WO2010001373A1

WO2010001373A1 - Handover method and system

Info

Publication number: WO2010001373A1
Application number: PCT/IB2009/053243
Authority: WO
Inventors: Zhensha Xiao; Chuang Lin; Rongfei Zeng
Original assignee: France Telecom SA
Current assignee: Orange SA
Priority date: 2008-06-30
Filing date: 2009-06-30
Publication date: 2010-01-07
Anticipated expiration: 2010-12-30

Abstract

A method for a User Equipment (UE) to select a first telecommunication network among a plurality of heterogeneous telecommunication networks available to said UE, said method comprising the acts of receiving information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks; deriving from said information a prediction of the QoS on each of the heterogeneous telecommunication network, and selecting if needed the first telecommunication network from the plurality of heterogeneous telecommunication networks taking into account the predictions of QoS.

Description

HANDOVER METHOD AND SYSTEM

Field of the Invention

The present invention relates in general to handovers in telecommunication networks and more specifically to handover predictions

Background of the Invention

A Vertical Handoff or Vertical Handover (VHO) consists, for a Mobile Terminal (MT), in changing the type of connectivity it uses to access a supporting telecommunication infrastructure (telecommunication network). A VHO involves switching between different heterogeneous networks. Existing VHO schemes comprise a few different solutions based on switching criteria such as either measurement of the Received Signal Strength (RSS) of the different radio access networks the MT may access or Quality of Service (QoS) provided by each different network to the MT for given applications.

A first solution uses RSS measurements. For example, for switching between a Wireless Local Area Network (WLAN) and a Global System for Mobile Communications/General Packet Radio Service (GSM/GPRS) network where communication data packets are exchanged using a Home Agent, this solution uses multiple tunnels of communication to minimize the communication data packet loss. The Home Agent (HA), the function of which being to route communication data packets to a MT, copies the same communication data packet destined to a given MT and sends various copies to multiple destinations through multiple tunnels or channels. This scheme uses timers and thresholds to perform the VHO properly. For instance, two thresholds for the Received Signal Strength (RSS), T_OFF and T_ON, are used to control the handover from and to the WLAN network and two timers, called dwell timers, TDWELLI and T_DWE_LL₂, are used to avoid the ping-pong effect (i.e. connecting untimely in turn to each of the networks while being on the edge of both network coverages). When the MT is using the WLAN network and detects that RSSWLAN < TOFF and this status persists for T_DWELLI, the MT performs a VHO, or handoffs, to the GSM/GPRS network. When the MT is using the GSM/GPRS network and detects that RSSWLAN > T₀N and this status persists for T_DWELL2, the MT handoffs to WLAN network. In this VHO scheme, the management of the MT interface is not efficient as the RSS measurement involves an energy consumption issue that is not taken into consideration. Indeed, both GSM/GPRS and WLAN interfaces need to be activated at all time in order to monitor continuously the signal strength of both GSM/GPRS and WLAN networks and decide whether to handoff to another network or not. This scheme focuses on latency to the detriment of cost and does not consider application QoS requirements.

Another solution to perform VHO between Institute of Electrical and Electronics Engineers (IEEE) 802.11-type networks and 3G networks is based on gateway/client architecture. A loosely coupled IOTA (Integration of Two Access technologies) gateway and a IOTA client software are used wherein the IOTA gateway manages the network interfaces and prevents network oscillation. This solution involves adding extra equipment to traditional telecommunication network architectures, focuses on latency to the detriment of cost and does not consider application QoS requirements.

Another solution proposes a protocol of VHO for overlay networks called HOPOVER (Handoff Protocol for Overlay Networks). An overlay network is a computer network which is built on top of another network. Nodes in the overlay can be thought of as being connected by virtual or logical links, each of which corresponds to a path, perhaps through many physical links, in the underlying network. HOPOVER enables smooth/soft handovers intra- and inter- network(s) through three procedures: pre-resource reservation, buffering and forwarding, but it only focuses on the radio part and does not consider application QoS requirements.

Another VHO solution is based on the Active Application Oriented Vertical Handoff (AAO) scheme. The AAO scheme performs VHO by considering application requirements. It allows the MT to actively decide when to handoff and which network to attach to according to the running application on the MT, i.e. it allows the MT to be served by the most suitable network for the given application.

The AAO schemes uses information on the QoS from a Location Service Server (LSS). The LSS is part of the Location-Based Services (LBS) domain. LBS provide some information of a MT according to its geographical position. The AAO scheme uses the concept of LSS, which originates from the paper "MIRAI: A Selection to

Seamless Access in Heterogeneous Wireless Networks" and which allows obtaining information on the QoS of a given MT. The LSS gathers for each MT QoS data on the networks each MT may be able to connect to in a given location at a given time. The AAO scheme presents an application QoS requirements evaluation method and describes two handoff decision algorithms, achieving efficient interface management and performing the handoff to the most suitable network at the right time. However, the AAO scheme has several drawbacks. First, this active scheme causes much latency, because the MT does not receive the information passively from the LSS when a handoff decision algorithm is executed, the MT needs to request the LSS for nearby networks information, the result is then sent back to the MT. As VHO is concerned, this increase of latency is a major drawback for the MT to perform VHO and thus to handoff communications when switching networks. Moreover, AAO does not provide a mechanism to deal with the inconstant situation of network like e.g. network congestion as the MT cannot request the LSS for new information on other networks when the network is congested. Furthermore, a VHO decision cannot be taken if the QoS requirements for a given application are not known yet, for example if when the MT boots up or changes application. These two major drawbacks make the AAO scheme not efficient enough and costly as VHO may not be performed in some cases even if QoS requirements are taken into account. Today there is a need for a VHO solution that can be easily implemented on the existing communication infrastructures, overcoming the drawbacks of the prior art.

Summary of Invention It is an object of the present system to overcome disadvantages and/or make improvement over the prior art.

To that extend, the invention proposes a method for a User Equipment (UE) to select a first telecommunication network among a plurality of heterogeneous telecommunication networks available to said UE, said method comprising the acts of:

- receiving information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks; - deriving from said information a prediction of the QoS on each of the heterogeneous telecommunication networks;

- selecting if needed the first telecommunication network from the plurality of heterogeneous telecommunication networks taking into account the predictions of QoS.

The method according to the invention allows a MT to perform handovers based on decisions that take into account predictions from up-to-date QoS of the different networks covering said MT. A prediction algorithm is performed in order to make a handover decision in advance for a given application, reducing thus the latency to a certain extent. Furthermore, the method according to the invention allows estimating and anticipating changes of network conditions so that the handoff accuracy is increased as it is based on a prediction algorithm as described hereafter.

The invention also relates to a system, a User Equipment and a computer program as described hereafter.

Brief Description of the Drawings

Embodiments of the present invention will now be described solely by way of example and only with reference to the accompanying drawings, where like parts are provided with corresponding reference numerals, and in which: Figure 1 schematically illustrates the method according to an embodiment of the present invention;

Figure 2 schematically illustrates the method according to an embodiment of the present invention;

Figure 3 schematically illustrates simulation results of the method according to an embodiment of the present invention.

Figure 4A and 4B schematically illustrates the system according to an embodiment of the present invention

Description of Preferred Embodiments The prior art AAO scheme also makes decision based on current information, in this case QoS data, which includes for instance bandwidth, latency, packet loss rate, etc... As networks situation changes quickly, a handover decision may not be suitable any more while executing the handover algorithm. In the method according to the invention, both current information (i.e. QoS data) and information obtained through a prediction scheme are used to anticipate or shorten handover decisions.

The method according to the invention concerns handover decisions between heterogeneous wireless networks, or Vertical Hand-Overs (VHO). There may be several heterogeneous networks in the same location that a Mobile

Terminal (MT) may connect to and where said MT may need switching from one wireless network to another in order to obtain a better QoS to satisfy its applications.

The method according to the invention uses information of QoS, or QoS data, on the networks a MT may connect to in a given place at a given time altogether with a prediction scheme. The goal is to perform the handover on the most suitable wireless network more accurately and quicker according to the up-to-date real value of network parameters of QoS (or QoS data) and to a prediction of said QoS data.

In an exemplary embodiment of the method according to the invention, the prediction scheme used is based on an Exponential distribution called Exponential- distribution based Prediction Algorithm (EPA), accessible to the man skilled in the art and described hereafter.

Figures 4A and 4B describe two illustrations of embodiments of the system according to the invention. The system according to the invention is illustrated on Figures 4A and 4B using two heterogeneous wireless networks. This illustration is not limiting as there may be a plurality (more than two) of heterogeneous wireless networks as long as the MT possesses the radio interfaces allowing said MT to connect to each of said heterogeneous wireless networks.

On both Figures 4A et 4B, a MT 430 is able to connect to two wireless communication networks, i.e. the MT 430 possesses the radio interfaces allowing said MT to connect to each of said wireless communication networks. Considering Figure 4A, a MT 430 is connected to a first communication network 410 through a Base Station (BS) 415. The MT 430 is able connect to a second communication network 420 through a BS 425.

Considering Figure 4B, a MT 430 is connected to a second communication network 420 through a Base Station (BS) 425. The MT 430 is able connect to a first communication network 410 through a BS 415.

On both Figures 4A and 4B, a Quality of Service (QoS) Provider Node (QPN)

440 collects QoS data from both communication networks 410 and 420 and sends to the MT 430 QoS data that will allow accordingly said MT to implement the present method, i.e. to take or not a handover decision between the communication networks 410 and 420.

More generally, the QPN 440 is a server comprising a database allowing the collection of QoS information on heterogeneous wireless networks the MT 430 may be able to connect to. The QPN 440 may be for instance an enhanced Home Location Registrer (HLR), a Mobility Manager, a Location Service Server (LSS) etc... In the illustration of embodiment of the method according to the invention illustrated hereafter on Figure 1 , the QPN 440 is a Location Service Server (LSS). The QPN 440 is able to get the location of the MT 430 (for example through a HLR, or directly from the MT itself) and/or derive the list of heterogeneous wireless networks the MT may be able to connect to in order to further send to the MT the corresponding QoS data.

The QPN 440 may for instance be:

- centralized as being one entity connected to each of the different heterogeneous wireless networks or, if not connected to each of the different heterogeneous wireless networks, one entity able of collecting the QoS data from said different heterogeneous wireless networks,

- distributed as being one different entity connected to each different heterogeneous wireless networks.

Once the QoS data are collected from the different heterogeneous wireless networks, the QPN 440 filters or selects and sends to the MT 430 the QoS data that correspond to the different heterogeneous wireless networks the MT 430 may be able to connect to. From the received QoS, the MT 430 is able to perform a handover prediction algorithm described hereafter using Figure 1 in the illustration of embodiment of the method according to the invention.

Figure 1 describes an illustration of an embodiment of the method according to the invention.

In an act 110, a MT receives from a Location Service Server (LSS) QoS data on the networks said MT may be able to connect at a given place at a given time (current place, current time). Said QoS data is called QoS data (i) (i being the current instant). The QoS data provided by the LSS may be for instance the bandwidth, the latency of available wireless networks said networks. In order to let the MT get the up-to-date information of its current and nearby networks, the LSS pushes or sends spontaneously a LSS_packet (an example of LSS_packet is shown hereafter in Table 1) to the MT periodically (for example every 10 seconds). The LSS_packet comprises the QoS data (i) for the MT. The LSS needs to know the location of the MT in order to be able to send it the QoS data associated with the networks the MT may be able to connect to at time t. This location information may be obtained by the LSS from another network entity that knows said location (e.g. the Home Location Register (HLR)) or for instance from a location-based service (e.g. GPS) etc...

Table 1 describes an example of LSS_packet:

After receiving this LSS_packet, the MT updates a table called MT_table stored on the MT which is shown in Table 2 hereafter by storing the received QoS data (i) in an act 120. The MT_table should be able to store the information received in at least the last two LSS_packets (at the time (i-1 ) and i), i.e. at least QoS data (i-1 ) and QoS data (i).

Table 2 describes an example of MT_table:

The up-to-date information to collect on a network in order to take a VHO decision comprises the Quality of Service information for said network. The QoS is defined through a number of network parameters such as, for example, latency (or time delay) for receiving a stream of a given application, bandwidth (e.g. available for a given application...), network congestion, throughput, cost etc... The QoS is defined in terms of levels of QoS for a given application for a given network with given parameters. Levels of QoS are defined as being ranges of values for a given network parameter. In the hereafter illustration of embodiment of the method according to the invention, a level of QoS is defined for a given application for a given network using a range of values for the parameters of said given network. In order to find the most suitable network for using a given application, these ranges of values or requirements should be defined for said application for a number of network parameter for said given network such as e.g. bandwidth, latency... The requirement of each parameter is a range instead of a constant, so upper bound and lower bound should be given.

The QoS may be defined for n network parameters where:

- // is a network parameter where {0<=j<=n},

- Uf_j is the upper bound (maximum) for network parameter f_j - Lή is the lower bound (minimum) for network parameter f_j.

Equation (1 ) is thus a function to quantify the level of QoS provided, for n network parameters, by a network m at the time /:

where the w_k are the weights of parameter f_n. w_k may be assigned by the application according to each parameter significance for said application. For instance, for massive data stream application, bandwidth requirement is more important than latency, so we may define e.g. w_bandwidth = 0.7 and wjatency = 0.3. fm,n(i) is the normalized value of the real value f'_m,n(i) in the network m at time / as described hereunder. This allows using a continuous and linear function when estimating the level of QoS with equation (1 ).

When a parameter is important to consider for an application, it is called a high factor (it may be e.g. bandwidth, throughput...). High factors are parameters with value that should be as high as possible. For a high factor, the normalization is:

For instance, if the real value f'_m,n(i) is lower than Lf_n, this factor does not meet the application requirement.

When a parameter is less (in comparison to a high factor) or not important to consider for an application, it is called a low factor (it may be e.g. latency, cost...). Low factors are factors with value that should be as low as possible. For a low factor, the normalization is:

In the present exemplary embodiment of the method according to the invention, the statistics of QoS evaluation change in a small time interval Δ/ , with Δ/ -» 0 ₍ obey an Exponential distribution.

The value of QoS_m(i+1) may be evaluated through the real value of network at the time i-1 and /.

\ -e -x

QoS_1n (z + 1) = QoS _m (0 + AQ_n, (i + 1) = QoS _m (Q + ^■Δ&,(0 e^A -l (3)

l - e -x

QoS_m(i + l) = QoS_m(i) + (QoS _m(i) - QoSJi - I)) e^λ - \ (4)

In an act 130, equation (3) (or (4)) is used to calculate the prediction value at the time i+1 of QoS in network m, i.e. predict QoS data (i+1 ) for network m the MT may be able to connect to at instant i using QoS data (i-1 ) and QoS data (i). This act is performed for each network the MT may be able to connect to at instant i using the information on each of these networks provided by the LSS. In an act 140, the prediction values for each network the MT may be able to connect to are compared in order to select the most suitable for the MT, in particular the most suitable for a given application the MT would privileged among others (e.g. data streaming, internet, voice communication etc.). The algorithm of the method according to the invention is executed while the

MT is moving in the network and each time the MT boots up or changes application. The LSS sends the information periodically, and the algorithm executes simultaneously so that the actual network situation may be considered.

When an application starts running on the MT, the requirements for said application are then compared to the QoS data received from the LSS. After calculating the prediction value ^° ™^"Λ^{/ + 1}\ the MT may try to find a nearby network which is more suitable for satisfying said application requirements. If it exists, this network may be chosen as the target network. Otherwise the MT may stay with the current network. The best network from all the nearby networks may not necessarily be chosen in case the MT is not allowed to (e.g. barred by the network operator), if there are rules implemented by the network operator to privilged some networks rather than others depending on the MT or on the application, or if the targeted network is not stable for instance. In this case, it may be possible to run the algorithm again to further satisfy those criteria.

An example of algorithm to illustrate the method according to the invention could be as follows:

Check the *ⁿ *ⁿ of the application and define the value of ^Wj

find J™™*'" ^ ^' ' and ^^<"™" ^π ^ of current network from the database

calculate J™™' "^¹ ' and f™™^»*v> ( Equation (I)(Z)) calculate Q^oS _™™, 0 - 1) _and QoS^_n, (/) ₍ Equation (1) )

calculate @^oS™™, 0 + ^l) ( Equation (4) ) repeat until all nearby networks are tested calculate ^QoS→^{(i + l)}

i_f QoS_n→(i + \)_>QoS_cumnt{i + 1) _then

select this nearby network as target network

calculate ^¹W, O ⁺ 2) _am/ QoS_curreπl(i ₊ 2)

if

_>QoS_clιrreM(i ₊ 2) _tnen

handoff to target network exit else target network is unstable -> remove this network from nearby networks list else remove this network from nearby networks list end stay in current network.

In the system according to the invention, the MT is adapted to:

- receive information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks it may be able to connect to;

- deriv from said information a prediction of the QoS on each of said heterogeneous telecommunication networks;

- select if needed a first telecommunication network from the plurality of heterogeneous telecommunication networks taking into account the predictions of QoS.

The present method may be implemented through an agent or an application or a computer program running on the MT for performing a VHO when QoS predictions allow it. Said agent, application or computer program may be for instance provided by an operator of one of the heterogeneous telecommunication networks or downloaded from a server...

Figure 2 describes a comparison between the AAO scheme from the prior art and an illustration of an embodiment of the method according to the invention based on the EPA. The method may be executed while the MT is moving in the network and each time the MT boots up or changes application. In the method according to the invention, the LSS sends the information of QoS periodically so that the MT always has up-to-date information in order to take VHO decisions. Figure 2 shows the comparison of latency between the method according to the invention and the AAO scheme of the prior art. The MT is connected to the Location Service Server (LSS) through a Base Station (BS) of a first network. The prediction time is the time when the method according to the invention is performed to decide or not of a VHO. Variable i is the current prediction time, i-1 means the previous prediction time, i+1 is the next prediction time. Each prediction time should be associated with information on the QoS (QoS data) obtained from the LSS by the MT for at least two networks in order to compare the QoS data on these two networks. On Figure 1 , the AAO scheme from the prior art involves a waste of time to transfer the messages between the MT and the LSS through the BS (MT^BS^LSS^BS^MT). In the method according to the invention, LSS packets comprising information on QoS are periodically sent to the MT so that the MT always has up-to-date information in order to take VHO decisions, reducing drastically the latency in regard to the AAO scheme of the prior art.

Figure 3 shows an example of simulation results using a Universal Mobile

Telecommunication System (UMTS) network and a Wireless Local Area Network (WLAN). Figure 3 shows the comparison between an example of embodiment of the method according to the invention and the VHO decision algorithm described in the AAO scheme from the prior art. The x-axis is the simulation time that lasts 130s, and the y-axis is the amount of transferred data at certain time. The positive value depicts MT is under UMTS and the negative value depicts MT is under WLAN. Obviously, the MT may switch from UMTS to WLAN timely and accurately using the method according to the invention. Moreover, the method according to the invention allows discovering network congestion and adapting to the variable network conditions as quickly as possible. The simulation scenario comprises a Corresponding Node (CN) (for example a Callee), a Base Station (BS) in UMTS, an Access Point (AP) in WLAN and a MT. The MT is a device with two interfaces: one for WLAN and one for UMTS. There is a one kilometer square plain area under UMTS coverage. Within this area, a WLAN hotspot of 250-meter diameter is built. The factors (QoS data) considered by the applications at the MT are bandwidth and latency.

Table 3 shows the value of these factors used in the simulation:

The simulation lasts 120s. From the beginning, the MT starts moving in a straight line with the speed of 5m/s. After 10s, the voice conversation application is running on the MT. It ends at 40s. Then, massive data streaming begins and ends at 80s and video streaming starts to run later. After 100s, the WLAN is overloaded and happens to be in congestion.

Positive values shown on Figure 3 indicate that the MT is under UMTS coverage and negative value indicate the MT is under WLAN coverage. The method according to the invention allows the MT to switch from UMTS to WLAN timely when the MT enters the WLAN region. Using the AAO scheme, the MT does not have any action because no application changed. Moreover, the AAO scheme completely neglects the real-time change of network conditions. So after 100s, only the method according to the invention allows discovering the network congestion accurately and handover to the UMTS network quickly.

Claims

1. A method for a User Equipment (UE) to select a first telecommunication network among a plurality of heterogeneous telecommunication networks available to said UE, said method comprising the acts of:

- receiving information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks;

- deriving from said information a prediction of the QoS on each of the heterogeneous telecommunication networks; - selecting if needed the first telecommunication network from the plurality of heterogeneous telecommunication networks taking into account the predictions of QoS.

2. A method according to the preceding claim, wherein said information on the QoS comprises at least a couple of successive values for a given telecommunication network.

3. A method according to any of the preceding claims, wherein the act of receiving information comprises the act of receiving one single message at a time comprising all the information on the QoS for each one of the available heterogeneous telecommunication networks.

4. A method according to any of the preceding claims, wherein the prediction of the QoS on each of the heterogeneous telecommunication networks is carried out using an Exponential-distribution based Prediction Algorithm.

5. A method according to any of the preceding claims, wherein the received information on the QoS corresponds to the UE location.

6. A User Equipment (UE) for selecting a first telecommunication network among a plurality of heterogeneous telecommunication networks available to said UE, said UE being adapted to:

- receive information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks;

- derive from said information a prediction of the QoS on each of the heterogeneous telecommunication networks;

- select if needed the first telecommunication network from the plurality of heterogeneous telecommunication networks taking into account the predictions of QoS.

7. A User Equipment according to the preceding claim, said User Equipment is further adapted to predict the QoS on each of the heterogeneous telecommunication networks using an Exponential-distribution based Prediction Algorithm.

8. A User Equipment according to any of the preceding claims 6 and 7, said User Equipment being further adapted to send its location to a Quality of Service Provider Node (QPN), said QPN being adapted to collect information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks corresponding to said UE location.

9. A telecommunication system comprising:

- at least one User Equipment (UE), - a plurality of heterogeneous telecommunication networks available to the at least one UE,

- a Quality of Service Provider Node (QPN) adapted to collect and send to said at least one UE information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks available to said at least one UE; the UE being adapted to:

- receive information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks from the QPN;

10. A system according to the preceding claim, wherein said information on the QoS comprises at least a couple of successive values for a given telecommunication network.

11. A system according to any of the preceding claims 9 and 10, the QPN being adapted to send one single message at a time comprising all the information on the QoS for each one of the available heterogeneous telecommunication networks.

12. A system according to any of the preceding claims 9 to 11 , wherein the prediction of the QoS on each of the heterogeneous telecommunication networks is carried out using an Exponential-distribution based Prediction Algorithm.

13. A system according to any of the preceding claims 9 to 12, wherein the QPN is adapted to collect information on the Quality of Service (QoS) on each of the heterogeneous telecommunication networks corresponding to said UE location.

14. A system according to any of the preceding claims 9 to 13, wherein the UE is adapted send its location to the QPN, said QPN collecting information on the QoS based on the received location.

15. A computer program providing computer executable instructions stored on a computer readable medium, which when loaded on to a data processor causes the data processor to perform a method for performing a handover according to claims 1 to 5.