tsconit +a+...

of 9/9
UMTS and LTE/SAE handover solutions and their comparison T.F.M. Hendrixen Faculty of Management and Governance University of Twente, the Netherlands [email protected]  ABSTRACT Handovers are procedures used to keep one or more communicating connection(s) of a mobile user alive, even when the user roams from one network access point to another access point. Handovers are therefore, very important for mobility support in wireless networks. This paper focuses on the efficiency of handovers between UMTS and the new 3GPP network technology LTE. The paper first explains the important requirements for efficient handovers. Once the requirements are clear, five possible handover solutions are explained. To verify if all solutions are efficient they are compared to the requirements. KEYWORDS LTE/SAE, UMTS, handover, efficiency 1. INTRODUCTION The demand for mobile internet increases every day. To keep up with user preferences, several types of networks are available and implemented. All these different networks have their own properties such as bandwidth, response time and coverage area. But due to developments in applications and services, larger bandwidths, lower latency, always alive connections, etc., are needed. To accommodate these needs, new better networks are being specified and developed. Nowadays to provide mobile internet Universal Mobile Telecommunications System (UMTS) / High-Speed Downlink Packet Access (HSPDA) technologies are used. At the same time the 3rd Generation Partnership Project (3GPP) is developing a new technology called 3GPP Long Term Evolution (LTE). 3GPP LTE aims to improve third generation (UMTS) technology to meet requirements in a time perspective of 2010 and beyond, see [1] [2] [3]. Some of the agreed requirements/targets of LTE are a significant increase in peak data rates (100 Mbps down and 50 Mbps up); a scalable bandwidth and a reduced latency for the user [4]. Examples of applications that are used by mobile users are video conferencing, email, messaging and even live TV. For most of these applications mobile users desire that their connections are maintained as their devices move from one access point to another. To provide this service handover mechanisms are used between access points [1].  To integrate a new technology like 3GPP LTE with the already available networks, vertical handover mechanisms have to be used. These mechanisms maintain network (internet) connectivity for the mobile user and switch to the best suitable technology available at that moment, see [5]. Because handover mechanisms aren’t a new way of keeping connections alive, several mechanisms have already been researched and developed. Some of these mechanisms are also used between UMTS and LTE, but are these proposed handovers efficient? Is there a better way to provide handovers between UMTS and LTE? This brings us to the main research question of this paper: - Are the handover solutions supported in UMTS and LTE/SAE efficient? To answer this main research question, sub questions are derived: 1. How do handover mechanisms work? 2. What are requirements for an efficient handover? 3. What handover solutions are used between LTE/SAE and UMTS? 4. Do the handover solutions comply with the requirements for handover efficiency? A literature study and a qualitative comparison analysis is used in order to answer these research questions. The paper is divided into sections, where each section answers a research question. In section 2 available handover concepts are explained. In section 3 the requirements for efficient handovers are discussed. Section 4 treats the solutions that might be or are used between UMTS and LTE. In section 5 the solutions are compared with the derived requirements. Finally, the conclusions and the possible future work activities are presented. 2. HANDOVER MECHANISMS As explained earlier, handovers are used to keep mobile clients connected to their service network, even when these clients roam from a network access point to another network access point. To supply this service there are two types of handovers, horizontal and vertical. For example: a mobile user is having a video phone call, which uses UMTS, while traveling by train. During the trip, the train moves out of the range of the currently used UMTS network access point, i.e., UMTS cell coverage area. To maintain the phone call, the phone switches from one UMTS network access point to another UMTS network access point, with a better signal quality. The type of handover used in this case is called a horizontal handover, since the handover occurs and is supported by the same wireless technology, i.e., UMTS. When the train reaches the next train station, the phone picks up a WLAN (Wireless Local Access Network) access point. The WLAN access point supports a much greater bandwidth which might provide a much better quality video call. At this point the mobile client switches to a network access Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citati on on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission. 11 th Twente Student Conference on IT, Enschede, June 29 th  , 2009 Copyright 2009, University of Twente, Faculty of Electrical Engineeri ng, Mathematics and Computer Science

Post on 06-Apr-2018

218 views

Category:

Documents

0 download

Embed Size (px)

TRANSCRIPT

  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 1/9

    UMTS and LTE/SAE handover solutions and their comparisonT.F.M. Hendrixen

    Faculty of Management and Governance

    University of Twente, the Netherlands

    [email protected] are procedures used to keep one or more

    communicating connection(s) of a mobile user alive, even when

    the user roams from one network access point to another access

    point. Handovers are therefore, very important for mobility

    support in wireless networks. This paper focuses on the

    efficiency of handovers between UMTS and the new 3GPP

    network technology LTE. The paper first explains the important

    requirements for efficient handovers. Once the requirements are

    clear, five possible handover solutions are explained. To verify

    if all solutions are efficient they are compared to the

    requirements.

    KEYWORDS

    LTE/SAE, UMTS, handover, efficiency

    1. INTRODUCTIONThe demand for mobile internet increases every day. To keep up

    with user preferences, several types of networks are available

    and implemented. All these different networks have their own

    properties such as bandwidth, response time and coverage area.

    But due to developments in applications and services, larger

    bandwidths, lower latency, always alive connections, etc., are

    needed. To accommodate these needs, new better networks are

    being specified and developed.Nowadays to provide mobile internet Universal Mobile

    Telecommunications System (UMTS) / High-Speed Downlink

    Packet Access (HSPDA) technologies are used. At the same

    time the 3rd Generation Partnership Project (3GPP) is

    developing a new technology called 3GPP Long Term

    Evolution (LTE). 3GPP LTE aims to improve third generation

    (UMTS) technology to meet requirements in a time perspective

    of 2010 and beyond, see [1] [2] [3]. Some of the agreed

    requirements/targets of LTE are a significant increase in peak

    data rates (100 Mbps down and 50 Mbps up); a scalable

    bandwidth and a reduced latency for the user [4].

    Examples of applications that are used by mobile users are

    video conferencing, email, messaging and even live TV. For

    most of these applications mobile users desire that their

    connections are maintained as their devices move from one

    access point to another. To provide this service handover

    mechanisms are used between access points [1].

    To integrate a new technology like 3GPP LTE with the already

    available networks, vertical handover mechanisms have to be

    used. These mechanisms maintain network (internet)

    connectivity for the mobile user and switch to the best suitable

    technology available at that moment, see [5]. Because handover

    mechanisms arent a new way of keeping connections alive,

    several mechanisms have already been researched and

    developed. Some of these mechanisms are also used between

    UMTS and LTE, but are these proposed handovers efficient? Is

    there a better way to provide handovers between UMTS and

    LTE? This brings us to the main research question of this paper:

    - Are the handover solutions supported in UMTS andLTE/SAE efficient?

    To answer this main research question, sub questions arederived:

    1. How do handover mechanisms work?2. What are requirements for an efficient handover?3. What handover solutions are used between LTE/SAE

    and UMTS?

    4. Do the handover solutions comply with therequirements for handover efficiency?

    A literature study and a qualitative comparison analysis is used

    in order to answer these research questions.

    The paper is divided into sections, where each section answers a

    research question. In section 2 available handover concepts are

    explained. In section 3 the requirements for efficient handovers

    are discussed. Section 4 treats the solutions that might be or are

    used between UMTS and LTE. In section 5 the solutions are

    compared with the derived requirements. Finally, the

    conclusions and the possible future work activities are

    presented.

    2. HANDOVER MECHANISMSAs explained earlier, handovers are used to keep mobile clients

    connected to their service network, even when these clientsroam from a network access point to another network access

    point. To supply this service there are two types of handovers,

    horizontal and vertical. For example: a mobile user is having a

    video phone call, which uses UMTS, while traveling by train.

    During the trip, the train moves out of the range of the currently

    used UMTS network access point, i.e., UMTS cell coverage

    area. To maintain the phone call, the phone switches from one

    UMTS network access point to another UMTS network access

    point, with a better signal quality. The type of handover used in

    this case is called a horizontal handover, since the handover

    occurs and is supported by the same wireless technology, i.e.,

    UMTS. When the train reaches the next train station, the phone

    picks up a WLAN (Wireless Local Access Network) access

    point. The WLAN access point supports a much greaterbandwidth which might provide a much better quality video

    call. At this point the mobile client switches to a network access

    Permission to make digital or hard copies of all or part of this work for personal or

    classroom use is granted without fee provided that copies are not made or

    distributed for profit or commercial advantage and that copies bear this notice and

    the full citation on the first page. To copy otherwise, or republish, to post on servers

    or to redistribute to lists, requires prior specific permission.

    11th

    Twente Student Conference on IT, Enschede, June 29th

    , 2009Copyright 2009, University of Twente, Faculty of Electrical Engineering,

    Mathematics and Computer Science

  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 2/9

    point supported by another wireless technology. This process is

    called a vertical handover. [5] [6]

    The central goal of the handover mechanisms is to maximize

    overall network utilization and allow each client to remain best

    connected at all times [7]. Best connected is specified as

    having the optimal connection for the service the client is using.

    A handover process can typically be divided into five parts:measurements, processing, reporting, decision and execution. A

    general description of the handover procedure is given below,

    see [8].

    To make sure that the connection stays maintained, a handover

    mechanism constantly scans the air for other access points and

    monitors the active connection. Once the service criteria arent

    fully supported anymore, or when a better suitable network is

    found, then a handover procedure is started. The next section of

    this paper explains this decision in detail, see [5].

    To give a more detailed view of a handover, a high level GSM

    handover example is given: In the first step of the handover

    procedure the client connects and registers with the new access

    point. Then the connection specific information is transmitted

    (reference id, status, etc). Only when the connection is

    successfully taken over, the client disconnects from the old

    access point, see [2] [9].

    With todays new technologies, vertical handovers become

    more popular. For example, a few years ago tri-band mobile

    phones were far from mainstream; nowadays quad-band phones

    with personal area network interfaces are common, and phones

    are even equipped with wireless LAN interfaces. With all these

    types of networks deciding which network(s) to attach to are not

    easy problems when considering all parameters involved: Radio

    resource sharing, multi-operator environment, security, end-to-

    end path optimality and energy efficiency.

    Making the right decisions involves several constraints and

    has to meet several objectives. Failing to satisfy the constraintscauses service interruptions for the mobile node; not meeting

    the objectives leads to, for example, inefficient use of battery

    power and network resources, see [7].

    3. REQUIREMENTS FOR AN EFFICIENTHANDOVER SOLUTIONTo provide an efficient handover there are several requirements

    and criteria. During the literature research many requirements

    and criteria for an efficient handover were found. Some of these

    requirements are listed below.

    3.1 Handover momentOne of the most important criteria for an efficient handover isthe handover moment or handover location. To provide a

    handover without degrading the quality of service (QoS) the

    location of the handover moment should be carefully planned.

    The optimal point for a handover is at a spot where the old and

    the new access point have an overlap in coverage area. Outside

    this area there might be a lot of noise which degrades the

    connection and slows the handover down. When a handover

    moment isnt chosen at the right point this might also lead to

    unnecessary handovers. This leads to the next criterion, see

    [10].

    3.2 Unnecessary handoversAn important requirement for an efficient handover mechanism

    is the amount of unnecessary handovers. The lower the numberof unnecessary handovers the more efficient the handover

    mechanism is.

    To decide the handover moment, handover mechanisms use

    signal strength algorithms to determine the distance to the

    access point. Once the client comes near the border of the

    coverage area the algorithm starts determining the handover

    point. Measurement errors in this process can lead to

    unnecessary handovers, which can lead to higher energy

    consumption and possible degradation of the supported QoS,

    see [8] [11].

    3.3 Handover delayThe duration of the handover procedure is an important

    criterion of the efficiency of a handover mechanism. When a

    handover takes too long, service disruption can be experienced

    or connections can timeout and will be lost. For example, a real

    time video call could experience a temporary disruption when a

    handover takes longer than 400 ms. If the delay is even longer,

    the call could be terminated entirely.

    To provide a seamless and efficient handover, this delay should

    be as short as possible. The delay is measured from the

    execution of the handover algorithm until the algorithm

    completes the handover procedure and the client is successfullyconnected to the other access point, see [5] [12].

    The International Telecommunication Union (ITU) has

    specified a handover delay QoS parameter for IP transport of

    telecommunication applications in [13]. The value of the

    required handover delay parameter changes for different

    application classes. This paper focuses on seamless handovers

    and for that reason only classes for real-time applications are

    used. The classes used are 0 and 1. Class 0 requires a delay of

    less than 100 ms and class 1 requires a delay of less than 400

    ms.

    3.4 Packet lossAn optimal handover mechanism provides handover without

    packet losses. No packet or/-data loss is almost impossible sothe less packet loss a mechanism generates, the more efficient

    the mechanism is. [12] The ITU has also got a QoS parameter

    for this and states that the probability for a packet loss shouldnt

    be more than 1 103, see [13].

    3.5 Scalability: how scalable is a handoversolution?The scalability of a handover solution is an important factor

    when implementing it into wireless network technologies with a

    high density of mobile users. When the number of supported

    users significantly increases will the handover mechanism still

    be efficient? Often a high density of mobile users generates

    errors.

    3.6 Complexity of the solutionThe complexity of a handover procedure is very important when

    talking about mobile devices. These devices are often battery

    powered and have limited resources. If the handover mechanism

    takes too many resources, smaller mobile devices cant use the

    handover.

    3.7 Modifications of the protocol standardsA solution which can be placed in the already available network

    architecture without changing the standard (signaling)

    procedures and protocols makes the solution easier to be

    deployed.

  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 3/9

    4. HANDOVER SOLUTIONSTo make a comparison of handovers between UMTS and LTE,

    the handovers used in the standards for both technologies are

    explained in the first two sections. After that three other

    possible solutions are explained.

    4.1 UMTS HandoversAt this moment UMTS uses the Universal Terrestrial Radio

    Access Network (UTRAN) for handovers. The architecture of

    UTRAN can be seen in Figure 1. The main component in

    UTRAN is the Radio Network Controller (RNC) which is

    connected to several base stations calledNode Bs. Every Node

    B can support several cells with clients connected. Within

    UTRAN, UMTS uses Wideband Code Division Multiple

    Access technology (WCDMA) to carry radio transmissions.

    WCDMA is emerged as the most widely adopted third

    generation air interface. Its specification has been created in the

    3rd Generation Partnership Project (3GPP) WCDMA provides

    three different types of handovers which will be explained

    below, see [14].

    RNS

    RNC

    RNS

    RNC

    Core Network

    Node B Node B Node B Node B

    Iu Iu

    Iur

    Iub IubIub Iub

    UTRAN

    Figure 1: UTRAN Architecture Intra Frequency handovers

    (from [15])

    4.1.1 Intra Frequency handoversIntra Frequency handovers are horizontal handovers between

    two access points on the same frequency. The handover

    initiation is started by an algorithm which measures the signal

    strength continuously. When the algorithm finds a stronger cell,

    the cell is added to its active set, or else into the neighbor list.

    First, the RNC queries the client for its active set. After the

    RNC receives the set, it chooses the best cell to switch to and

    orders the client to handover. These handovers are the most

    common ones and are performed to provide roaming and

    mobility support to the users, see [14].

    4.1.2Inter Frequency Handovers within WCDMAInter Frequency handovers within WCDMA are handovers

    between different frequencies in the same network. These

    handovers are used to provide a higher capacity on the cell.

    The RNC triggers the client to start measuring and identify cells

    and put them on the neighbor list. During the measurements and

    identification the client finds cells with a shared channel. The

    client indentifies the cells as separate single channel cells for

    compatibility and reports the measurements to the RNC. The

    RNC sends the command to connect to the cell with the lowest

    load or best signal, see [14].

    4.1.3 Inter System or Inter RAT HandoversInter System or Inter RAT handovers are vertical handovers

    between different network types i.e. UMTS to LTE. At the start

    of the UMTS deployment, this mechanism was mostly used to

    provide continuous coverage. A graphical representation of the

    handover is given in Figure 2.

    Figure 2: UMTS to LTE Handover

    The handover mechanism is triggered by the RNC. First the

    RNC sends a command to measure and report the signal

    strength from the cells in the vicinity. Between UMTS andLTE, the client performs E-UTRAN measurements which will

    be explained in the next sub section. Once those measurements

    are received by the RNC, it evaluates the measurement and

    decides if and which cell to switch to. Once the decision has

    been made, the RNC sends the command to perform the

    handover, see [15]. If the connection to the LTE network is

    successful, the Serving GPRS support node (SGSN) will

    forward packets from existing connections through the Serving

    GateWay (S-GW) towards the client and vice versa.

    4.2 LTE HandoversLTE uses Evolved UMTS Terrestrial Radio Access network

    (EUTRAN) for handovers, see [16]. EUTRAN is the

    evolvement of UTRAN which was developed as a multipleaccess method with a functional split between the radio access

    and core network in the network architecture, see Figure 3. Due

    to this split all radio functionality (RLC/MAC) is placed in the

    base station, also called eNodeB (an evolved UTRAN Node B).

    This means that the eNodeB is responsible for decisions of

    horizontal handovers. Due to the use of eNodeB with RNC

    capabilities there is no need for a separate RNC and thus

    handovers between two cells on the same RNC arent needed

    anymore which makes EUTRAN a flat network architecture, see

    [14].

    Figure 3 LTE/SAE Architecture (from [15])

  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 4/9

    4.2.1 LTE Intra EUTRAN HandoverBecause of the flat architecture, LTE has only one horizontal

    handover type called the intra EUTRAN handover. This

    handover is triggered by the eNodeB. The eNodeB makes the

    handover decision on signal measurements made by the client.

    Once the source eNodeB decides to initiate the execution of the

    handover, it requests a preparation on the target eNodeB. The

    target eNodeB can perform admission control to check whether

    the established QoS bearers of the client can be accommodated

    in the target cell, see [17].

    Next, the source eNodeB sends the handover command to the

    client, this command includes all information that is necessary

    for the client to access the target cell. At the same time the

    source eNodeB closes the connection with the client and starts

    to forward the data that has not yet been successfully send to the

    client toward the target eNodeB.

    Meanwhile, the client starts to execute the handover and tries to

    connect at the target eNodeB. To perform this connection, the

    client needs approximately 30 ms. This time is needed to

    synchronize with the eNodeB to be able to commence data

    transmission.Once connected, the client sends the handover complete

    message through the target cell towards the target eNodeB. This

    message is used by the target eNodeB to check that the client is

    connected to the right cell. At this point the data connection is

    completed. The target eNodeB reports the successful switch to

    the source eNodeB to release the client, see [3] [14].

    The LTE specification [18] states that due this fast handover

    this mechanism supports seamless mobility support across the

    cellular network speeds up to 500km/h are supported

    (depending on the frequency).

    4.2.2 LTE Inter RAT HandoverTo provide handover compatibility with other 3GPP network

    types i.e. UMTS, LTE uses a component of the SystemArchitecture Evolution (SAE) called Evolved Packet Core

    (EPC). A subcomponent in the EPC called S-GW provides

    mobility for inter eNodeB handovers like EUTRAN to UTRAN.

    Figure 4: LTE to UMTS Handover sequence

    A LTE client constantly reports signal measurements to the

    connected eNodeB. When the eNodeB determines the necessity

    to initiate a handover, the handover preparation procedure is

    initiated. This is done by sending the HANDOVER

    REQUIRED message to the Mobility Management Entity

    (MME) located in the EPC, see Figure 4. When the message is

    received, the MME starts preparation by starting the Resource

    allocation procedure. This procedure is responsible for getting

    resource information from the access points in the vicinity.

    With this information the handover decision is made and the

    MME sends the HANDOVER command with the neededinformation to the eNodeB which sends it towards the client. At

    this point the client disconnects from the source eNodeB and

    connects to the other access point. This other access point, in

    turn will report to the source eNodeB that the handover was

    successful, see [15]. Once connected the connection is

    redirected via the SGSN to the S-GW which is connected to the

    packet data network via the packet data network gateway (P-

    GW). No network connections are lost due to the routing and

    forwarding mechanism in the S-GW, see [19]. A graphicalrepresentation of the handover is given in Figure 5.

    Figure 5: LTE to UMTS handover

    4.3 Non UMTS or LTE standard handovermechanismsThis section provides an overview of other possible solutions

    that could be used to provide a vertical handover between the

    UMTS and LTE wireless technologies.

    4.3.1 Mobile IP (MIP)Mobile IP has two versions, version 4 and version 6, both willbe explained.

    IP version 4 (IPv4) assumes that a node's IP address uniquely

    identifies the node's point of attachment to the Internet. To

    support this, a node must be located on the network indicated

    by its IP address in order to receive data destined to it;

    otherwise, data to the node would be undeliverable. If a node

    wants to change its point of attachment in IPv4 it always loses

    its connection (i.e.by IP address change). To support this

    handover mobile IP defines such a mechanism, which enables

    nodes to change their point of attachment to the Internet

    without changing their permanent IP address, see [20].

    However, when the mobile node changes its network point of

    attachment, then this node gets via Mobile IP a temporarily IP

    address called Care of Address (COA). In Mobile IP thepermanent IP address is called Home Address. During the

    period in which the mobile node remains attached to the

    network via the same network point of attachment, a binding

    between the COA and the Home Address is maintained by the

    Mobile IP functionality.

    Mobile IP uses three basic functional entities: the mobile node

    (MN), the Home Agent (HA) and the Foreign Agent (FA) , see

    Figure 6. When a MN detects that it has moved to a foreign

    network, it obtains a care-of address (COA) on the foreign

    network. This address identifies the MN in the other network.

    Once the MN has its COA, the MN registers it with its HA

    trough the currently connected FA. At this point, the COA and

    the home address are bound together. When now data is sent

    towards the HA of the MN, the data is tunneled towards theMNs COA and finally to the MN itself, this is also called

  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 5/9

    triangular routing [5]. The MN itself sends packets directly to

    the host. In this way the permanent IP address, i.e., Home

    Address, of the MN stays the same, so TCP connections can be

    kept alive, see [20].

    Figure 6: Mobile IP routing (from [20])

    In IP version 6 (IPv6) mobility support shares many features

    with mobility support in IPv4. But IPv6 benefits from the

    experiences gained from IPv4 and so IPv6 has some major

    improvements. One of the most important and relevant

    improvement is the route optimization. Mobility support inIPv6 sends its COA towards the host to inform its IP address

    change. This action optimizes the route of the data flow,

    because now the data doesnt need to be tunneled anymore

    through the HA, see [21].

    4.3.2 Media Independent Handover (MIH)MIH is a handover mechanism that is standardized by IEEE and

    is specified in IEEE 802.21. MIH is a handover mechanism that

    focuses on service continuity and provides interworking

    between any wireless network access technology, i.e. IEEE 802

    systems (e.g., IEEE 802.11 and IEEE 802.16e), but also

    between IEEE 802 and non-IEEE 802 systems (e.g., cellular

    networks like 3G).

    Figure 7: MIH layer (from[6])

    MIH is a framework which is positioned between the IP layer

    and the network specific link layer. As you can see in Figure 7

    the layer has three communication lines to relay information

    between the two layers, see [6]. To explain this in more detail

    the following scenario is described, see Figure 8.

    Figure 8: MIH Scenario

    In the scenario presented in Figure 8, the 3G operator and the

    LTE operator have a roaming relationship and both core

    networks contain MIH entities. While a mobile node is within

    the UMTS network, it can query the information server to

    obtain available LTE network information. This can be done

    without activating and scanning through the LTE interface. Not

    scanning can save battery power significantly. Using the

    information provided by the information server, the mobile

    node can activate its LTE interface and have the guarantee that

    there is an available network. Then, the node can connect to the

    LTE network while the session is still active through the UMTS

    interface. When successfully connected, the node can use MIH

    commands to handover. The use of the MIH services allows

    much of the time-consuming preparation work to be done

    before the handover takes place which significantly reduces

    handover latency and packet losses, see [22].

    4.3.3 Session Initiation Protocol (SIP)SIP is an application-layer control protocol that can establish,

    modify, and terminate multimedia sessions. SIP can also invite

    participants to already existing sessions, such as multicast

    conferences. It transparently supports name mapping andredirection services, which supports personal mobility. This

    means that users can maintain a single externally visible

    identifier regardless of their network location, see [23].

    As stated in [23], SIP supports five facets of establishing and

    terminating multimedia communications:

    - User location: determination of the end system to be usedfor communication;

    - User availability: determination of the willingness of thecalled party to engage in communications;

    - User capabilities: determination of the media and mediaparameters to be used;

    -

    Session setup: "ringing", establishment of sessionparameters at both called and calling party;

    - Session management: including transfer and termination ofsessions, modifying session parameters, and invoking

    services.

    With these functionalities, clients can always be found after a

    handover and connections can be re-established and maintained.

    SIP doesnt provide services, but provides primitives that can

    be used to implement services. In particular, SIP can locate a

    client and send a data object to it, the data object can contain

    everything e.g. service related data. [23] also states that IPv4

    and IPv6 are supported.

  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 6/9

    Figure 9: SIP Handover

    A SIP session, see Figure 9, is started with SIP INVITE

    message to the universal resource indicator (URI) which is

    associated with the user device. This URI has a fixed address in

    the format sip:[email protected] The message is sent trough a

    SIP proxy server that registers the IP address of the user devices

    and forwards the message. After that, the communication can

    take place via the Real Time Protocol (RTP). When a client

    changes from network architecture, the client has to re-INVITE

    to the corresponding node and register at the proxy server. Once

    this is done, both nodes can communicate again because the

    proxy server knows where to find them, see [24].

    5. COMPARISON AND ANALYSIS OFHANDOVER SOLUTIONSIn this section the vertical handover solutions explained earlier

    will be compared with the requirements found in section 3.

    Every solution is explained separately except the Inter RAT

    Handovers from UMTS and LTE, which are treated together.

    The reason why they are treated together is because both

    handovers principles are very similar. First a table, see Figure

    10, which summarizes the solutions compared to the

    requirements is shown. The properties good, fair and bad are

    interpreted for each requirement in the following way:

    Handover Momento Good: handover on the perfect moment, without

    service degradation;

    o Fair: handover on the moment where both cells arestill in range, but with a little service degradation

    (i.e. higher packet loss);

    o Bad: handover too early or too late, with highservice degradation or even service loss;

    o Fair*: the handover moment quality for the MIPbase protocol could not be easily defined, but it is

    important to emphasize that the Internet Engineering

    Task Force (IETF) has specified and is currently

    specifying several handover enhancements that can

    be applied in combination with the MIP baseprotocol, such as [25].

    Unnecessary Handoverso Good: no unnecessary handovers;o Fair: some unnecessary handovers but without

    service degradation;

    o Bad: too much unnecessary handovers which causesservice degradation;

    o Fair*: the unnecessary handovers quality for theMIP base protocol could not be easily defined, but it

    is important to emphasize that the IETF has

    specified and is currently specifying several

    handover enhancements that can be applied in

    combination with the MIP base protocol, such as

    [25];

    Handover Delayo Good: handover delay is below 100 ms (class 0 of

    ITU);

    o Fair: handover delay is below 400 ms (class 1 ofITU);

    o Bad: handover delay is over 400 ms (class 1 of ITU);o Fair*: the handover delay for the MIP base protocol,

    is Bad, but it is important to emphasize that the

    IETF has specified and is currently specifying

    several handover enhancements that can be applied

    in combination with the MIP base protocol, such as

    [25].

    Packet Losso Good: no packets are lost during the handover;o Fair: some packets may be lost during the transfer,

    but below the threshold of 1 103;

    o Bad: packet loss is higher than the threshold of1 103.

    Scalabilityo Good: very good usable in environments with both

    high and low density of users;

    o Fair: moderate usable in environments with bothhigh and low density of users;

    o Bad: not usable in environments with both high andlow density of users.

    Complexityo Good: Just a couple of simple procedures and

    resources needed which can be done by any device

    and makes it easy to implement;

    o Fair: Some more procedures and resources areneeded but can still be done by almost any device

    and makes it moderate to implement;

    o Bad: a lot of procedures and resources needed, hardfor most mobile devices and hard to implement.

    Modifications to the protocolo Good: very easy to implement without modifications

    to the protocols used;

    o Fair: moderate to implement, some changes have tobe made to the protocols used;

    o Bad: hard to implement, a lot of changes have to bemade to the protocols used.

  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 7/9

    Handover

    moment

    Unnecessary

    Handovers

    Handover

    delay

    Packetloss

    Scalability

    Complexity

    Modification

    s

    toprotocol

    Inter

    RAT

    Good Good Fair Good Good Fair Good

    MIP Fair* Fair* Fair* Fair Fair Fair Fair

    MIH Good Good Fair Fair Fair Bad Fair

    SIP - - Fair Fair Fair Good FairFigure 10: Comparison of Handovers

    5.1 Inter RAT HandoverDue to the fact that the handover moment of an inter RAT

    handover is decided by measurements done by the client which

    reports them to the source NodeB or eNodeB, the handovermoment can be chosen at the ideal time and location. This also

    means that if the decisions made on the measurements are done

    right, there is almost no chance of unnecessary handovers. This

    method has also a disadvantage, using the client to measure

    signal strength will take some processing power and consumes

    battery.

    As stated in [18] the handover delay during a EUTRAN to

    UTRAN handover doesnt take longer than 300 ms for real-time

    quality services. This is below the ITU threshold parameter for

    real time services and thus will go unnoticed by the user. LTE

    and UMTS support specific features to increase reliability and

    packet loss. Therefore, it could be assumed that the packet

    loss criterion can be graded with Good.

    The provided handover solution needs sophisticated procedures,

    but compared to the large number of features and procedures

    supported by LTE and UMTS, this complexity can be

    considered as non-significant. The flat architecture provided

    due the split has also got its advantages. The biggest advantage

    is that it makes the solution highly scalable, see [14].

    5.2 Mobile IP (MIP)Due to the use of triangular routing, the QoS of real-time

    communications can be degraded. This is because the routing

    adds extra hops which can increase packet delay. In addition the

    redundant routing in MIP results in traffic increase on the home

    agent, see [26]. MIP handovers can take up to seconds, see

    [27]. Due to these high handover delays, there is a high chance

    that the handover will be noticed by the user. Long packet

    routes and high packet delays create a bigger probability of

    packet losses, see [26]. Due to the fact that each home domain

    can be managed by a different home agent, makes the MIP

    solution fairly scalable. A scalability drawback however, is that

    the capacity of a home agent can limit the number of mobile

    users that can be supported in a home domain. Servers could

    easily be placed in the core network to support MIP.

    In order to use MIP for UMTS LTE handovers, some minor

    modifications on the standards might be needed.

    A MIP handover between UMTS and LTE is initiated when the

    MN detects that it has moved to a foreign network, see [20].

    Because of this, MIP cannot control the handover moment

    which means it is not responsible for unnecessary handovers.

    Note however, that the Internet Engineering Task Force (IETF)

    has specified and is currently specifying several handover

    enhancements that can be applied in combination with the MIP

    base protocol, such as [25].

    5.3 Media Independent Handover (MIH)MIH is dependent on an extra layer in the network stack. The

    use of an extra layer between the link layer and the IP layer

    makes the MIH solution a complex solution. The use of an

    extra layer has also, a significant advantage: The scalability ofMIH depends on how it will be integrated within the LTE and

    UMTS network. Moreover, this also depends on the way of

    how the MIHF entities, see Figure 8, are distributed and used.

    Due to the use of an additional protocol layer, between the IP

    layer and the different Link layers, modifications on the

    available standards might be needed.

    Due to the fact that MIH uses a make-before-break handover

    procedure, the information about the target network can be

    gathered before the original connection is lost. This

    significantly improves the calculation accuracy of the handover

    moment and decreases the number of unnecessary handovers.

    The delay of MIH handovers is dependent on the used protocol

    in the IP layer. When implemented with Mobile IP, MIH

    decreases the delay to approximately 200ms which also results

    in low packet losses, see [6].

    5.4 Session Initiation Protocol (SIP)According to [24], a SIP based handover performed will take

    approximately 227 ms, in average. This time, depends of course

    on the used network scenario, and is mainly consumed by the

    reconnection, re-invite and termination + setup of the RTP.

    With this procedure, the proxy SIP server already knows the

    ends of the connection and with this information the connection

    can be re-established quickly, which shortens the handover

    delay significantly. A short handover delay means a low packet

    loss.

    Deployment of a SIP based handover solution is not complex.Due to the fact that SIP is an application layer handover

    mechanism, it is easy to implement it on different platforms.

    SIP is using a client server based architecture. This means that

    the number of users that will be supported by the SIP based

    handover solution will depend on the capacity and on the way

    of how the SIP servers are distributed within the network.

    A SIP handover between UMTS and LTE is initiated when the

    client changes from subnet, see [24]. SIP is an application layer

    protocol and therefore, it is not able to accurately calculate the

    handover moment and to prevent unnecessary handovers.

    6. CONCLUSION AND FUTURE WORKIn this paper the efficiency of vertical handover mechanismssupported between UMTS and LTE have been researched. Due

    to the fact that handover mechanisms maintain network

    connections over different wireless technologies and network

    architectures they are one of the most important mechanisms for

    the support of the mobility of a user.

    In this paper several handover solutions where compared with

    the requirements for handover efficiency. These handover

    efficiency requirements were found in the papers that studied

    and analyzed handover mechanisms such as [3]. The most

    important requirements where: handover moment, handover

    delay, packet loss, scalability, complexity and protocol

    modifications.

    The research of this paper first focused on the already usedhandover mechanisms between UMTS and LTE. After that,

    other possible solutions where used to make a comparison and

  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 8/9

    to analyze if there were other mechanisms that might

    outperform the standardized Inter-RAT vertical handover.

    These additional handover solutions are mobile IP, media

    independent handover and session initiation protocol. These

    solutions were chosen because they are often used and

    researched.

    Finally the solutions found were compared to the requirementsfor handover efficiency found earlier. This comparison gave

    some interesting results. As can be seen in Figure 10, for some

    of the criteria, the inter-RAT handover mechanism is

    outperforming other handover mechanisms. None of the studied

    handover mechanisms was found to be more efficient than the

    standard Inter-RAT handover solution. During the research

    several other solutions where identified. Most of these solutions

    were combinations or improved copies of the basic solutions

    described in other papers, see [24][26][28][29][30]. These

    handovers solutions might be very interesting for further

    research. Finally, to test the handover efficiency of the different

    handover solutions in a more accurate way, quantitative

    experiments and analysis maybe needed, which can be

    performed using, for all handover solutions, the same testenvironment and workload parameters.

    ACKNOWLEDGMENTSI would like to thank my supervisor dr. ir. Georgios

    Karagiannis. He helped me with the structure of my research

    and always reviewed my work. I would also like to thank my

    reviewers. They provided useful comments and helped me with

    my English.

    REFERENCES[1] Banerjee, N.; Wei Wu; Das, S.K., Mobility support in

    wireless Internet, Wireless Communications, IEEE,

    vol.10, no.5, pp. 54-61, Oct 2003[2] 3GPP TS 23.009 V8.0.1 (2009-01), Handover

    procedures, January 2009

    [3] Racz, A.; Temesvary, A.; Reider, N., HandoverPerformance in 3GPP Long Term Evolution (LTE)

    Systems, Mobile and Wireless Communications Summit,

    2007. 16th IST, pp.1-5, July 2007

    [4] Bachl, R., Gunreben, P., Das, S., & Tatesh, S., The longterm evolution towards a new 3GPP* air interface

    standard.Bell Labs Technical Journal, vol.11, no.4, pp.

    25-51, March 2007

    [5] Taylor, L.; Titmuss, R.; Lebre, C., The challenges ofseamless handover in future mobile multimedia networks,

    Personal Communications, IEEE , vol.6, no.2, pp.32-37,April 1999

    [6] Wright, D.J., Maintaining QoS During Handover AmongMultiple Wireless Access Technologies, Management of

    Mobile Business, 2007. ICMB 2007. International

    Conference on the , pp.10-10, July 2007

    [7] Hollos, D.; Poyhonen, P.; Strandberg, O.; Aguero, R.;Pentikousis, K.; Blume, O., A Study of Handover

    Strategies for Mobile Multiaccess Ambient Networks,

    Mobile and Wireless Communications Summit, 2007. 16th

    IST, pp.1-5, July 2007

    [8] Anas, M.; Calabrese, F.D.; Ostling, P.-E.; Pedersen, K.I.;Mogensen, P.E., Performance Analysis of Handover

    Measurements and Layer 3 Filtering for Utran LTE,Personal, Indoor and Mobile Radio Communications,

    2007. PIMRC 2007. IEEE 18th International Symposium

    on , pp.1-5, September 2007

    [9] 3GPP TS 48.008 V8.5.0 (2008-11), Mobile SwitchingCentre - Base Station System (MSC-BSS) interface; Layer

    3 specification, November 2008

    [10]Christophorou C.; Pitsillides A., MBMS Handovercontrol: A new approach for efficient handover in MBMSenabled 3G cellular networks, Computer Networks, vol.

    51, no. 18, pp. 4897-4918, December 2007

    [11]Christophorou, C.; Pitsillides, A., An Efficient HandoverAlgorithm for MBMS Enabled 3G Mobile Cellular

    Networks,Computers and Communications, 2006. ISCC

    '06. Proceedings. 11th IEEE Symposium on , pp. 187-193,

    June 2006

    [12]Kim H.; Yeom H., An efficient multicast mechanism fordata loss prevention, Advanced Communication

    Technology, 2005, ICACT 2005. The 7th International

    Conference on, vol.1, pp. 497-502, February 2005

    [13]ITU-T Y.1541 Network Performance Objectives forIPBased Services, May 2002

    [14]Holma H.; Toskala A., WCDMA for UMTS: HSPAEvolution and LTE,4th edition, John Wiley & Sons,

    2007.

    [15]Hammer, M.; Salkintzis, A.; Tanaka, I.; Wong, C., Voicecall handover mechanisms in next-generation 3GPP

    systems,Communications Magazine, IEEE, vol.47, no.2,

    pp.46-56, February 2009

    [16]3GPP TS 36.300 V8.7.0 (2008-12) , Evolved UniversalTerrestrial Radio Access (E-UTRA) and Evolved

    Universal Terrestrial Radio Access Network (E-

    UTRAN), December 2008

    [17]Amirijoo, M.; Frenger, P.; Gunnarsson, F.; Kallin, H.;Moe, J.; Zetterberg, K., Neighbor cell relation list and

    measured cell identity management in LTE, Network

    Operations and Management Symposium, 2008. NOMS

    2008. IEEE, pp.152-159, April 2008

    [18]UTRA-UTRAN Long Term Evolution (LTE) and 3GPPSystem Architecture Evolution (SAE)

    ftp://ftp.3gpp.org/Inbox/2008_web_files/LTA_Paper.pdf,

    access date: 22 April 2009

    [19]Salkintzis, A.; Hammer, M.; Tanaka, I.; Wong, C., "Voicecall handover mechanisms in next-generation 3GPP

    systems," Communications Magazine, IEEE, vol.47, no.2,

    pp.46-56, February 2009

    [20]RFC3344, c. Perkins, IP mobility support for IP4 nodes,August 2002

    [21]RFC3775, D Johnson, C. Perkins, J. Arkko, MobilitySupport in IPv6, June 2004

    [22]Taniuchi, K.; Ohba, Y.; Fajardo, V.; Das, S.; Tauil, M.;Yuu-Heng Cheng; Dutta, A.; Baker, D.; Yajnik, M.;

    Famolari, D., IEEE 802.21: Media independent handover:

    Features, applicability, and realization,Communications

    Magazine, IEEE, vol.47, no.1, pp.112-120, January 2009

    [23]RFC3261, J Rosenberg, et. Al., Mobility SIP: SessionInitiation Protocol, June 2002

    [24]Jover, F.; Reid, G.; Jover, X., A faster HandoverMechanism using SIP,Mobile and Wireless

    Communications Summit, 2007. 16th IST, pp.1-5, July2007

    ftp://ftp.3gpp.org/Inbox/2008_web_files/LTA_Paper.pdfftp://ftp.3gpp.org/Inbox/2008_web_files/LTA_Paper.pdfftp://ftp.3gpp.org/Inbox/2008_web_files/LTA_Paper.pdf
  • 8/3/2019 TSConIT +a+ +Hendrixen%2C+T.F.M.%2C+UMTS+and+LTE%2FSAE+Handover+Solutions+and+Their+Comparis

    http:///reader/full/tsconit-a-hendrixen2ctfm2cumtsandlte2fsaehandoversolutionsandtheircomparison2cuniversityoft 9/9

    [25]RFC5268, R. Koodli, Ed. , Mobile IPv6 Fast Handovers,June 2008

    [26]Seta, N.; Miyajima, H.; Zhang, L.; Hayashi, H.; Fujii, T.,All-SIP Mobility: Session Continuity on Handover in

    Heterogeneous Access Environment, Vehicular

    Technology Conference, 2007. VTC2007-Spring. IEEE

    65th , pp.1121-1126, April 2007[27]F. Zhu and J. McNair, Cross layer design for Mobile IP

    handoff, inProc. IEEE Vehicular Technology Conference

    (VTC 2005-Spring), vol. 4, pp. 22552259, March 2005

    [28]Hwang Y.; Park A., Vertical Handover Platform overApplying the Open API for WLAN and 3G LTE Systems,

    Vehicular Technology Conference, 2008. VTC 2008-Fall.

    IEEE 68th , pp.1-5, September 2008

    [29]Kwon D.; Kim W.; Suh Y., An efficient mobile multicastmechanism for fast handovers: A study from design and

    implementation in experimental networks, Computer

    Communications, vol. 31, no.10, pp. 2162-2177, June

    2008[30]Yoo S.; Shin S., Fast Handover Mechanism For Seamless

    Multicasting Services in Mobile IPv6 Wireless Networks,

    Wireless Personal Communications, 2007, vol. 42, no. 4,

    pp. 509-526, September 2007