
recalculate a routing table that a router includes therewithin.
This is because OSPF uses a complicated algorithm (i.e.,
Di j kstra' s algorithm) for calculating an optimum route and thus
the CPU and memory resources of the router are significantly
consumed.
[0420] Also in S-wire algorithm, it takes about one second
to switch a route in response to a fault because it takes a certain
time period to calculate the weights.
In a mission-critical environment, it is desirable to reduce
the time period to switch a route in response to a fault as much
as possible. In some applications of a network, even one second
may be too long. However, arithmetic calculations, such as the
route cost calculation in STP, the calculation according to
Dijkstra's algorithm in OSPF, and the weight calculation in
S-wire algorithm, apparently require a certain time period.
[0421] On the other hand, in this embodiment, no arithmetic
calculation is required to update the routing table 103 and the
100p detection table 105, as described above. This is because
what are managed by the routing table 103 and the 100p detection
table 105 are not pieces of numeric data, such as the weight
and distance, but are the states represented discretely.
[0422] The reason why practically sufficient performance is
attained although the ports are managed using the states rather
than the numeric data is that, in this embodiment, the adjacent
nodes are connected to each other in a wired manner and thereby
the communication quality is stable. Thatis, it isnotnecessary
for this embodiment to consider variation in communication
quality, while arithmetic operations for making the variation
in communication quality reflected are typically performed for

wireless communication, which significantly varies in
communication quality, because it is preferable to select an
appropriate route according to the variation in communication
quality.
[0423] As described above, in this embodiment, arithmetic
calculation is unnecessary to update the routing table 103 and
the 100p detection table 105. Accordingly, the routing engine
102 can easily be realized by a logical operation circuit (more
specifically, e.g., the FPGA 113) performing on/off switching.
Thus, the routing engine 102 in this embodiment, compared with
that realized by software control by the MPU 114, operates at
a higher speed.
[0424] Therefore, according to this embodiment, a time period
required to switch a route upon occurrence of a fault, in terms
of a time period taken to update the routing table 103 in a single
certain node apparatus 100, is only several microseconds long.
[0425] It is a matter of course that there is a case where
backtracking as exemplified in Figure 2B occurs before a route
is switched in the entire wired ad hoc network. In this case,
a time period taken to return a frame from a node apparatus to
another is not zero. However, in almost every case, the route
is switched in the entire wired ad hoc network within a time
period much shorter than one second. That is, unless the number
of occurrences of backtracking is extraordinarily large, the
route is switched within a time period much shorter than one
second even from a viewpoint of the entire wired ad hoc network.
[0426] According to this embodiment, use of the 100p detection
table 105 enables each node apparatus 100 to detect a 100p with
no arithmetic operation and in a self-contained manner and to
learn the destination port corresponding to the route where a
100p occurs . The expression "in a self-contained manner" herein
indicates that information pertaining to the network topology
is not exchanged with other node apparatuses.
[0427] Thus, even though each node apparatus 100 behaves in
a self-contained manner, the depth-first search by distributed

microseconds long) is taken for convergence in which a new route
is selected. In other words, this embodiment does not require
preliminary network design to keep the convergence time period
within an extent that is not problematic in practice. Thus,
this embodiment realizes appropriate performance and redundancy
without bothersome design and setting operations.
[0437] Further, this embodiment also has an advantage that
enables network resources to be fully utilized. This advantage
is better understood when it is compared with .STP and RSTP.
In STP and RSTP, a port connected to a redundant link is
set to the blocking state (i.e., is set as a blocking port) in
a normal situation in order to prevent a frame from endlessly
100ping in a network physically connected in a 100p configuration.
In other words, besides paths used in the normal situation, there
are backup paths that are not used in the normal situation but
are used only when a fault occurs. Accordingly, the network
resources (i.e., the backup paths) are idle in the normal
situation.
[0438] In contrast, according to this embodiment, the links
in the wired ad hoc network are not distinguished between those
for the normal situation and those for a situation in which a
fault occurs. Accordingly, it is possible to fully utilize the
network resources without leaving them idle.
[0439] The present invention is not limited to the above
embodiment, but may be modified in various manners. Some of
examples thereof will be described below.
In the above embodiment, the routing engine 102 is realized
by the FPGA 113. However, the routing engine 102 may be realized
by an ASIC (Application Specif ic Integrated Circuit) . Instead,
the routing engine 102 may be realized by the MPU 114 executing
a firmware program. The firmware program may be stored in an
arbitrary computer-readable storage medium and provided.
[0440] In the above embodiment, various data examples are
represented in a table format. The table format is exemplified
as an example of a preferable data format that enables realization

of high speed search using hardware, i.e., the CAM 115. In some
embodiments, the node apparatus 100 may hold various pieces of
data in a format other than the table format.
[0441] The table storage unit 135 in Figure 6 may be realized
by a combination of the CAM 115 and the SRAM 116 as described
above, but may also be realized only by the CAM 115 or only by
the SRAM 116. The table storage unit 135 can also be realized
by a combination of any other one or more types of storage devices.
[0442] The format of the ad hoc frame 400 in Figure 3 is an
example. The order of the fields and the length of each field
in the ad hoc header 410 are arbitrary according to embodiments.
A specific value of the type 413, a method according to which
the node apparatus 100 generates the value of the FID 414, an
initial value of the TTL 415, and an algorithm for calculating
the FCS 417 may arbitrarily be determined according to
embodiments.
[0443] In Figure 2A, for instance, the sensor 204 is
illustrated outside the node apparatus 100e. However, in some
embodiments, a sensor may be built into the node apparatus. For
instance, the node apparatus may include a built-in sensor, in
addition to pieces of hardware in Figure 5.
[0444] In lieu of the general LAN port 106 in Figure 5, the
built-in sensor may output data (i.e., data of a result sensed
by the built-in sensor) in the Ethernet frame format to the L2SW
unit 112. Instead, the built-in sensor may output the sensed
result data to theMPU 114, andthe MPU 114 may generate an Ethernet
frame 420 including the data of the result sensed by the built-in
sensor and may output the Ethernet frame 420 to the FPGA 113.
[0445] In Figure 2A, the sensor 204 is connected to the node
apparatus 100e via the general LAN port 106e. However, the
connection interface between the sensor 204 and the node
apparatus 100e may be any interface other than the general LAN
port 106e.
[0446] In the above embodiment, identity between broadcast
frames is judged using the broadcast management table 104 in

consideration of compatibility between the wired ad hoc network
and an external network. However, in a wired ad hoc network
that is not connected to an external network and that is used
in an isolated manner, each node apparatus may not include the
broadcast management table 104, and steps S210, S211, and S214
in Figure 11B may be omitted.

We Claim:
1. A first node apparatus in a network including a plurality
of node apparatuses including the first node apparatus and a
second node apparatus that are connected in a wired manner, the
first node apparatus comprising:
a plurality of ports, each of which is a port to connect,
in the wiredmanner, the first node apparatus to another different
node apparatus other than the first node apparatus among the
plurality of node apparatuses;
100p detection information storage means for storing loop
detection information that associates, with first
identification information for uniquely identifying a first
frame, destination port distinguishing information for
distinguishing, from among the plurality of ports, a first port
that is a destination port for a case where the first node apparatus
has transmitted the first frame;
routing information storage means for storing routing
information that associates, with each of the plurality of node
apparatuses, state information that is information for
indicating whether frame transmission from each of the plurality
of ports is feasible or not;
receiving means for receiving, from the second node
apparatus, a second frame that includes second identification
information for uniquely identifying the second frame;
routing information updating means for updating, when the
second identification information is identical with the first
identification information, the state information associated
by the routing information with a destination node apparatus
that is one of the plurality of node apparatuses and that is
a destination of the second frame so as to indicate that frame
transmission from the first port is not feasible; and
transmitting means for selecting a second port from which
transmission of the second frame is feasible from among the
pluralityof ports, according to the state informationassociated
by the routing information with the destination node apparatus,

and for transmitting the second frame from the second port.
2. The first node apparatus according to claim 1, further
comprising
link-down monitoring means for monitoring whether the
plurality of ports are in a link-down state or not,
wherein, when the receiving means receives the second frame,
the routing information updating means updates the state
information associated by the routing information with the
destination node apparatus so as to indicate that frame
transmission froma port judgedby the link-down monitoring means
to be in the link-down state is not feasible.
3. The first node apparatus according to claim 1 or 2, further
comprising
load monitoring means for monitoring a load on the first
node apparatus, wherein
when the load exceeds a predetermined criterion,
the load monitoring means generates a first pause
frame for requesting one or more adjacent node apparatuses
connected to the first node apparatus in the wired manner among
the plurality of node apparatuses to stop frame transmission
to the first node apparatus, and
the transmitting means transmits the first pause
frame to the one or more adjacent node apparatuses,
when the receiving means receives the second frame from
the second node apparatus after receiving, from one of the one
or more adjacent node apparatuses, a second pause frame for
requesting to stop frame transmission, the routing information
updating means updates the state information associated by the
routing information with the destination node apparatus so as
to indicate that frame transmission from a third port which is
one of the plurality of ports and which is connected to a pause
requesting node apparatus being a source of the second pause
frame is not feasible.

4. The first node apparatus according to any one of claims
1 to 3,
wherein the transmitting means returns the second frame
to the second node apparatus when judging that there is no port
from which transmission of the second frame is feasible among
the plurality of ports according to the state information
associated by the routing information with the destination node
apparatus.
5. The first node apparatus according to any one of claims
1 to 4, further comprising
broadcast management information storage means for
storing, in association with first broadcast source information
for identifying a source of a first broadcast frame received
by the receiving means, time information that indicates a time
at which the first broadcast frame is processed, wherein
when the receivingmeans receives a second broadcast frame,
the transmitting means discards the second broadcast frame if
second broadcast source information for identifying a source
of the second broadcast frame is identical with the first
broadcast source information and if a difference between the
time information and a current time is within a prescribed time
period.
6. The first node apparatus according to any one of claims
1 to 5, wherein
the first identification information includes:
first node-apparatus-identifying information for
uniquely identifying, among the plurality of node apparatuses,
a first source node apparatus that is one of the plurality of
node apparatuses and that is a source of the first frame, and
first frame-identifying information for uniquely
identifying each of a plurality of frames transmitted by the
first source node apparatus serving as the source, and

the second identification information includes:
second node-apparatus-identifying information for
uniquely identifying, among the plurality of node apparatuses,
a second source node apparatus that is one of the plurality of
node apparatuses and that is a source of the second frame, and
second frame-identifying information for uniquely
identifying each of a plurality of frames transmitted by the
second source node apparatus serving as the source.
7. The first node apparatus according to any one of claims
1 to 6, wherein
the state information is information that associates, with
each of the plurality of ports,
one of one or more prescribed transmittable states
which indicate that transmission is feasible, or
one of one or more prescribed non-transraittable
states which indicate that transmission is not feasible,
the transmitting means includes a programmable logic
device defined by a lookup table
whose input is a first bit pattern that indicates
which of a prescribed number of states constituted by the one
or more prescribed transmittable states and the one or more
prescribed non-transmittable states each of the plurality of
ports is in, and
whose output is a second bit pattern that indicates
port identification information for identifying each of the
plurality of ports, and
using the programmable logic device, the transmitting
means determines the second port from the state information.
8. A method executed by a first node apparatus in a network
including a plurality of node apparatuses including the first
node apparatus and a second node apparatus that are connected
in a wired manner, the method comprising:
transmitting a first frame from a first port among a

plurality of ports that are included in the first node apparatus
and that are for connecting, in the wired manner, the first node
apparatus to different node apparatuses other than the first
node apparatus among the plurality of node apparatuses;
storing, as 100p detection information, destination port
distinguishing information for distinguishing the first port
from among the plurality of ports, and first identification
information for uniquely identifying the first frame, in
association with each other;
receiving, from the second node apparatus, a second frame
including second identification information for uniquely
identifying the second frame;
when the second identification information is identical
with the first identification information,
referring to routing information that associates,
with each of the plurality of node apparatuses, state information
that is information for indicating whether frame transmission
from each of the plurality of ports is feasible or not, and
updating the state information associated by the
routing information with a destination node apparatus that is
one of the plurality of node apparatuses and that is a destination
of the second frame so as to indicate that frame transmission
from the first port is not feasible;
selecting a second port from which transmission of the
second frame is feasible from among the plurality of ports,
according to the state information associated by the routing
information with the destination node apparatus; and
transmitting the second frame from the second port.
9. A program that causes a computer to execute a process,
the computer controlling a first node apparatus in a network
including a plurality of node apparatuses including the first
node apparatus and a second node apparatus that are connected
in a wired manner, the process comprising:
transmitting a first frame from a first port among a

plurality of ports that are included in the first node apparatus
and that are for connecting, in the wired manner, the first node
apparatus to different node apparatuses other than the first
node apparatus among the plurality of node apparatuses;
storing, as 100p detection information in a storage
apparatus, destination port distinguishing information for
distinguishing the first port from among the plurality of ports,
and first identification information for uniquely identifying
the first frame, in association with each other;
receiving, from the second node apparatus, a second frame
including second identification information for uniquely
identifying the second frame;
when the second identification information is identical
with the first identification information,
referring to routing information that is stored in
the storage apparatus and that associates, with each of the
plurality of node apparatuses, state information that is
information for indicating whether frame transmission from each
of the plurality of ports is feasible or not, and
updating the state information associated by the
routing information with a destination node apparatus that is
one of the plurality of node apparatuses and that is a destination
of the second frame so as to indicate that frame transmission
from the first port is not feasible;
selecting a second port from which transmission of the
second frame is feasible from among the plurality of ports,
according to the state information associated by the routing
information with the destination node apparatus; and
transmitting the second frame from the second port.

ABSTRACT

A first node apparatus transmits a first frame from a first
port; stores, as 100p detection information, information for
distinguishing the first port and first identification
information for identifying the first frame, in association with
each other; and receives a second frame from a second node
apparatus. When second identification information for
identifying the second frame is identical with the first
identification information, the first node apparatus updates
port state information stored in association with a destination
node apparatus of the second frame so as to indicate that
transmission from the first port is not feasible. According
to the updated information, the first node apparatus then selects
a second port from which transmission is feasible, and transmits
the second frame.