Handover between 2G, 3G and WLANs

Iqbal Hussain, Rabiullah Khattak

II.
AbstractThis study project mainly deals with handover between 2G, 3G and WLANs. First main part deals with the handover between GSM and WCDMA and vice versa which is a typical example of the handover between 2G and 3G. In section the main problem associated with such handovers are discussed and then for the solution the required construction and design of the mobile terminal and transmission format. So that user is unaware of such handover and still preserving the same data rate. The handover decision is discussed in section III. In the section IV the measurement and reporting about the handover and re-selection of cell are discussed. Section V is reserved for the protocol flow diagram and sequence how this handover will take place. Section VI deals with handovers procedures, algorithms and metrics between WLAN and GPRS and vice versa. A procedure based on IP layer for handover between GPRS and WLAN is discussed and the same procedure can be extended to describe the handovers between any two types of wireless networks

MAIN PROBLEM

I.

INTRODUCTION

Third generation mobile networks have new set of communication protocols due the rapid increase in data services and development in hardware and software. As the commercial companies compete with each other to increase their market potential, so to save 2G customers and to bring a new 3G customers, there should be ways (maybe not all these ways will be standard), to do that, so the subscriber can move from one cell (covered by one network) to another (covered by another or the same) seamlessly, the customer may actually will be unaware of what actually done and this is called Handover or Handoff. Wide band code division multiple access (WCDMA) is mainly used for high bandwidth and data requirements services as due to the higher bandwidth Bit error rate is lowered and so error free communications become possible. Due to reducing the bit error probability high order modulation can be used so enhance the data rate as well. WCDMA is useful for audio and video streaming purposes. Seamless handover between the WCDMA and Global system for mobile communication (GSM) is very important both from operator and user point of view. From operator point of view if the hand over is performed in the existing infrastructure then it will save a lot of money and from the user point no communication break should occur whether it is 2G environment or 3G environment. So the user terminal has the capability of inter working to handover between WCDMA and GSM with out any interruption. In section VI handover between GPRS and WLAN is discussed and various solutions are suggested.

In order to avoid complications and make possible the feasible implementation of such mobile terminal and network the mobile terminal must communicate only one the technology either WCDMA or GSM at a time. It will lead to less expensive mobile terminal and avoid huge signalling in the network. In order to utilize the existing infrastructure in more efficient manner message of one technology must be enclosed in the format of other technology. For example to make a handover from WCDMA to GSM then the GSM handover message must be enclosed in the packet format of the WCDMA and after receiving this packet the mobile terminal extract the message of handover of GSM and pass it through the interfaces nodes of the networks[3]. The same procedure will be used when handover is to be made from the GSM to WCDMA. Handover message length is also very important factor because if the handover message is very long then it will degrade the performance of the networks. In case of handover from WCDMA to GSM then the length of the handover message must be short so that the system performance is not degraded. Similarly in case of handover from GSM to WCDMA it must be noted that due to the limited data rate of GSM it make very difficult to carry the WCDMA long messages. Therefore in the handover messages only short references should be made. The mobile terminal has a combined portion where it can receive both WCDMA and GSM messages and also contain some part which can individually process the WCDMA or GSM as shown in Figure 1.

WCDMA handover message to GSM

Common WCDMA and GSM part of the Mobile Equipment GSM handover message part is extracted in WCDMA part

GSM WCDMA specifice specific part GSM handover part of the of the mobile message Mobile equipment equipment

Handover message passed to the GSM part Figure 1: WCDMA handover message to GSM

There must be a mechanism in the mobile terminal such that when it communicating on WCDMA technology it performs measurement for the other GSM technology and vice versa so that whenever the handover is required it takes the decision on the measurements [3]. As the WCDMA technology is working on continuous transmission and reception so to perform measurement for the GSM technology is very difficult task. For this purpose a gap is introduced in the transmission and reception in the continuous communications of WCDMA. However for the constant data rate services e.g. voice the data rate before MOBILITY MANAGEMENT BETWEEN WCDMA AND GSM the gap and after the gap is increased as shown in Figure 2. III. GSM measurements take place in the interval of gap between transmission and reception. There are two types of mobility management one is the Mobile terminal-controlled mode and other one is the Network-controlled mode. In the mobile terminal-controlled mode the mobile terminal take the decision to which cell it will connect but however the network factors also have a lot influence on the decision of re-selection of cell but the final decision is taken by the mobile terminal. In the network-controlled mode the network take decision on the basis of the measurements from the mobile terminals. The network select the most optimum cell to which the mobile terminal is to be connect in the sense of present load and efficient utilization of the resources. There are two procedure available in which it is described that how the network will handle the re-selection of the cell for the mobile terminal. One procedure is handover which is normally used for the circuit switching

Instead of introducing air gap the measurements task can be accomplished by using two receivers in the terminal however all these effect will increase the battery consumption. Standby mode of terminal is very important for the power saving in the mobile terminal. However if we use the dual receiver then it will mainly effect the standby time. However the monitoring frequency of the radio channels for the other technology e.g. WCDMA is lowered when the quality of the present technology e.g. GSM is adequate. Also no such hard rules are existing for the reselection of the cells between WCDMA and GSM as exist for the re-selection of cells in standard GSM.

such as voice. In handover the network first reserve the resources in the cell to which this terminal will have to be connected and after the reservation the mobile terminal is order to connect to the corresponding selected cell. The cell change order procedure is used in the packet switching environment in which case the mobile terminal not only forward the current serving cell information but also of the other cell information as well. When some pre-defined condition is satisfied then the re-selection decision is issued by the network. IV. RE-SELECTION OF CELL BETWEEN WCDMA AND GSM
MT

BSS

RBS

BSC A Gb

Um BSS : Base Sation Sub-System RBS : Radio Base Station Sub-System. BSC : Base Station Controller. RNC : Radio Network Controller Uu

Core Network

The architecture of the GSM and UMTS are shown in Figure 3. Base station subsystem (BSS) consists of several Base station controllers (BSC) and Radio base station subsystem (RBS). Several RBSs are controlled by a single BSC. Similarly several nodes B are controlled by a single Radio network controller. The priority of the WCDMA or GSM cell is based upon the signal strength. When the mobile terminal is the WCDMA cell then it can re-select the cell either in GSM or WCDMA network in idle mode or in connected mode. When the mobile terminal has dual receivers then it can select the GSM cell or other any WCDMA cell which has high priority than the present WCDMA cell, by high priority means the cell to which the selection is made must have signal to noise ration than the present cell signal to noise ratio. Monitoring of the best cell is performed continuously by mobile terminal or start when the signal to noise ratio of the current cell is lowered from a certain threshold. When the signal to noise ratio measured by mobile terminals of certain cell exceed some threshold value then the mobile terminal is free to re-select that cell. The network received the information transmitted by the mobile terminal about the target cell. As soon as the network received these information then instructions issued by network toward mobile terminal about power managements. These instructions includes that how much power mobile terminal will use when accessing the target cell and also tell about the radiated power by mobile terminal when using that target cell.

NODE B

Iu RNC

UTRAN

Figure 3: GSM and WCDMA nodes and infterface invovled in cell re-location and handover process.

Frequent re-selection can cause heavy signalling and may degrade the network performance [2]. So to avoid this frequent re-selection in case of very dynamic mobile terminal penalty time and temporary offset are used. In order to track the dynamic mobile terminal WCDMA uses broadcast signalling messages. When the mobile terminal is providing packet-switched data in the WCDMA network and based on the measurement the re-selection decision has been taken in the favour of GSM cell. Then the mobile terminal request the base station subsystem of the target cell in the GSM network for route updates. The base station subsystem forward the corresponding route update request to the core network of GSM to which it attached. The core network request the serving Universal Mobile Telecommunication system (UMT) Terrestrial Radio Access Network (UTRAN) for the queuing data for mobile terminal on the downlink. The UTRAN forward the corresponding PDUs to the core network. After that the core network request the release connection from the UTRAN and after succeful release of the connection from the UTRAN core network forward the route updates to the mobile terminal.

MT

UTRAN

Core Network

BSS

Message from MT to BSS about routing updates

SRNS context Request SRNS context response SRNS data forward command

Forward PDUs Iu release command Iu release complete

Routing updates accept

Figure 4: WCDMA to GSM re-selection

V.

WCDMA TO GSM CELL CHANGE ORDER

WCDMA to GSM cell change order is shown step by step in Figure 5 and discussed in detail in [1].In this situation the mobile terminal takes measurements of the target GSM cell and inform network accordingly. The signalling required for this cell change order is similar to that of the cell re-selection signalling. But in this case the request of cell change order is initiated by the UTRAN to the core network.

As soon as the mobile terminal receives the handover message from the GSM cell then it moves toward the target cell and establish a radio connection. This handover is detected by the GSM Base station subsystem and core network. After this handover the mobile terminal issues the message of the handover complete and GSM BSS A. HANDOVER FROM WCDMA TO GSM responds to this handover message toward core network by confirming the completion of handover. Then the core When the quality of the current WCDMA cell becomes network issue the release of the corresponding mobile degraded and quality of the neighbouring cell of GSM terminal from the UTRAN and UTRAN confirm this exceeds a certain threshold, then the following messages release by a positive acknowledgement flow between the two technologies. Mobile terminal takes the measurements of the B. HANDOVER FROM GSM TO WCDMA neighbouring GSM cell and concludes that the quality of the target cell exceeds a certain value and reports it to its BSS sends messages the mobile terminal to take corresponding UTRAN. UTRAN initiate the handover measurement about the WCDMA cells. The messages request to the core network for the reservation of the from the base station sub-system includes the criteria of resources in the corresponding GSM cell. GSM cell the measurements and information about the reporting to allocate the radio resources for the handover and forward the base station sub-system. this message to the UTRAN via WCDMA air interface. If the criteria for handover are fulfilled then BSS initiate This message includes all the information of the radio for the allocation of resources to the WCDMA system by resources. requesting to the core network as shown in Figure 6. The core network forward this message to the UTRAN and then after the allocation resources of the WCDMA cell is completed by the core network then the UTRAN issue, the handover command to the core network by relocation request acknowledge. Core network then forward this message to the mobile terminal via Base station Subsystem which includes the configuration details. After receiving this message in the mobile terminal, the mobile terminal is

switched to the frequency used by the WCDMA cell and start synchronization with WCDMA system. After successful completion of handover the mobile terminal issue the hand over complete command and then resources of the GSM system is released.
MT UTRA N CN BSS

Measurement information Measurement Report Handover required Relocation request Relocation request Acknowledge Handover command

(Transmission Control Protocol/Internet Protocol) level, transport level (TCP), network level (IP) and underlying protocol level [6] One of the solutions is to incorporate inter-technology roaming protocol layer between application and transport protocol layer. One approach used for this solution is the introduction of MOVE [4] and Mobile TCP/IP [5]. Other approaches for this solution are related to modification of the current protocols placed above the TCP/IP stack. On of them is related to the work on Xinterface mobility. In this approach, a Mobility Gateway is used to provide mobility at the transport layer. Examples of this approach are indirect TCP and MOCKS. Internetworks handover solution can also be found in the layers below IP [6]. At the network layer the most significant approach is the IETF (Internet Engineering Task Force) Mobile IP. Network layer implementations of this approach have been implemented in several projects like Monarch [7], MosquitoNet [8] and Deadalus/Barwan [9]. In [10] network layer handoff strategies have been discussed and it is expressed how Fuzzy Logic and Neural Networks class of algorithms can be used to control Intertechnology handoff procedure. Handovers between GPRS (General Packet Radio Service) and WLAN (Wireless Local Area Network) is used as an example in [10]. A. DIFFERENT SCENARIOS FOR INTERCONNECTING THE SERVICES OF WLAN AND GPRS In [11] different scenarios are presented to implement inter-technology roaming. There are five different scenarios that make it possible to implement inter-tech roaming between WLAN and GPRS as shown in Figure 7.

Inter-system to UTRAN Command

Physical layer synchronisati on

Relocation Detect

Handover to UTRAN complete

Relocation Complete

Clear Command Clear complete

Figure 6: Handover from GSM to WCDMA

VI. HANDOVER BETWEEN GPRS AND WLAN Today telecom networks play a significant role in various fields of life. The vast expansion of high speed local area networks and demand for multimedia services has led the broadband telecom networks to concentrate on research, development and standardization activities all over world. In case of heterogeneous networks, mobility is very important. Mobility can be defined as the accessing of network resources by mobile users at any time and at any place within the coverage area of network. To provide mobility in heterogeneous networks, we have to consider every layer of communication networks in order to get power control in physical layer, traffic management in the data link layer, mobility management in the network layer and communication optimization in the transport and application layers. For mobility in heterogeneous networks, both operator and software industries are looking for solutions, so that implementation can be done with least changes to current infrastructure. There are several possible levels to solve the problem of mobility in heterogeneous network: above TCP/IP

In the first two scenarios WLAN is implemented as a base station in the GPRS network. In these architectures GPRS acts as a master and WLAN is used as a slave. In the third scenario Mobile IP is used for implementation of inter-tech roaming. In the fourth scenario GPRS is used as an AP (Access Point) in WLAN to implement the inter-tech roaming, in this case WLAN acts as a master and GPRS acts as a slave. In the fifth scenario Mobility Gateway is used to interconnect to the two networks. In the 3rd and 5th scenarios GPRS and WLAN are interconnected as peers in a larger network. In order to get the interconnectivity between the networks, we want to do little changes in the existing networks and technologies particularly at the lower layers such as MAC (Medium Access Control) and physical layers. Therefore the 3rd and 5th scenarios are preferred because they treat GPRS and WLAN as peer networks, without any changes to GPRS or WLAN specific protocols. In the 3 and 5 scenarios from service scenarios point of view when wideband local area service is present, the terminal will use it. In case of non-availability, the terminal roams to lower speed underlay wireless data service. The test case presented in [10], the user has an application running on the terminal and moves out of the range of WLAN. The protocols involved in the 3rd and 5th scenarios are shown in the Figure 8 and Figure 9. The mobile data service should notice the availability and non-availability of networks and should fall back from WLAN mode to an underlay, low speed mobile data service (GPRS or CDPD (Cellular Digital Packet Data)). In order to control these activities daemon program is required for monitoring of the network resources so that when one network is unavailable, routing table should be adjusted and make the activation of mobile IP registration procedure.
Mobile

B. HANDOVER PROCEDURE BETWEEN GPRS AND WLAN Fig.10 shows the block diagram of Handover (HO) from WLAN to GPRS and handoff from GPRS to WLAN. Figure 12 is related to the interactions among the algorithm, metrics and outcomes of executing algorithm. The issues related to HO architecture (step-by-step description of the procedure to relate the elements), HO metrics and the HO decision mechanism and algorithm are discussed as below:

rd

th

Working WLAN

Working GPRS

HO Algortthm A B

Near WLAN?

No

C Ready for GPRS Yes No Prepare C

Yes Ho Algorithm

Prepare

B Wait M<N(Sec)

A Prepare

Wait N(Sec)

Handover Handover to WLAN

Application TCP IP
SNDCP LLC LLC RLC RLC MAC MAC PHY PHY LI bis Fr. Rel BSSGP BSSGP Fr. Rel LI bis IP BB L2 BB L1 IP BB L2 BB L1 PDN L1 PDN L1 LLC UDP UDP Base Station SGSN IP IP SNDCP GTP GTP PDN L2 803.3 LLC IP GGSN Home or Foreign Agent

Figure 10: Handoff Procedure for inter-tech Mobility

803.3 MAC 803.3 PHY

C. HANDOVER FROM WLAN TO GPRS The following steps take place when a mobile goes out from the coverage of WLAN and enters the coverage of GPRS as shown in Figure 11. 1. Signal received from the AP in the WLAN is strong. 2. Signal from the AP becomes weak 3. HO algorithm at the mobile host decides to handoff from WLAN to GPRS.

Access Point

On the LAN IP IP 802 LLC 802.3 MAC 802.3 PHY

802 LLC 802.11 MAC 802.11PHY

802 LLC 802.11 MAC 802.11P HY 802.3 MAC 802.3 PHY

802 LLC 802.3 MAC 802.3 PHY

Figure 8: Interconnecting WLAN and GPRS by using Mobile IP(Scenario 3)

4. Start/update Mobile IP: Foreign Agent (FA) in the Mobile Host (MH) gets activated using the Mobile-IP protocol and MH uses the new IP address. 5. The Home Agent (HA) in the GPRS is informed about the new IP address through the Mobile IP protocol.

4. Update Mobile IP: FA in the MH is deactivated by the Mobile IP protocol and the regular IP address is used. 5. HA in the GPRS is informed by the MH through the Mobile IP protocol. The HO mechanism is shown in Figure 12 above. The most significant part of this mechanism is the algorithm based on FL (Fuzzy Logic) and Neural Network (NN) class of algorithm [12, 13]. The HO algorithm shown in Figure 12 has three possible outputs: A, B and C. When a mobile host is in the coverage of underlay network (WLAN) the output A (Relax) shows that MH is receiving a strong signal in the underlay network (WLAN) i.e. link quality is satisfactory and there is no chance for handoff to be happened. The output B (Alert) prepares the system for handover to be happened in the near future as link quality of underlay network (WLAN) is decreasing gradually. The output C (Handoff) invokes the handoff procedure to the overlay network (GPRS) as the handoff is needed when output is given by the mechanism. Mechanism shown in Figure 12 does not show the possibility of failure because it is assumed that overlay network is always available. If a MH is in the overlay network (GPRS), it checks its position on regular basis, so as to measure its proximity to underlay network. The optimum way to find the underlay network is to compare the geographical coordinates. Once the MH realizes the underlay network availability, it then starts searching for the presence of WLAN pilot and invokes the fuzzy or neural net engine to calculate the quality of connection. The output in this case is the same as in the case of handoff from WLAN to GPRS except if the connection indicator is unstable (B), no handoff preparation are made. MH waits till stability (A) of link and then handoff to underlay network is performed while output C shows that no handoff should be performed.

Monitoring

Measurements

Algorithm

Stable WLAN A

The difference between the above approaches is that a user does not loose its connection while working in the overlay network. Intra-overlay handoff (within GPRS) is Figure 12: Handoff Mechanism supported by the lower GPRS layers. So a user of overlay network layer rarely checks availability of underlay network hopefully aided by the geolocation information to D. HANDOVER FROM GPRS TO WLAN avoid useless search if the MH is e.g. in a different country. If the location is provided by geographical In the reverse situation where a MH moves from GPRS coordinates, the MH can save underlay network to WLAN the following steps happen: coordinates e.g. at the time when it left it last time and start searching for the underlay network as soon as it 1. Signal from the WLAN does not exist. approaches this area. A user controlled handoff procedure 2. MH detects a signal from the AP of a WLAN. is also possible i.e. a user clicks a button in the screen of 3. HO algorithm decides on the HO from GPRS to laptop when he arrives to his office environment and WLAN. mobility management works automatically.
B C

Unstable WLAN

Poor WLAN

For seamless operation of the system, handoff for GPRS to WLAN algorithm is the most important because the MH experiences severe degradation of the service quality as it moves away from the underlay network coverage and it has to handoff very fast to maintain higher layer connection. The triggering algorithm has been divided into two parts as overlay (WLAN) and underlay (GPRS) in order to facilitate the features such as power saving by power downing the unused interface cards, loading of the interface drivers on demand and the proposed algorithm also allows for an alert or possible several levels of alerts, which would make the system to be able to handle upcoming handoff.
VII.

discussed, however this procedure can be extended to make handover between any type of two wireless networks e.g. 3G and WLAN. REFRENCES [1] Gertie Alsenmyr, Joakim Bergstrm, Mattias Hagberg, Anders Miln, Walter Mller, Hkan Palm, Himke van der Velde, Pontus Wallentin and Fredrik Wallgren, Handover between WCDMA and GSM, Ericsson Review No. 1, 2003. [2] Client Smith, Daniel Collins, 3G Wireless Networks, Second Edition, 2007. [3] Roland Heicker, Stefan Jelvin and Bodil Josefsson, Ericsson Seamless Network, Ericson Review No.2, 2002. [4] Meggers et al., 1996, Roaming Between GSM and Wireless LAN, ACTS Mobile Telecommunications Summit, Granada, Spain, vol. 2, Nov. 27- 29, pp. 828 834. [5] X. Qu, J. Xu Yu, and R. Brent, Implementation of a Portable-IP System for Mobile TCPDP, ANU Computer Science Technical Report, November 1997.

METRICS

Several handoff metrics have been used in various voice and data mobile networks. The most popular of them is the received signal strength (RSS) handoff metric. In order to get the optimum use, the RSS is combined with some parameters such as hysteresis margin, traffic, velocity advantages and disadvantages in a variety of handoff algorithms. Handoff metrics have a close relation with the type of [6] R. Pichna, J. Vallstrom, M. Ylianttila, J. Talvitie, algorithm that may be ultimately employed to make Current Wireless Inter-technology Roaming Efforts, handoff decision [14]. The broad class of handoff metrics Proceedings of the 1 Euromicro Summer School in used include RSS, beacon packets, SNR, BER and packets Mobile Computing, Oulu, Finland, August 1998, pp. error rates with parameters such as hysteresis margin and 1 19-131. traffic. The reason for using such a broad class of handoff [7] A. Hills and D. Johnson, A Wireless Data Network metrics is the algorithms used in handoff are based on Infrastructure at Carnegie Mellon University. IEEE Neural Network (NN) and Fuzzy Logic (FL) which can Personal Communications, 3( 1):56-63, February 1996. make highly sensitive decision based on large number of [8] M. Baker, X. Zhao, S. Cheshire, J. Stone, inputs. In NN and FL the decision is made on a pattern Supporting Mobility in MosquitoNet, Proc. of 1996 match the between current inputs and previously observed USENIX Technical Conference, January 1996. situations. Future work regarding the handoff is to [9] S. Seshan, H. Balakrishnan, R. Katz, Handoffs in concentrate on investigating the performance and Cellular Wireless Networks: The Daedalus sensitivity of NN and FL algorithms to different HO Implementation and Experience, Kluwer metrics. International Journal on Wireless Communication Systems, 1996. VIII. CONCLUSION [10] M. Ylianttila, R. Pichna', J. Vallstrom', J. Makela, As the good interworking between GSM, which is a A. Zuhedi', P. Krishnamurthy, and K. Pahlavan Handoff typical 2G network, and WCDMA, which is a 3G network procedure for Heterogeneous Wireless Networks Global enables users to seamlessly handover between these two Tele-communication Conference - Globecom'99, Future different generation networks. The difficulties between the Wireless Communication System. [11] P. Krishnamurthy, K. Pahlavan, A. Zahedi, J. interworking between WCDMA and GSM can be Vallstrom, J. Talvitie, R. Pichna, and M. Ylianttila, overcome by dual mode mobile terminal. Another way to Handoff in 3G Non-Homogeneous Mobile Data minimize the complexity by using the compressed mode Networks, in the Proceedings of MTT-S European transmission. Cell re-selection, cell change order and WirelessP8 in Amsterdam, October 1998. handover between WCDMA and GSM play important role [12] N. Tripathi, Generic Adaptive Handoff Algorithms in this regard. using Fuzzy Logic and Neural Networks, Ph.D To make handover between WLAN and GPRS, there are Thesis, Virginia Polytechnic Institute and State various procedures, algorithms and metrics. For handover University, August 1997, 262 p. between WLAN and GPRS, a general level procedure is

[13] A. Hatami, P. Krishnamurthy, K. Pahlavan, J. Makela,R. Pichna, M. Ylianttila, J. Talvitie, Fuzzy Logic,Neural Network and Other Algorithms for Handoff inWireless Networks, to appear in Proc. of Third International ICSC Symposium on Fuzzy Logic and Applications, Rochester, June 1999. [14] A. Hatami, P. Krishnamurthy, K. Pahlavan, M. Ylianttila, J. Miikela, R. Pichna, Analytical Framework for Handoff in Non-Homogeneous Mobile Data Networks, in Proc. of PIMRCP9, Osaka, Japan, September 1999. Global