Fast Handover In Wireless Communication Systems Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

In modern day communication network it is important for a wireless network to perform well for real time communications applications like VoIP or video. Different type of wireless network systems have been developed and integrated to offer the internet access to the end users. So the main issue is to manage the mobile nodes while moving across the networks with session continuity and minimum handover latency. MIPv6 has been proposed by the IETF to support this mobility, MIPv6 provide the solution but problem is its inability to support the seamless handover because it gives long latency and packet loss. So to reduce the layer 3 latency FMIPv6 has been proposed by using the address pre-configuration method with the layer 2 triggers. Current research gives the layer 2 trigger procedure for seamless handover but does not provide the exact timing and proper criteria of L2 trigger. This report discusses the exact timing and criteria of layer 2 trigger to reduce the handover latency and packet loss. Two modes of FMIPv6 have been discussed reactive and predictive. This discussion leads predictive mode of FMIPv6 has a better performance with handover in ipv6 this has been showed by the experiment conducted and results derived from it.

In WiFi networks each mobile node is attached with an access point (AP) which provides the access to the fixed network infrastructure. Each AP covers a limited area. But when mobile node moves it may need to change the AP. So the control of this mobile node is handed over to other AP. This process of changing from one AP to other AP is called handover. In this process all the control goes under the other AP.

In last few years, wireless Local Area Network (WLAN) emerged as a very competitive technology with wide bandwidth and low cost. Users of wireless technology are growing day by day because it is easy to install and provide mobility. Now the main issue is handover between APs to maintain the mobility. Especially during the real time communication such as VoIP, though problem is not only handover, actual problem is to provide handover techniques which gives minimum delay during the time of switching from one AP to other AP without disrupting the real time multimedia services.

To solve this problem of handover many techniques in [1]-[5] have been proposed by designing new algorithms or developing new network protocols. They proposed different techniques by using network layer (L3), L2 and physical layer. To manage the movement of mobile nodes between wireless IPv6 network, Mobile IPv6 [5], is designed. This Mobile IPv6 provides unbroken connectivity while mobile node moves from one wireless point of attachment to another point in different wireless network. To notify the position/location of mobile node IPv6 set up a message exchange with the correspondents by binding between the addresses. Mobile IPv6 provides a handover solution but it gives delay during the process of handover while the Fast Mobile IPv6 gives less delay with out packet loss. In this report we will discuss the different approaches used for the handover while switching from one AP to other AP in different wireless subset. We will compare them and then suggest the best which gives fast and secure handoff with low latency.

1 Handover Latency

In MIPV6 the time interval when the mobile node cannot send or receive any kind of traffic during handover is called as handover latency. This is classified into two kinds L2-handover latency and L3-handover latency.

L2 Handover Latency

L2 latency is the time period when the air-link with current AR is disconnected and also the mobile node is not connected to the air-link of new AR which it wants to get attached. According to this scenario the whole process of handover can be divided into three phases




In scanning of new AP either mobile node sends a ProbeRequest (active scanning) message or the AP send a Beacon messages (passive scanning). After scanning the all channel one AP is selected by the mobile node by using the received signal Strength Indicator (RSSI), Signal to Interference Ratio, link quality etc. then the mobile node exchanges IEEE 802.11 authentication message with the selected AP. If the mobile node is authenticated by the AP then node sends it the Reassociation Request message. In this phase new AP and old AP exchange the information with the help of Inter-Access Point Protocol (IAPP) [6]. This delay of scanning, authentication and re-association is called L2 delay.

L3 Handover Latency

L3 latency is the sum of total time of movement detection latency, Binding Update latency and acquisition of new CoA latency. This will add up to overall latency and lot of packets can be lost.

Solution to L2 Handover Latency

From the study of previous work we learned that channel scanning is the most time consuming phase in handover as in [19] more than 90% delay is the cause of scanning. Previous most of the work is done to reduce the L2 delay [10,11] such as in [8] a new scheme Neighboring Graph [7] was introduced. Same as to reduce the latency in [12] a new scheme Backhaul-Aided Seamless Handoff (BASH ) was introduce. In this scheme a Mobile Node do not need to scan all the channels during the handoff, it only send a broadcast probe request message to the neighboring Mesh Routers in Backhaul. Qualified mesh router responses the old mesh router of Mobile Node, directing to attach with the new mesh router. These were the some schemes to reduce the Layer 2 handover latency.

Concept of Mobile IP

We would like to discuss the concept of Mobile IP Protocol which is one solution to provide continual connectivity to the mobile users regardless of their physical movement developed by IETF. There are 3 functional entities in Mobile IP namely Home Agent (HA), Foreign Agent (FA) and Mobile Node (MN). Every mobile node has 2 addresses one is permanent home address (HoA) which is assigned in its home network and second is the temporary care-of address (CoA) which it obtain temporarily in the network in which it move. The Mobility agent (FA) in the visited network provides CoA to the mobile node which can be either unique or shared. When MN moves to a new network it register its CoA with its Home Agent (HA) which keeps the binding of HoA with CoA. All the packets meant for the Mobile node are received by HoA first where HA intercept all the received packets and tunnel them to the MN via FA. It is not a satisfactory solution for mobile users having the large distance between the visited network and home network (cause of signaling delay) because it require the registration of visited network before start sending the packets and due to long distance the long handoff delay and packet loss occurs. Fast Handovers for Mobile (FMIP) and Hierarchical Mobile (HMIP) is used to overcome this problem in Mobile IP

Approaches of Handoff

There are different approaches or techniques for seamless handoff, the two important issues to be considered for the seamless handoff are i.e. the detection of network condition for handoff and maintenance of the connection. In [13] hey introduce Connection Manager (CM) and Virtual Connectivity (VC). CM QoS based handoff metric by using FFT (Fast Fourier Transform) Signal decay detection scheme which helps to make accurate handoff decisions when moving from WLAN to 3G and available bandwidth and access delay measured by MAC layer are used when moving from 3G to WLAN to reduces the probability of wrong or undesired handoffs, whereas VC use Local Connection Translation(LCT) to maintains the mapping between the original connection and current connection information and Subscription/Notification (S/N) service to support mobility.

One Handoff approach is based on the Link layer trigger [13] by using link layer information and performing network layer handoff in advance. In this scheme it’s the responsibility of Mobile node to perform Link layer Find Network layer handoff simultaneously and that reduce the handoff latency. It has one drawback that it requires exact handoff initiation for both layers. One Handoff approach is named Handoff Approach based on Dual Link (HADL) [14] they use link layer information while moving from one subnet to other for impending handoff and start handoff process which is based on the dual link, in the handoff in this approach the mobile node maintain one link with the old network while connecting to new network. The important thing in the process of handoff which should be considered is the selection on the network to be switched and this thing is more important when the handoff is between heterogeneous networks, there are various Vertical Handoff decision algorithms for this and these algorithms use handoff metric to for making the decision which network is best to use. In one of the algorithms [15] they assigned cost function to available networks on the basis of bandwidth, delay and power requirement and selection is based on the network with lower cost value. Other decision algorithms are Analytical Hierarchy Process (AHP) [16], Grey Relational Analysis (GRA) and fuzzy Multiple Attribute Decision Making (MADM) [17]. There are two methods proposed by MADM namely Simple Additive Weighting (SAW) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). We will discuss MIPv6 and FMIPv6 with two different mode of reactive and predictive.


The solution provided by Internet Engineering Task Force (IETF) for network mobility is the Mobile IPv6. in MIPv6, MN uses two address, permanent address and CoA to communicate from the temporary location. MN send the binding update (BU) message with the new CoA to the HA. The one enhanced feature in MIPv6 is route optimization which is used to avoid the triangular routing, in this way MN has direct binding with the CN. MIPv6 provide both L2 and L3 latencies because it deals layer 2 and layer 3 sequentially. So this delay can not be compensated in real time application such as video conferencing and VoIP. There were proposed many solutions to reduce the latency of MIPv6. Fast Mobile IPv6 [2] gives the concept of pre-acquisition of CoA with the help of Layer 2 trigger.


The aim of the FMIPv6 is to reduce the handover latency of MIPv6. in FMIPv6 when a MN find a new Access Router (AR), MN trigger a L2 mesasge to its old AR with the MAC address of new AR and ask it to set Binding Update and make the bidirectional tunnel with it. In this way before the end of layer 2 handover, ARs have done BU and acquisition of CoA. One precaution is taken that MN should not use the CoA unless it receives the acknowledgement.

Now we will discuss the messages which take place during the initiation of trigger to L3 handover. In this scheme MN send a Router_Solicitation_For_Proxy to the PAR also for response it send a Proxy_Router _Advertisement which contain a new CoA of the new AR for the MN. The PAR sends a handover initiation message to get the new CoA to the new AR. The new AR replies with the handover acknowledge which contain the CoA. And new AR register the MN in the neighboring cache with this CoA.MIPv6 also provide stateless address configuration to give the flexibility in fast handover.

1.0 Figure FMIPv6 Showing Handover [18]

Just before moving the MN send a FBU (Fast Binding Update) message to the PAR. On receiving this message it create a tunnel with new AR to send the message. And send a Binding Update Acknowledgement. Now the layer 3 handover is ready. Now the trigger Link_Down initiate the layer 2 handover. When the L2 process is done the Link_up trigger is fired to instruct the MN to send the Fast Neighboring Advertise (FNA) to the new AR. After the Successful process of FNA by the new AR, all buffered messages are delivered to the MN which were stored during the process of hand over. In this way no packet is lost. As we discussed before that both layer 2 and layer 3 handover goes parallel. Now two conditions can happen. If the layer 2 handover time exceeds the FBU/FBack messages, the handover mode is predictive. and vice versa the mode will be reactive handover. If handover went into reactive mode then MN after L2 messaging perform the address configuration.

2.0 Implementation and Experiment

The FMIPv6 modes i.e. Predictive or Reactive is obtained by time sequence of the L2 handover and the Exchange of Binding updates and also the messages that precedes these handover. It is regarded as Reactive if the L2 handover occurs earlier than the updates and messages. These tests were based on the IEEE 802.11b AP s linux based. The network is designed as shown in the above figure. The systems contain intermediate router, sniffer systems, and a mobile node. Here three identical systems are used PCMCIA-slot on board to make it easy to supply them with our WLAN cards they are used as they have Intel prisim2 chipset [last 1] also they are compatible with the host AP drivers. These drivers provide power control. There are also additional system monitors each wireless of ARs.

Figure 2.0: Experimental setup of the Network

3.0 Results

The Overall handover latency has L2 and L3 handover is as shown in table. The aim of this report is to compare different schemes of handover while switching from one AR to other AR. There are two level of handover, layer 2 handover and layer 3 handover. Here are little discussions about layer 2 handover more emphasize on overall latency. There is comparison in the latency of handover by comparing Mobile IPv6 and FMIPv6. FMIPv6 is further divided into reactive mode of handover and predictive mode of handover. And the other cases of FMIPv6 are Anticipated Handover and Tunnel-Based Handover. From the comparison of MIPv6 and FMIPv6 it shows FMIPv6 gives less latency as compare to MIPv6. as shown in the table given below predictive mode of FMIPv6 gives low latency which is less than 1 second.

L2 latency (ms)

Sum of latency

473.9 ms

L3 latency (ms)

Sum of latency

2695 ms

Sum of total handover latency

3134 ms

Table 4.1: MPIPv6 handover latency

Reactive mode handover latency

Sum of latency L2 &L3 latency

2516 ms

Predective mode L2 & L3 handover latency

Sum of latency L2 & L3 latency

958 ms

Table 4.2: FMIPv6 handover latency

4.0 Conclusion

From the above argument it can be concluded that FMipv6 in predictive mode has the best latency time in handover and also it is effective for real time application s such as voip and video communication. The Results show that it has an over all latency of 958ms in the implementation carried out which is way better than other modes discussed hence this method of handover is preferred.