METHOD FOR WIFI CONNECTIVITY LOSS ANTICIPATION
FIELD OF THE INVENTION
The present invention relates to short Wireless Local Area Network
(WLAN) connectivity.
BACKGROUND OF THE INVENTION
WiFi (for Wireless Fidelity also referred to as IEEE 802. 11) is now fully
integrated in the communication landscape and is becoming the dominant
Wireless Local Area Networking (WLAN) standard. Developed in different
versions (i.e. 802. 11a/b/g/n) offering coverage up to hundreds of meters
with a theoretical throughput up to about 50 bps, WiFi networks are
widely deployed by means of numerous WiFi Access Points (AP) scattered
across different environments including business, public and residential
environments.
While roaming within the range of a WiFi access point, a mobile device
equipped with a WiFi interface may remain constantly connected, through
this access point, to wide-area networks (such as Internet or Intranet) for
different communication services including delay-constrained ones (such
as voice-over-IP, IPTV, or streaming).
However, as soon as the mobile device leaves the coverage of this WiFi
access point, an interruption occurs, causing the loss of the WiFi network
connection. Therefore, handover techniques for maintaining seamless
connections during mobility are proposed so that, while moving, WiFi users
do not experience substantial interruption in their ongoing communications
via WiFi access points.
Nevertheless, although different solutions have been proposed to improve
the handover process between WiFi access points, the handover triggering
is often based on radio strength measurements (such as Signal to
Interference and Noise Ratio SINR, or Received Signal Strength RSS). In
fact, when the radio strength level becomes lower than a predefined
threshold, the mobile station may either roam to another WiFi access point
or switch the current WiFi connection(s) to another network offering a
better connectivity (LTE for instance).
Such solutions lack robustness because radio strength measurements are
not precise enough and generally depend on hardware platform. Moreover,
radio strength level may decline quickly due to attenuation and fading
(notably, because of multi-path propagation) with the risk for the mobile
station to not having enough time to move current WiFi connection(s) on
another radio network without interruption thereof.
One object of the present invention is to propose a method and algorithm
to anticipate WiFi connectivity loss between a mobile station and a WiFi
infrastructure in order to make all necessary actions to save or move
existing network streams to another network that proposes a better
connectivity.
Another object of the present invention is to provide a method for intra-
WiFi access points handover for a seamless mobility.
Another object of the present invention is to provide an efficient trigger for
intra-WiFi handover.
Another object of the present invention is to propose a metric for intra-
WiFi access points handover that provides gains in term of delays and
system overall throughput.
SUMMARY OF THE INVENTION
Various embodiments are directed to addressing the effects of one or more
of the problems set forth above. The following presents a simplified
summary of embodiments in order to provide a basic understanding of
some aspects of the various embodiments. This summary is not an
exhaustive overview of these various embodiments. It is not intended to
identify key of critical elements or to delineate the scope of these various
embodiments. Its sole purpose is to present some concepts in a simplified
form as a prelude to the more detailed description that is discussed later.
Various embodiments relate to methods for anticipating the loss of
connectivity between a mobile device and a wireless short-range access
point among a plurality of such wireless access points, said methods
comprising the following steps:
collecting the identifiers of access points to which the mobile device
was successively associated without loss of connectivity and the
duration of each of these associations so that reproducing a plurality
of paths corresponding to the mobile device movements within the
coverage areas of the said access points to which the said mobile
device was successively associated without loss of connectivity,
each path comprising a root access point at which a connectivity is
originated and a dead-end access point at which the said
connectivity is lost;
- detecting the current access point to which the mobile device is
currently associated;
- identifying the paths comprising the said current access point;
computing the risks of losing the connectivity when following each of
the said identified paths from the said current access point to the
root access point or to the dead-end access point of each of the said
identified paths, the risk of losing connectivity when following a path
from a first access point to a second access point being the inverse
of the sum of the collected durations of association corresponding to
the access points comprised in the said path from the said first
access point to the said second access point.
In accordance with a broad aspect, the above methods further comprise
an identification step of at least an application in the mobile device
using the connectivity provided by the access point to which the said
mobile device is currently associated;
- a determining step of a threshold value above which, for the
identified application, the connectivity is considered lost;
- when a computed risk is greater than the determined threshold
value, a triggering step of a handover of the identified application.
In accordance with another broad aspect, the threshold value is the time
required to handover the connection used by the identified application to
another wireless short-range access point.
In accordance with another broad aspect, the threshold value is the time
required to handover the connection used by the identified application to a
long-range communication network.
Various embodiments relate to systems for anticipating the loss of
connectivity between a mobile device and a wireless short-range access
point among a plurality of such wireless access points, said system
comprising:
means for collecting the identifiers of access points to which the
mobile device was successively associated without loss of
connectivity and the duration of each of these associations so that
reproducing a plurality of paths corresponding to the mobile device
movements within the coverage areas of the said access points to
which the said mobile device was successively associated without
loss of connectivity, each path comprising a root access point at
which a connectivity is originated and a dead-end access point at
which the said connectivity is lost;
means for detecting the current access point to which the mobile
device is currently associated;
means for identifying the paths comprising the said current access
point;
means for computing the risks of losing the connectivity when
following each of the said identified paths from the said current
access point to the root access point or to the dead-end access
point of each of the said identified paths, the risk of losing
connectivity when following a path from a first access point to a
second access point being the inverse of the sum of the collected
durations of association corresponding to the access points
comprised in the said path from the said first access point to the
said second access point;
In accordance with a broad aspect, the above systems furthers include
means for identifying at least an application in the mobile device
using the connectivity provided by the access point to which the said
mobile device is currently associated;
means for determining a threshold value above which, for the
identified application, the connectivity is considered lost;
- when a computed risk is greater than the determined threshold
value, means for triggering a handover of the identified application.
In accordance with another broad aspect, the paths are obtained from a
site map.
In accordance with another broad aspect, the wireless short-range access
point is a WiFi access point.
Various embodiments further relate to a computer program product for
performing the above methods.
While the various embodiments are susceptible to various modification and
alternative forms, specific embodiments thereof have been shown by way
of example in the drawings. It should be understood, however, that the
description herein of specific embodiments is not intended to limit the
various embodiments to the particular forms disclosed.
It may of course be appreciated that in the development of any such actual
embodiments, implementation-specific decisions should be made to
achieve the developer's specific goal, such as compliance with systemrelated
and business-related constraints. It will be appreciated that such a
development effort might be time consuming but may nevertheless be a
routine understanding for those or ordinary skill in the art having the
benefit of this disclosure.
BRIEF DESCRIPTION OF THE DRAWINGS
The objects, advantages and other features of the present invention will
become more apparent from the following disclosure and claims. The
following non-restrictive description of preferred embodiments is given for
the purpose of exemplification only with reference to the accompanying
drawing in which
FIG. 1 is a schematic diagram illustrating an environment for
deploying various embodiments;
FiG.2 is a schematic diagram illustrating functional components for
handover triggering according to various embodiment;
- FIG. 3 is a schematic diagram illustrating an implementation of
various embodiments in a representative environment.
DETAI LED DESCRI PTION OF ILLUSTRATIVE EMBODIM ENTS
With reference to figure 1, there is shown a plurality of WiFi access points
A 1 -A1 0 deployed in an environment 10 . These WiFi access points A 1-A1 0
may be any access nodes for residential or business local area and/or
public or commercial hotspots connectivity.
As illustrative examples of the environment 1 0 , one can mention a
company, a university, a hotel, an airport, a shopping mall, a downtown
location, a pedestrian street, a cultural center, a district, or more generally
any area or building that may be covered, at least partially, by a plurality
of WiFi access points A1-A1 0 .
A user 1 provided with a WiFi capable mobile device 2 may benefit, in the
environment 10 , from different communication services including both
disconnection-tolerant and disconnection-sensitive applications such as
video streaming, Instant messaging, Web browsing, ftp, messaging (Mail),
Mo P (Multimedia over IP such as voice, data and video),
uploading/downloading files.
By "mobile device" is meant here any user equipment provided with a WiFi
interface and a WiFi connectivity manager permitting to automatically
connect this user equipment, through the WiFi interface, to a WiFi access
point A1-A1 0 . A laptop computer, a smartphone, a mobile telephone, or a
Personal Digital Assistant (PDA) are examples of such mobile device 2.
When the user 1 moves in the environment 10 while maintaining at least an
active connection through the WiFi access points A 1-A1 0 , this connection
is handed off between the WiFi access points A 1-A1 0 that the user 1
moves through their respective coverage areas. It results in that, while
moving in the environment 10 , a "path" corresponding to the user
movement within the coverage area of the WiFi access points A1-A1 0 is
built of nodes representing the WiFi access points to which the user has
been associated (i.e. has connected thereto). These paths cross each over
at cross-road nodes.
For example, if the user 1 moves successively close to the WiFi access
points A10 , A9, A7, A6 and through which he/she had, respectively,
maintained a Web browsing service (or simply having an established WiFi
connectivity without necessarily being used by an application), then a path
A 10-A9-A7-A6 corresponding to the user movement in the environment 10
is built. These access points A 10 , A9, A7, A6 may be manually of
automatically selected for the connection.
The term "move" is to be understood broadly to include any change,
whatever continuous or discrete in time, in location of the user 1 in the
environment 10 such as "walking from the railway station to the office,
working in the office, going for lunch in a restaurant close to the office,
take the elevator to the third floor, going back to the railway station".
It is to be noted that, some of the WiFi access points A 1-A1 0 may be
situated next to a coverage hole or at the border of the environment 10
(such as next to the entrance/exit of the environment 10) so that they
correspond to "dead-end" access points in paths that a user can follow
within the environment 10 . Whereas, other access points A 1-A1 0
correspond to "safe places" (from a connectivity point of view) where it has
been detected that the user 1 stays long times with no risk of WiFi
connectivity loss (such as, at home or in office).
Accordingly, the identifiers of access points A1-A1 0 to which the mobile
device 2 was successively associated without loss of connectivity and the
duration of each of these associations are collected so that reproducing a
plurality of paths corresponding to the mobile device 2 movements within
the coverage areas of the access points A1-A1 0 to which the mobile
device 2 was successively associated without loss of connectivity, each
path comprising a root access point at which a connectivity is originated
and a dead-end access point at which this connectivity is lost. Thus, paths
followed by the user 1 within the environment 10 and the durations of
association of the mobile device 2 to each of the nodes (i.e. access points)
of these paths are collected and registered.
The topology of the previously followed paths (i.e. collected paths) by the
user 1 in the environment 10 may be
- either automatically and progressively built and stored by the iFi
connectivity manager that detects WiFi access point A1-A1 0
connection;
- downloaded from a server; or even
- configured manually by the user.
Advantageously, if the paths topology is automatically built by the mobile
device 2 , these paths are consistent with the mobile device radio
performance and to the user movements.
Further, while the user is moving in the environment 10 , the WiFi
connectivity manager is configured to continually or intermittently
examines the paths followed by the user 1 and the type of ongoing network
streams used by the applications running on the mobile device 2 so as to
decide whether to trigger a handover procedure, toward another WiFi
access point A 1-A1 0 (or, alternatively, toward another radio access
networks such as 3G or 4G wireless networks), before starting the loss of
at least an ongoing network stream.
Then, at each node of a current path (in other words, at each WiFi access
point A1-A1 0 to which the mobile device 2 is currently associated, this
WiFi access point may be the first, an intermediate, or the last node of an
ongoing path), the WiFi connectivity manager is configured to update the
paths if a new WiFi access point A 1-A1 0 is detected or to estimates the
path that has the best probability to be followed.
To that end, the current access point to which the mobile device 2 is
currently associated is detected and the paths that comprise the current
access point are identified from the collected ones. Then, the risks of
losing the connectivity when following each of the identified paths from the
current access point to the root access point or to the dead-end access
point of each of the identified paths, are computed. The risk of losing
connectivity when following a path from a first access point to a second
access point is the inverse of the sum of the collected durations of
association corresponding to the access points comprised in this path from
the first access point to the second access point.
In other words, at each node of the current path being followed by the user
1 while moving in the environment 10 , the WiFi connectivity manager
calculates a risk of the current connectivity to be lost. This risk depends
on the delay necessary to go from the current WiFi access point to which
the mobile device 2 is associated to the root WiFi access point or the
dead-end WiFi access points of each collected path comprising the current
WiFi access node (i.e. the last and first WiFi access points in previously
followed path by the user 1 where a connectivity is originated or a
connectivity loss is occurred). This delay, given by the sum of association
durations, may be calculated based on the distance between the WiFi
access points (i.e. the nodes) in the path divided by the user moving
speed which could be evaluated with RSSI variation . Moreover, this delay
may be calculated using means of delays measured during previous paths.
For example, the delay necessary to go from the access point A 1 to the
access point A2 may be the arithmetic mean of
the delay to go from the access point A 1 to the access point A2; and
the delay to go from the access point A 2 to the access point A 1 (i.e.
the reverse path).
The risk is a dynamic value which estimates the probability to lose an
established WiFi connection between the mobile device 2 and a WiFi
access point A 1 -A1 0 . This risk is calculated for each current path and
depends on the current user location and speed. The value is dynamic and
is calculated at each new WiFi access point association.
In one embodiment, the computed instantaneous risk is weighted by
predefined external parameters which may depend on the hour in the day,
the day in the week (for example, public holiday or not).
After each calculation, the risk of each path, from the current user location
and speed, is compared to a threshold value. If the risk becomes greater
or equal to this threshold, a handover of ongoing network stream(s) is to
be triggered. This threshold is dynamic and depends on the applications
currently utilizing the WiFi connection. This threshold is calculated based
on the time necessary to handle properly these applications in case of
connectivity loss and the influence of connectivity loss on these
applications (i.e. whether WiFi disconnection-tolerant application or not).
In one embodiment, if no real time communication application is running,
the threshold is infinite. This is the case, for instance, of browsing
applications. Whereas, if the applications which are running are real time
applications such as voice or video communication, this threshold may
correspond to the time necessary to transfer the session over another
radio connection if available (such as a data cellular network) or to end
properly the sessions (redirection over a voice/video messaging system for
instance).
If the applications which are running are not deeply impacted by the
temporary network connectivity loss, the threshold could be estimated as
the delay necessary to re-establish the session plus the delay during
which the user accepts that the service is unavailable. For instance, the
threshold for "Instant Messaging" application could be a tenth of seconds.
When the calculated risk becomes greater than a threshold, the WiFi
connectivity manager anticipates WiFi connectivity loss by taking adequate
decisions. These decisions could be:
- warn the user;
trigger an handover procedure: switch current media-streams on
another and safer radio connectivity so that they will not be lost;
- stop, postpone or hibernate applications which could be damaged by
this connectivity loss.
Accordingly, a current value of the threshold is estimated by detecting the
active applications and their ongoing network streams to which the
corresponding current threshold value is to be applied. For each active
application or for each ongoing network stream (an application may have
more than one network stream such as an Instant Messaging application:
text, audio and video), the threshold corresponds to the inverse of the
handover delay which would impair the application. For example, if the
handover of Web browsing service between a WiFi access point and a
cellular network lasts less than 5 seconds (D_browsing_Handover = 5),
then the threshold would be T_browsing = 1/5. More generally, one can
have:
D_MMol P_Handover = 1 second, then the threshold for Multimedia
over IP applications T_MMolP is equal to 1;
- D_video_Streaming_Handover = 2 seconds, then the threshold for
video streaming application T_video_Streaming is equal to 0,5. This
value may depend on the buffer length;
- D_chat_Handover = 5 seconds, then the threshold for chat
applications T_chat is equal to 0,2;
- D_mail_Handover = 10 seconds, then the threshold for mail
applications T_mail is equal to 0, 1.
In one embodiment, the application thresholds are computed based on
default values and/or on measurements made during real handovers. The
default values correspond to general user experience.
Figure 2 depicts an illustrative embodiment of the WiFi connectivity
manager 30. In fact, the WiFi connectivity manager 30 is configured to
anticipate the WiFi connectivity loss and subsequently trigger the
handover towa rd anothe r WiFi access point or anothe r communicati on
netwo rk. To that end , the WiFi connect ivity manager 30 compri ses
a threshold estimato r 22 in charge of est imati ng a threshold above
which a ha ndover should be tri gge red , thres hold calculat ion be ing
based on current act ive st re ams. The detecti on of thes e act ive
stre ams may be done at netwo rk leve l identify ing act ive socket , or at
the mob ile devi ce leve l by means of an API used by the applicati ons
to indicate act ive strea ms. Altern ativel y, the thre shold estimato r
uses the mobile devi ce OS framework fac ilit y to detect acti ve
stream s, or uses QoS marke r in st re ams to ident ify their re al-t ime
depende ncy and accord ingly dete rmines corres ponding thres hold
values;
- a r isk calcu lato r 24 in charge to calculate in re al time the risk of
losing the WiFi con nect ivity ;
- a handover dec isio n make r 23 configured to com pare the ca lculate d
r isk to the curre ntl y estimated thresh old, and subseq uent ly take the
adequate decis ion. Once the computed r isk is greate r than the
estimated threshold, the handove r dec ision make r 23 trig ger the
handove r management module 28 for exe cut ing the handover towa rd
a WiFi acce ss point via the WiFi layer 26 or towa rd a cellular mobi le
communicati on network via the cellular PHY/MAC laye r 27 ;
- a path detector 25 in charg e of collect ing and detect ing , through the
WiFi layers 26, the paths that the user is following .
Of cou rse , the above described WiFi con nect ivity manager 30 may include
othe rs int erfa ces with diffe rent modules of the mobile dev ice .
In one embodiment , the pat h dete cto r 25 is fed with paths prov ided by the
use r and/o r down loaded from site maps .
In one embodiment , the threshold estimator 22 embed s an API (Applicat ion
prog ramming Interfa ce) to detect ongoing netwo rk st reams (such as an
uploading/d own loading , a web browsing, or a video stre aming) via the WiFi
connect ivity. Then, a threshold value per act ive application or more
genera lly per ongoing network stre am 211-213 (for example , MolP, Chat ,
video streaming, browsing, mail, social networks). It is to be noted that the
threshold estimator 22 may further offer the user the ability to specify the
threshold values.
With reference now to figure 3 , an illustrative embodiment of the above
disclosure is presented for teaching how the risk and threshold could be
calculated and how disconnection-sensitive applications can be handled.
In the example of figure 3, the environment is a building 30 comprising a
plurality of WiFi access point A31-A44.
A user enters the building 30 and walks along the path P 1 (the dashed
line) to his office 3 1 where he starts working: his mobile device is
automatically associated to A41 as soon as it is under the radio coverage
of A41 . The WiFi connectivity manager records a new path P 1 starting with
root access point A41 (changing the mobile device from a disconnected
state to a connected state). This WiFi access point A41 is identified in the
path P1 with its BSSID (Basic Service Set ID). Subsequently, a timer is
armed as soon as the WiFi connectivity is established for measuring the
duration of this association. The user walks to his office 3 1. The mobile
device maintains a WiFi connection, successively, through the access
points A31 , A32 and A33 according to handover known process. When the
smartphone looses A41 association to establish a better connectivity
through A31 , A31 is appended to the path initialized above with A41
association. The timer is stopped and the value D1_31 corresponding to
the delay necessary to reach A31 is stored within the path (i.e. the
association duration to A41 ) . The WiFi connectivity manager knows now
that it takes around D1_31 to go from A41 to A31 when taking path P 1 .
When the user is in his office 3 1 , the path P 1 is made of A41 -A31 -A32-
A33 with delays [D1 _31 , D1_32, D1_33].
After having checked his e-mail during a certain time period (for example,
1 hour or two), the user goes from his office 3 1 to a meeting room 32. This
creates the path P2 (the dotted line) made of access points A33-A34-A35
with delays [D2_34,D2_35]. As the time period of connection to the WiFi
access point A33 is quite long (i.e. several hours) , A33 is considered as a
"safe place". The reverse path P2r (namely, from the meeting room 32 to
the office 3 1) is automatically created. P2r is A35-A34-A33 with
corresponding delays [D2_35, D2_34].
The path P2 (respectively, P2r) may be the continuity of the path P1
(respectively, P2) resulting in a single path, or a new path originated from
the "safe place" covered by A33 (respectively, A35).
For instance, after a 2 hours meeting, the user comes back to his office
3 1 . This does not create a new path as the access points which ensure
WiFi connectivity between the meeting room 32 and the office 3 1 are the
same A35-A34-A33. The delay between A34 and A33 is used to update
D2_34 as this new value corresponds really to the time necessary to walk
from A34 to A33 whereas the first value corresponded to the time elapsed
between the moment at which the user arrived in his office 3 1 and the
moment at which he left his office 3 1 . For the same reason, the delay
D2_35 measured at step2 is updated with the new value which
corresponds roughly to the meeting duration. But this long value indicates
that A35 may be considered as a "safe place" .
The user leaves the building 30, to go for lunch, through the exit under the
WiFi access point A38 radio coverage (the dashed-and-dotted line). This
creates a new path P3 A33-A34-A35-A36-A37-A38 with the corresponding
delays [D3_34, D3_35, D3_36, D3_37, D3_38, D3_38loss]. The new delay
D3_35 measured between A34 and A35 (i.e. the association duration to
A34) is compared with D2_35 and both paths are updated with mean delay
if the new value is neither significantly lower nor greater than the previous
one. D3_34 is the delay measured in path P2, namely D2_34. The last
delay D3_38loss is the delay measured between the time at which the
mobile device associated to the dead-end access point A38 and the time at
which the WiFi connection is lost. The reverse path P3r is automatically
created: P3r=A38-A37-A36-A35-A34-A33 with delays
[D3_38, D3_37, D3_36, D3_35, D3_34]. When the WiFi connectivity with A38
is lost, this access point is considered as a dead-end access point. The
time of the day at which this connectivity is lost (namely, the duration of
the association to A38) is registered by the iFi connectivity manager.
This may be used to weight the risk based on user habits.
The user comes back from lunch. He enters the building 30 through the
same door and walks along path P3 to his office 3 1. Delays between A38
and A37, A37 and A36, A36 and A35, A35 and A34, A34 and A33 are
updated according to the same algorithm as above: If the new value is
significantly greater than the previous one, it is not taken into account. If
the new value is significantly lower than the previous one, it replaces it.
Otherwise, the new value is the mean of the previous one and the current
one.
As illustrative example of the risk computation, the user is in
communication with a client but has to go to the railway station to catch a
train. At the very beginning of this use case, the user is sat in his office
3 1 . The call is a video session and has been established in SIP over the
WiFi connection offered by the WiFi access point A33.
At that time there are 3 paths P1 r (reverse path of P1 ) , P2, and P3 that the
user can follow.
In one embodiment, the risk for each of these paths is calculated as
follows:
- the risk R1r when following P1 r is given by 1/(D1_41 loss + D1_33 +
D1_32 + D1_31 ) . D1_41 loss is null in this case as it is not yet
experienced connectivity loss when connected to A41 . Here, D1_31
includes D1_41 loss;
- the risk R2 when following P2 is null as there is no dead-end access
point at the end of P2. Then R2 is equal to 0 ;
- the risk R3 when following P3 is given by 1/ { D3_38loss + D3_37 +
D3_36 + D3_35 + D3_34).
These three values are compared to a threshold which is calculated based
on current running connection-based application 2 1 1-21 3 . Accordingly,
these thresholds take into account the characteristics of the ongoing
steams. In fact, the QoS required for each stream has an impact on
handover threshold. Real-time streams require early and quick handover to
avoid data loss while other streams such as browsing are disconnectiontolerant.
For instance, the threshold T may be approximate to
1/D_video_Handover where D_video_Handover is the delay necessary to
switch the current video session from the WiFi network to the cellular data
network (D_video_Handover is about 1 second for instance).
In one embodiment, initial values or default values of these thresholds per
application type are chosen so that all risks are below these initial
threshold values.
At first step, when the user moves toward the building exit and the mobile
device is associated to the WiFi access point A32, an instantaneous path
Pi=A3-A2 and an associated delay Di_32 representing the speed at which
the user is moving are created. As this instantaneous path Pi is only
comprised in P 1r among all history paths (i.e. P 1 , P 1r, P2, P2r, P3, P3r),
then this path P1 r is the most probable to be followed.
At second step, the new risk is now R1r=1 /(D1_41 loss + D1_32 + D1_31 )
and is compared to the threshold T relative to the ongoing network stream
(for example, the stream of a video). T has not changed as the video call
is still running on the mobile device. The measured Di_31 may be
compared with D1_32 to estimate if the user is moving faster or slower
than the mean values previously calculated, improving by the way the
accuracy of the risk R1r .
At third step, the mobile device is associated to WiFi access point A31 .
According to path P 1 topology, there is only one access point left before
the WiFi connectivity is lost. The risk, when associated to the access point
A31 , is R1r = 1/(D1_41 loss + D1_31 ) which grows rapidly. Consequently,
when the mobile device is associated to WiFi access point A31 , the WiFi
connectivity manager starts a timer which measures the delay Di_a since
this association. The risk is now calculated according to the foilowing
formula: R1 r= 1/(D1 _41 loss + (D1_31 -Di_a)).
At forth step, when the current risk becomes greater or equal to the
threshold T, the WiFi connectivity manager triggers a handover of the
disconnection-sensitive application 2 11-21 3 (such as a video call) over a
cellular data network. This may happen either while the terminal is still
connected through the WiFi access point A31 or when connected to the
dead-end WiFi access point A41 (when associated to A41 , R1 r is first
calculated using formula of the third step and a timer is started at A41
association to measure D1_41 loss for P1 r calculation update).
Alternatively, the risk levels may be configured by the user, depending on
locations within the environment.
In one embodiment, the information related to the mobile device
movements within the environment 10 is correlated to location based and
"movement-based" applications and not only handover. A "movement
based" application is an application which does not only rely on the user
location but also on the direction is walking in and the speed he uses for
that.
Advantageously, the above-described method and system enable secured
determination of triggering point for handover, preventing any multimedia
communication loss (i.e. triggering handover of multimedia streams before
losing these streams) and ensuring multimedia sessions continuity (being
of particular interest to multimedia application providers). Further, the
proposed system ensures accurate handover decisions as it takes into
account the characteristics of ongoing applications (disconnectionssensitive
or -tolerant).
It is to be noted that the above described method and system may be applied for any other wireless LAN standard, in the same way as for the WiFi.

CLAIMS
1. A method for anticipating the loss of connectivity between a
mobile device and a wireless short-range access point (A31 -A44) among a
plurality of such wireless access points (A31 -A44), said method comprising
the following steps:
collecting the identifiers of access points (A31 -A44) to which the
mobile device was successively associated without loss of
connectivity and the duration of each of these associations so that
reproducing a plurality of paths corresponding to the mobile device
movements within the coverage areas of the said access points to
which the said mobile device was successively associated without
loss of connectivity, each path comprising a root access point at
which a connectivity is originated and a dead-end access point at
which the said connectivity is lost;
- detecting the current access point to which the mobile device is
currently associated;
identifying the paths comprising the said current access point;
- computing the risks of losing the connectivity when following each of
the said identified paths from the said current access point to the
root access point or to the dead-end access point of each of the said
identified paths, the risk of losing connectivity when following a path
from a first access point to a second access point being the inverse
of the sum of the collected durations of association corresponding to
the access points comprised in the said path from the said first
access point to the said second access point.
2. The method of claim 1 wherein it further comprises
a step of identification of at least an application in the mobile device
using the connectivity provided by the access point to which the said
mobile device is currently associated;
- a step of determining of a threshold value above which, for the
identified application, the connectivity is considered lost;
- when a computed risk is greater than the determined threshold
value, a step of triggering of a handover of the identified application.
3. The method of claim 2, wherein the threshold value is the time
required to handover the connection used by the identified application to
another wireless short-range access point.
4. The method of claim 2 , wherein the threshold value is the time
required to handover the connection used by the identified application to a
long-range communication network.
5 . The method of any of claim 1 to 4, wherein the computed risk is
weighted by predefined parameters.
6 . A system for anticipating the loss of connectivity between a
mobile device and a wireless short-range access point (A31 -A44) among a
plurality of such wireless access points (A31 -A44) , said system
comprising:
- means for collecting the identifiers of access points (A31 -A44) to
which the mobile device was successively associated without loss of
connectivity and the duration of each of these associations so that
reproducing a plurality of paths corresponding to the mobile device
movements within the coverage areas of the said access points to
which the said mobile device was successively associated without
loss of connectivity, each path comprising a root access point at
which a connectivity is originated and a dead-end access point at
which the said connectivity is lost;
means for detecting the current access point to which the mobile
device is currently associated;
means for identifying the paths comprising the said current access
point;
means for computing the risks of losing the connectivity when
following each of the said identified paths from the said current
access point to the root access point or to the dead-end access
point of each of the said identified paths, the risk of losing
connectivity when following a path from a first access point to a
second access point being the inverse of the sum of the collected
durations of association corresponding to the access points
comprised in the said path from the said first access point to the
said second access point.
7 . The system of claim 6 , wherein it further comprises
- means for identifying at least an application in the mobile device
using the connectivity provided by the access point to which the said
mobile device is currently associated;
means for determining a threshold value above which, for the
identified application, the connectivity is considered lost;
- when a computed risk is greater than the determined threshold
value, means for triggering a handover of the identified application.
8 . The system of claim 7 wherein the identified application is a
disconnection-sensitive application.
9 . The system of claims 6 or 7 , wherein the paths are obtained from
a site map.
10 . The system of any of claims 6 to 8 , wherein the wireless shortrange
access point is a WiFi access point.
11. Computer program implemented on a processing unit of a
computer, said program including code sections for performing instructions
corresponding to the steps of a method according to any of claims 1-5.