Method and Apparatus for Transporting a Client Layer Signal over an Optical Transport Network (OTN)
The invention is based on a priority application EP 04026902.9 which is hereby incorporated by reference.
Field of the Invention
The present invention relates to the field of telecommunications and more particularly to a method and apparatus for transporting a client layer signal over an Optical Transport Network (OTN).
Background of the Invention
Current transmission networks are mainly based on the Synchronous Digital Hierarchy abbreviated as SDH, see ITU-T G.707, 12/2003. A new hierarchy, the Optical Transport Hierarchy abbreviated as OTH has been standardized in ITU-T G.709 03/2003, which is incorporated by reference herein. The purpose of OTH is to deal more economically with very large bandwidths, which are termed Optical Channel Data Units and abbreviated as ODUs. Currently defined are ODU1 (~2.7Gbit/s), ODU2 (-10.7 Gbit/s), and ODU3 (-43 Gbit/s). New efficient overlay networks can be defined, build and used by network operators based on this hierarchy. The architecture of optical transport
networks is described in ITU-T G.872 (11/2001), which is also incorporated by reference herein.
As Ethernet is more and more upcoming as the primary transport format for data signals, 1 Gbit/s Ethernet signals would be a natural client signal for the business of OTH network operators.
The currently standardized method to transport a 1 Gbit/s Ethernet signal through an Optical Transport Network includes to map the 1 Gbit/s Ethernet signal into a concatenation of SDH Virtual Containers (SDH VCs) and than to map the framed SDH transport signal into an ODU. This can be seen for instance from ITU-T G7041 (12/2003) on page 47. The existing mapping of 1 Gbit/s signals into SDH Virtual Containers (SDH VCs) is, however, expensive as it requires to operate a functionally independent SDH network. This in many cases would not be necessary for backbone network operators only dealing with large capacities.
On the other hand, a mapping of a single 1 Gbit/s Ethernet into the smallest OTH entity (-2.7 Gbit/s) would, however, encompass an enormous waste of bandwidth.
As 1 Gbit/s Ethernet signals are currently gaining much importance as a main client signal to be transported in long-haul networks, there exists a need to ensure a cost effective transport of these signals.
It is therefore an object of the present invention to provide a method and related network element, which allows a more efficient transport of client signals over an optical transport network and which is particularly suited for the transport of 1 Gbit/s Ethernet signals.
Summary of the Invention
These and other objects that appear below are achieved by the definition of a new OTH entity referred to as Optical Channel Data Unit-0.
In particular, a network element for an optical transport network is designed to handle optical transport signals structured in accordance with the Optical Transport Hierarchy. The Optical Transport Hierarchy provides at least three multiplex layers k with k=l, 2, and 3 and defines corresponding Optical Channel Data Units-k. The Optical Channel Data Units-k with k=l, 2, and 3 are of a size that four transport signals built from Optical Channel Data Units of a lower layer can be multiplexed into one transport signal built from Optical Channel Data Units of the next higher layer. Each Optical Channel Data Units-k has an overhead area and a payload area. The network element supports at least one of the multiplex layers k=l, 2, or 3 and corresponding Optical Channel Data Units-k, k=l, 2, or 3, respectively. According to the invention, the network element has at least one I/O port for processing a transport signal built from Optical Channel Data Units-0, which are of a size that two transport signals built from Optical Channel Data Units-0 can be multiplexed into one transport signal built from Optical Channel Data Units-1, and means for mapping a client layer signal into the payload area of said Optical Channel Data Units-0.
The invention therefore solves the economic burden of either to operate an additional underlying SDH network or to waste more than 50% of bandwidth in the case of mapping it into an ODU1. This doubles the capacity of 1 Gbit/s Ethernet signals that can be transported individually in an OTH signal.
Brief Description of the Drawings
A preferred embodiment of the present invention will now be described with reference to the accompanying drawings in which:
figure 1 shows the multiplexing of the new entity ODUO into ODU1,
figure 2 shows the multiplexing of ODUO into ODU2,
figure 3 shows the multiplexing of ODUO into ODU3,
figure 4 shows an ODUO and its associated transport unit OTUO with a GFP-T encapsulated 1GE client layer signal,
figure 5 shows the mapping of ODUO into an OPU1 tributary slot TS1,
figure 6 shows the mapping of ODUO into an OPU2 tributary slot TS3,
figure 7 shows the mapping of ODUO into an OPU3 tributary slot TS32, and
figure 8 shows Multiplex Structure Identifier encoding for payload type value 0x21 defined for ODUO.
Detailed Description of the Invention
The basic multiplex entity in an Optical Transport Network (OTN) as specified in ITU-T Rec. G.872 is the Optical Channel Data Unit-k (ODUk) with k=l, 2, or 3. An ODUk when represented in the form of a table composed of columns and rows has an overhead area and a payload area. The Overhead area of an ODUk is composed of 16 columns by 4 rows, where columns 1 to 7 of row 1 are reserved for frame alignment purposes, columns 8 to 14 of row 1 are reserved for OTU-specific overhead (Optical Channel Transport Unit) and columns 15 and 16 are reserved for OPU-specific overhead (OPU: Optical Channel Payload Unit), while the remainder of the overhead area is available for the ODU overhead. The payload area, on the other hand, has 3808 columns by 4 rows. Either a client layer signal can be mapped into the payload area or a number of ODUs of a lower layer which are interleaved to what is called an Optical Channel Data Tributary Unit Group (ODTUG). Transmission of an ODUk is row by row. The Optical Transport Hierarchy (OTH) that describes the multiplexing scheme in more detail is specified in ITU-T G.709 (3/2003).
An 1000BASE-X signal (specified in IEEE 802.3) with its bit rate of 1.25 Gbit/s ± 100 ppm after 8B/10B encoding can be mapped as a client layer signal into an OPU1 (Optical Payload Unit-1) via GFP-T encapsulation (see ITU-T Rec. G.7041) for transmission over the Optical Transport Network (OTN). 1000BASE-X signals are commonly referred to as 1 Gbit/s Ethernet or simply 1GE signals.
As the real bandwidth of an OPU1 is, however, approximately 2.5 Gbit/s, this mapping would waste bandwidth and granularity. The bandwidth efficiency would be only about 50%. This is not too concerning on metro WDM line systems where low cost and simple mappings are an advantage. When carried however through an OTH network with DWDM line systems, a 100% increase in bandwidth efficiency and ODUk multiplexing granularity will make such network more economical given the increasing importance of 1GE connections.
Therefore, a basic idea of the present invention is the definition of a new OTH entity referred to hereinafter as Optical Channel Data Unit-0 (ODU0) with a capacity of approximately 1,22 Gbit/s. This new entity fits perfectly into the existing OTH multiplexing structure, allowing the transport of two times a 1 Gbit/s Ethernet client layer signal within the capacity of one ODU1, while being individually switchable (ODU0 switching).
For this purpose, each of the existing OPUk Tributary Slots (TS) is split into two. Each new Tributary Slot represents a bandwidth of 238/(239-k)*1.244160 Gbit/s ±20 ppm.
This split results in the following values:

(Table Removed)

Table 1: Bandwidth of existing Optical Payload Units (OPUk)
With k > 1 the OPUl Tributary Slots are the smallest ones, each having a bandwidth of 1.244160 Gbit/s ± 20 ppm. The new ODUO must fit the smallest OPU Tributary Slot. The ODUO bit rate is therefore 1,244160 Gbit/s ± 20 ppm.
Such an ODUO can be fitted into the OPUl, OPU2 and OPU3 Tributary Slots with -1 /0/+11+2 ODU justification scheme (as specified in clause 19 of recommendation G.709). The mapping into an OPU2 and OPU3 Tributary Slot requires the addition of some additional fixed stuff columns as described below.
With the ODUO bit rate known, the OPU0 payload bit rate can be
determined with the above formula:
OPUk payload bit rate = 238/(239-k)*l ,244160 Gbit/s
with k=0: = 238/239 * 1,244160 Gbit/s
= 1,238954309 Gbit/s ± 20 ppm.
This bit rate is lower than the bit rate of the 8B/10B encoded 1000BASE-X (1GE) signal (i.e. 1.25 Gbit/s ± 100 ppm), and it is therefore not possible to fit the 1GE bit stream directly into this ODUO payload. Instead the Transparent Generic Framing Procedure (GFP-T) encapsulation technique as specified in ITU-T Rec. G.7041 is used.
Figures 1, 2 and 3 illustrate the mapping of a 1GE signal via GFP-T into an ODUO, the multiplexing of 2, 8 or 32 ODUOs into an ODU1, ODU2 or ODU3, respectively, and the allocation of OPUk Tributary Slots in OPU1, OPU2 and OPU3.
In the upper part of figure 1, an ODUO 101 is shown with its payload area 103 and its overhead area 104. The overhead area 104 has an area 105 reserved for OPU-specific overhead, an area 106 reserved for ODU-specific overhead, an area 107 reserved for alignment-specific bytes, and an area 108 for OTU-specific overhead. A GFP-T encoded lGbit/s Ethernet signal (1GE) 102 is mapped into the payload area 103.
In the middle part of figure 1 it is shown schematically how two ODUOs 101, 101' are multiplexed and mapped into the payload area 111 of an ODU1 110 . The ODU1 110 contains a payload area 113 and an overhead area 114. Similar to the ODUO, the overhead area 114 has an area 115 reserved for OPU-specific overhead, an area 116 reserved for ODU-specific overhead, an area 117 reserved for alignment-specific bytes, and an area 118 for OTU-specific overhead.
In the bottom part of figure 1, the physical assignment of bytes into the two interleaved tributary slots in the payload area is shown. Each ODUO occupies each second tributary slot (TS1, TS2) in the payload area 113. The OPU1 overhead 115 contains one column of justification control and opportunity overhead for each of the two tributary slots in a 2-frame multiframe format. This is not illustrated in figure 1.
It has to be noted that the ODUO floats in one half of the OPU1 payload area. An ODUO frame will cross multiple ODU1 frame boundaries. A complete ODUO frame, which has 15296 bytes, requires the bandwidth of one tributary slot in 15296/7616=2,008 ODU 1 frames. This is not illustrated in figure 1.
The mapping will be explained in more detail with respect to figures 5 to 7 below.
Figure 2 shows how in the same way 8 ODUO are multiplexed and mapped into ODU2. The ODU2 payload is subdivided into 8 bytewise interleaved tributary slots TS1-TS8, which are assigned to the 8 ODUOs. It has to be noted that the ODUO floats in 1/8 of the payload area of the ODU2. An ODUO will cross multiple ODU2 frame boundaries. A complete ODUO frame (15296 bytes) requires the bandwidth of one tributary slot in 15296/1904=8,0336 ODU2 frames. This is not illustrated in figure 2. The OPU2 overhead contains one column of justification control and opportunity overhead for each of the 8 tributary slots in a 8-frame multiframe format. This is also not shown in figure 2.
Figure 3 shows schematically how 32 ODUO are multiplexed and mapped into ODU3. The ODU3 payload is subdivided into 32 bytewise interleaved tributary slots TS1-TS32, which are assigned to the 32 ODUOs. Similarly to the examples before, the ODUO floats in 1/32 of the payload area of the ODU3. An ODUO will cross multiple ODU3 frame boundaries. A complete ODUO frame (15296 bytes) requires the bandwidth of one tributary slot in 15296/476=32,1345 ODU2 frames. This is not illustrated in figure 3. The OPU3 overhead contains also one column of justification control and opportunity overhead for each of the 32 tributary slots in a 32-frame multiframe format, which is also not shown in figure 3.
While figures 1 to 3 show only mapping of lower layer ODUs of the same type into higher layer ODUs, it should be clear to those skilled in the art, that also a mixed mapping would be possible. Therefore, a suitable mix of transport signals built from ODUO and ODU1 can be multiplexed into one transport signal built from ODU2. Equally, a suitable mix of transport signals built from ODUO, ODU1 and ODU2 can be multiplexed into one transport signal built from ODU3.
Moreover, figures 1 -3 show only 1GE signals as client layer signals. While this is indeed the preferred embodiment of the invention, it should be clear, however, that once the ODUO has been introduced, it might be used for a variety of client layer signals having a suitable data rate.
In order to transport the ODUO over a fiber in a OTM-0.0 signal, or over a wavelength in a multi-wavelength optical signal (OTM-nr.m (m = 0, 1, 2, 3) or OTM-n.m), an Optical Channel Transport Unit-0 (OTUO) is defined that carries the ODUO. The OTUO is an OTUk with k=0. The OTUO bit rate is: 4080/3824 * 1,244160 * 1,327451046 Gbit/s ± 20 ppm.
Figure 4 illustrates the mapping of an ODUO 101 into such OTUO. The OTUO contains an additional area 109 for forward error correction code (FEC) as well as OTUO overhead and alignment bytes, which are entered into the reserved fields 107, 108 in the overhead area 104.
ODUO into OPUk Tributary Slot Mapping
An OPUk Tributary Slot represents a bandwidth of 238/(239-k)*l ,244160 Gbit/s ± 20 ppm. The ODUO has a bandwidth of 1,244160 Gbit/s ± 20 ppm. The OPU1 Tributary Slot bandwidth equals the ODUO bandwidth, whereas the OPU2 and OPU3 Tributary Slot bandwidth is larger than the ODUO bandwidth. The ratio is the same as for the case of STM-N into ODUk, and the ODUO into OPUk Tributary Slot mapping design with fixed stuff columns known from the mapping of STM-N into OPUk can be re-used here.
Figure 5 shows an ODUO mapped into one of the two ODU1 tributary slots. In particular, an ODUO is mapped into an OPU1 Tributary Slot without any fixed stuff columns. The OPU1 Tributary Slot has 1904 columns, all carrying ODUO bytes. In addition the OPU1 overhead in columns 15 and 16 of the overhead area of the ODU1 is shown. The
overhead bytes have the same meaning and function as described in ITU-T G.709.
Figure 6 shows the mapping of an ODU0 into a tributary timeslot of an ODU2. The ODU0 is mapped into an OPU2 Tributary Slot with two fixed stuff columns, i.e., a total of 64 fixed stuff bytes per 8 OPU2 frames. The OPU2 Tributary Slot has 476 columns. A preferred way to organize the fixed stuff columns is with 158 columns of ODU0 bytes, 1 column of fixed stuff bytes, 158 columns of ODU0 bytes, 1 column of fixed stuff bytes and 158 columns of ODU0 bytes (see Figure 6).
Figure 7 shows the mapping of an ODU0 into a tributary timeslot of an ODU3. The ODU0 is mapped into an OPU3 Tributary Slot with one fixed stuff column, i.e., a total of 128 fixed stuff bytes per 32 OPU3 frames. The OPU3 Tributary Slot has 119 columns, and can preferably be organized as 59 columns of ODU0 bytes, 1 column of fixed stuff bytes and 59 columns of ODU0 bytes (see Figure 7).
ODU0 Justification Control and Opportunity
Each OPUj Tributary Slot has associated Justification Overhead (JOH) consisting of three bytes of Justification Control (JC) overhead, one byte of Negative Justification Opportunity (NJO) overhead and one byte of Positive Justification Opportunity (PJO) overhead. The JC and NJO overhead is located in the OPUj OH, as shown in Figures 5 to 7. The PJO overhead location is located in row 4, column 1 of the tributary slot.
With two ODUOs per ODUl, there is one justification opportunity per two ODUl frames (see Figure 5) for each of the ODUOs. For the case of eight ODUOs into OPU2, there is one justification opportunity per eight ODU2 frames (see Figure 6) for each of the ODUOs, and for thirty-two
ODUOs into OPU3 there is one justification opportunity per thirty-two ODU3 frames (see Figure 7) for each of the ODUOs.
The ODUO has a bit rate which is 50% of the STM-16 bit rate (STM: Synchronous Transport Module used for SDH), 12.5% of the STM-64 bit rate and 3.125% of the STM-256 bit rate. The OPU1 Tributary Slot bandwidth is 50% of the OPU1 payload bandwidth, the OPU2 tributary slot bandwidth is 12.5% of the OPU2 payload bandwidth and the OPU3 tributary slot bandwidth is 3.125% of the OPU3 payload bandwidth. Given that the ODUO to OPUk tributary slot bandwidth relation is similar to the STM-N to OPUk payload bandwidth relation, the ODUO mapping into an OPUk tributary slot will require the same justification capabilities as for the STM-N mapping into OPUk payload. This implies that a -1/0/+1 justification scheme suffices, requiring one NJO and one PJO byte. The general ODU multiplexing justification scheme supports a -1/0/+1/+2 justification scheme with one NJO byte and two PJO bytes. Multiplexing of ODUO will therefore typically not make use of the second PJO byte.
Multiplex Structure Identifier
The OTN supports the multiplexing of multiple ODUk types into the payload of an ODUj (j>k). The allocation of an ODUk onto an OPUj tributary slot is flexible and is indicated in the encoding of the OPUj Multiplex Structure Identifier (MSI) overhead field.
The existing MSI definition in ITU-T G.709 is not capable to identify the ODUO as a tributary signal. Therefore the MSI overhead definition needs to be extended and this extended multiplex structure is to be identified by an additional Payload Type (PT) value, e.g. value 0x21 (ODU multiplex structure with ODUO capability). The existing PT value of 0x20 is renamed into "ODU multiplex structure ODUk, k>=l".
The MSI definition associated with PT value 0x21 supports up to 32 OPUj Tributary Slots, in which an ODU0 occupies 1 tributary slot, an ODU1 occupies 2 tributary slots and an ODU2 occupies 8 tributary slots.
The MSI indicates the content of each tributary slot of an OPU. The generic coding for each tributary slot is shown in Figure 8. One byte is used for each tributary slot. Bits 1 and 2 indicate the ODU type transported in the tributary slot. Bits 3 to 8 indicate the tributary port of the ODU transported. This is of interest in case of flexible assignment of ODUs to tributary slots (e.g. ODU2 into OPU3). In case of fixed assignment the tributary port number corresponds to the tributary slot number.
A potential future ODU4/OPU4, that may have 128 tributary slots, will require a 7-bit tributary port field, and the encoding of ODU type and tributary port number will then require at least 9 bits.
Impact on Network Elements
The impact on network elements will be discussed below. First of all, network elements terminating a transport signal built from or containing ODUO must be capable to handle this new type of signal, i.e., to terminate and process the overhead, adapt the rate by byte justification, encapsulate as well as extract payload signals (i.e., GFP-T encoded 1GE signals) and generate in reverse direction new ODUO. This will typically be done in certain I/O ports of the network element. Thus, the I/O section of at least some network elements in the Optical Transport Network will be modified accordingly. The implementation of these modifications as such will be readily apparent to those having ordinary skill in the art having read and understood the above specification and definition of the ODUO structure and its overhead.
Moreover, network elements are required that allow the multiplexing of ODUO into higher layer ODUk (k>0) as described above. Therefore, multiplexers will be provided in certain network elements which are capable of multiplexing ODUO into higher layer ODUk. Such multiplexers will need to perform a rate adaptation as explained above as well as insertion of two or one fixed stuff columns per frame in the case of ODU2 and ODU3, respectively.
In addition, the definition of ODUO allows to switch ODUO signals independently which leads to greater flexibility in the network. This switching will typically be performed by network elements referred to as Optical Crossconnects. These network elements have switch matrices that allow to switch ODUk entities in space and time and potentially also wavelength from any to any I/O port. In order to allow switching at ODUO layer, some of these network elements will thus have a switch matrix that is adapted to switch transport signals from any to any I/O port at the granularity of the new ODUO entities. How this can be implemented using state-of-the-art ASIC technology will be readily apparent to those having ordinary skill in the art.
It should be noted, that the aforementioned modifications can also be combined in one network element. Moreover, it should be noted that an Optical Transport Network is not required to support all multiplex layers defined for the OTH. Instead, an Optical Transport Network can be designed to handle for instance only ODU1 or only ODU1 and ODU2 but no ODU3. Consequently, the network elements according to the invention will have to support at least one of the three previously existing layers k=l, 2, or 3 and in addition the new layer k=0. Obviously, future network elements will additionally support higher multiplex layers (k>3) that may be defined in the future once the need for higher bit rates will arise.

What is claimed is
1. A method of transporting a client layer signal over an Optical Transport Network, said Optical Transport Network being designed for the transport of optical transport signals structured in accordance with an Optical Transport Hierarchy, said Optical Transport Hierarchy providing at least three multiplex layers k with k=l, 2, and 3 and defining corresponding Optical Channel Data Units-k, wherein said Optical Channel Data Units-k with k=l, 2, and 3 are of a size that four transport signals built from Optical Channel Data Units of a lower layer can be multiplexed into one transport signal built from Optical Channel Data Units of the next higher layer; wherein each Optical Channel Data Unit-k comprises an overhead area and a payload area; wherein said Optical Transport Network supports at least said multiplex layer k=l and corresponding Optical Channel Data Units-!; and wherein said method comprises the steps of generating a transport signal built from Optical Channel Data Units-0, which are of a size that two transport signals built from Optical Channel Data Units-0 can be multiplexed into one transport signal built from Optical Channel Data Units-1, and mapping said client layer signal into the payload areas of said Optical Channel Data Units-0.
2. A method according to claim 1 wherein said client layer signal is a
1 Gbits/s Ethernet signal and wherein said method comprises the step
of encapsulating said 1 Gbits/s Ethernet signal according to a Transparent Generic Framing Procedure.
3. A method according to claim 1 comprising the step of multiplexing two transport signals built from Optical Channel Data Units-0 into one transport signal built from Optical Channel Data Units-1.
4. A method according to claim 1 comprising the steps of multiplexing eight transport signals built from Optical Channel Data Units-0 into one transport signal built from Optical Channel Data Units-2, and filling two of 476 columns of a tributary slots of said Optical Channel Data Units-2 with fixed stuff bytes.
5. A method according to claim 1 comprising the steps of multiplexing 32 transport signals built from Optical Channel Data Units-0 into one transport signal built from Optical Channel Data Units-3, and filling one of 1 19 columns of a tributary slot of said Optical Channel Data Units-3 with fixed stuff bytes.
6. A network element for an Optical Transport Network, said network element being designed to handle optical transport signals structured in accordance with an Optical Transport Hierarchy, said Optical Transport Hierarchy providing at least three multiplex layers k with k=l, 2, and 3 and defining corresponding Optical Channel Data Units-k, wherein said Optical Channel Data Units-k with k=l, 2, and 3 are of a size that four transport signals built from Optical Channel Data Units of a lower layer can be multiplexed into one transport signal built from Optical Channel Data Units of the next higher layer; wherein each Optical Channel Data Units-k comprises an overhead area and a payload area; wherein said network element supports at least said multiplex layer k=l and corresponding Optical Channel Data Units-Land wherein said network element comprises at least one I/O port for processing a transport signal built from Optical Channel Data Units-0, which are of a size that two transport signals built from Optical

Channel Data Units-0 can be multiplexed into one transport signal built from Optical Channel Data Units-1, and means for mapping a client layer signal into the payload areas of said Optical Channel Data Units-0.
7. A network element according to claim 6 comprising an Ethernet interface for receiving a 1 Gbits/s Ethernet signal, means for encapsulating said 1 Gbits/s Ethernet signal according to a Transparent Generic Framing Procedure and means for mapping said encapsulated 1 Gbits/s Ethernet signal into the payload areas of said Optical Channel Data Units-0.
8. A network element according to claim 6, comprising a switch matrix adapted to switch transport signals from any to any I/O port of said network element at a granularity of said Optical Channel Data Units-0.
9. A network element according to claim 6, comprising at least one multiplexer for multiplexing two or more transport signals built from Optical Channel Data Units-0 into one.transport signal built from Optical Channel Data Units of a higher layer.
10. A network element for an optical transport network substantially as hereinbefore described with reference to the accompanying drawings.