VoIP Mobility

Pavan Kundhavaram

Contents Introduction.

VoIP Mobility.

Issues.

Conclusion.

References.

VoIP Introduction

VoIP is Voice Over IP.

VoIP is the routing of voice conversations over the

internet or any other IP-based network.

Voice Calls are transmitted over Packet switched Network instead of Public Switched Telephone Network(PSTN).

VoIP allows users to travel anywhere in the world

and still make and receive phone calls.

How It Works

VoIP converts the voice signal from telephone into a digital signal that travels over the internet then converts it back at the other

end .

A broadband connection is required in order to place VoIP call.

VoIP Mobility

Mobility include :

Terminal Mobility.

User Mobility.

Service Mobility.

Issues Optimizing the Handover Delay .

Mobility Management for VoIP Traffic.

VoIP Seamless Handover.

Issue#1:Optimizing Handover Delay

The support of IP-based real time services in the next-generation systems require coupling of mobility with QOS.

Coupling host mobility and quality of service is one of main challenges.

The mobile node can experience disruptions or intermittent disconnections of on going real-time session due to handovers.

Issue#1 contd.. The time or duration of such interruption is called disruption

time.

The handover delay is the time interval from when the handover process starts to when the first data packet is received by MN.

The handover delay can heavily affect the user satisfaction .

Proposed Solution

First step : Simple model that takes into account the delay increases between the different entities involved in the handover.

Second step: Considers FER of the wireless link and the retransmissions strategies of the different protocols to overcome the losses.

Simple Model for Analysis

Analysis contd.. the delay between the MN and the Radio Access Network

(RAN) is tmr.

the delay between the MN and the Access Router (AR)is ts.

the delay between the MN and the FA/MAP is tmf.

the delay between the MN and its HA is assumed to be th.

the delay between the MN and the CN is tmc.

the delay between the MN’s home network and the CN is thc.

Assumptions ts<th.

In MIPv4 FACoA instead of CCoA is used therefore MN’s incoming and outgoing traffic is relayed by the FA.

The MN sends regularly solicitations after leaving one network.

For each binding update (BU) message sent ,Binding Acknowledge (BA) is expected to be received.

For MIPv6 registration we do not consider the time needed by Duplicate Address Detection (DAD) process.

MIPv4 handoff The MN detects the IP subnet by exchanging Agent

Solicitation and Agent Advertisement messages which takes 2tmf.

Then, the MN sends a MIP Registration Request to the HA and gets a Registration Reply, which takes 2th.

At this MN starts receiving downlink packets.

The MIPv4 handoff takes 2tmf + 2th .

MIPv6 handoff The MN detects the IP subnet by exchanging with AR Router

Solicitation and Router Advertisement messages that takes 2ts.

MN sends to HA a Binding Update and gets a Binding Acknowledge that takes 2th.

Finally, the MN sends to the CH a Binding Update and gets a Binding Acknowledge that takes 2tmc.

The MIPv6 handoff delay is 2ts + 2th + 2tmc .

Second Step Assumptions A random error process.

An Agent/Router Advertisement is sent only if a Agent/Router Solicitation has been previously received.

An Registration Reply/Binding Acknowledge is sent only if a Registration Request/Binding Update has been received previously.

Error correction mechanisms and processing /queuing times are not considered here.

Second Step contd.. Probability of the frame being erroneous in the air link is p.

For k frames in MIP packet ,the packet loss rate is (1 − (1 − p)k) .

We denote τ as the inter frame time. D as the frame propagation delay through the RAN.

Propagation delay from MN to RAN for a MIP message is D + (k − 1)τ .

Adaptive Retransmission timer The retransmission timers for all the MIP-based protocols

follow the exponential back-off mechanism.

Tr(1) be the initial back-off timer.

The back-off timer upon the ith transmission Tr(i) doubles after each retransmission

Tr(i) = 2 (i−1) · Tr(1)

contd.. the initial retransmission timer Tr(1) is a crucial

parameter which should be optimized

It should not be too short.

It should not be too long.

It is proportional to the transmission time of the messages involved in the handover transaction.

Back-off interval timer

Retransmission Probability

The probability of retransmission q is the probability of a transaction having failed

The probability of having a retransmission of Solicitation is: q = 1− ((1 − p)k1+k2 )

Average Handover Delay

Let Nm be the maximum number of transmissions.

The average delay for a successful transaction is the average delay for successfully transmitting and receiving the corresponding acknowledgement of an MIP message.

The average handover delay TtMIP is given as: TtMIP = ∑ Tt(i)MIP

contd..

Tt(i)MIP =1/1 − qNm · [(1 − q) (D + (k − 1)τ ) +(1 − q)q(Tr(1) + D + (k − 1)τ ) +(1 − q)q2 · (3Tr(1) + D + (k − 1)τ) + · · · +(1 − q)qNm−1 · ((2Nm−1 − 1)Tr(1) +D + (k − 1)τ )] = D + (k − 1)τ − Tr(1) + ((1 − q)(1 − (2q)Nm))/(1 − qNm)(1 − 2q)) · Tr(1)

MIPv4 Handover Delay MIPv4 average handover delay is

TtMIPv4 = Tt(AgSol) + Tt(AgAdv) + 2trf

+2trh + Tt(RegReq) + Tt(RegRep).

Where trf is the delay between the RAN and the FA

(trf =tmf − tmr) and trh is the delay between the RAN

and the HA (trh = th − tmr).

MIPv6 Handover Delay MIPv6 average handoff delay is as follows:

TtMIPv6 = Tt(RSol) + Tt(RAdv) + 2trf

+2trh + 2trc + 2Tt(BU) + 2Tt(BA).

where trc is the delay between the RAN and the CN trc = tmc − tmr.

where trc is the delay between the RAN and the CN trc =tmc − tmr).

Numerical Results

Disruption time vs. FER

Issue#2: Mobility Management for VoIP Traffic

IP based mobility management traditionally operate at the network layer and provide basic connectivity to the MN as they change their point of attachment.

MIP ensures ubiquitous connectivity by allowing MN to retain its permanent home address(PHoA) and by tunneling packets to temporarily care of address(CoA).

These solutions are necessary for VoIP application in dynamic tactical battlefield networks.

Issue contd.. MIP potentially high update latency makes it

unsuitable for supporting seamless handoffs during ongoing call.

SIP at application layer offers many advantages over corresponding MIP but suffers a drawback of absence of mobility management hierarchy.

SIP and MIP use flat hierarchy in which every change in MN requires generation of global binding updates.

Updates incur high latency and make rapid handoffs impossible.

Draw backs of Flat Architecture

On every change in subnet.

MN refreshes its configuration information (COA) .

Generate global bindings to update remote nodes with new COA.

In absence of hierarchy every update travel all the way to the remote node.

Update process can have high latency because of communication delay.

If there is packet loss latency becomes much higher at intermediate hops.

Solution:DMA Architecture The DMA Architecture is based on two-level mobility

management hierarchy.

IDMP is used as the protocol for managing mobility within a domain. The Mobility Agent(MA) is similar to MIP foreign Agent (FA)

excepts it resides higher in network hierarchy and acts as a domain-wide point for packet redirection.

A Subnet Agent (SA) is similar to MIP FA and provides subnet-specific mobility services.

IDMP Architecture

contd.. Under IDMP MN has two concurrent CoAs:

Global Care of Address(GCoA).

Local Care of Address (LCoA).

Packets from a remote CN are forwarded to the GCoA and are intercepted by the MA.

The MA tunnels these packets to the MN’s current LCoA. Global binding updates are generated only when the MN

changes domains and obtains a new GCoA, This approach drastically reduces the global signaling load.

Elements of DMA Architecture

Dynamic technique The DMA architecture defines a dynamic technique for

assigning an MA to an MN when it first moves into the domain.

The architecture assumes the presence of multiple MAs and applies a load balancing technique for distributing the mobility load across the multiple MAs.

A central node called the Mobility Server (MS) implements different load balancing and MA-allocation strategies.

Contd.. The architecture also uses the Differentiated Services

framework to dynamically provision domain resources and provide an MN QoS guarantees as it moves within the domain.

DMA requires MN to obtain a new LCoA if network mobility is confined to single mobility domain.

A group of 200 soldiers communicating with 5CNs would generate 1000 simultaneous global binding updates under flat architecture but only 200 local updates under DMA approach.

Signal Flow of VoIP Mobility

Issue#3:VoIP Seamless Handover

The period from when the MN last receives data traffic via its old IP subnet to when it receives it new IP subnet is handover delay.

Delay is divided into four sub-delays:

Layer 1/Layer 2 radio link switching delay.

L2 access re-authentication delay.

IP layer binding delay.

Application layer authentication and registration delay,

Contd..

The Inter-AP handoff is reduced Inter-Access point protocol is proposed.

L2 re-authentication delay could be reduced during inter-AP roaming.

IP layer binding delay is due to allocation of dynamic IP address via DHCP followed by routing path update to new AP.

DHCP delay in Mobile IP and application layer authentication and registration delay in SIP mobility is a challenge.

Solution:VPN Technology

The MN is identified by its static private IP address regardless of its current point of attachment to the subnets.

This allows the MN to use the same IP address during handover.

When the mobile host hands off to any other AP the new AP receives session information in advance hindering further messages.

The delay of re-authentication for the MN is reduced.

Link Layer

The packet loss and end to end transmission delays can be reduced.

MN moves from one subnet to another subnet without interruption only if

MN should communicate simultaneously with multiple APs.

The network must duplicate and correctly merge the IP flows from the CN to the MN through different APs.

Multi-Homing Concept

The multi-homing feature enables the MN to support seamless handover by simultaneous binding of two different addresses.

The packets are multicast to MN and MIP agents without need to tunnel packets to the NAR form the PAR as in Mobile IPv6 networks.

The packet loss is reduced during the handover.

Mobile Agent Technology

The MA is software component which can be transferred from one network element to another while carrying on its status of execution.

MA technology can diminish network traffic and can maintain load balancing thus improving network performance specially in mobile environment.

MA technology in VoIP services includes reducing control packets, processing the SIM-based authentication via the VPN tunnel at new location of attachment and secure packet tranmission.

Mobile Agents to support seamless VoIP service

Both IP layer binding delay and application layer authentication and registration delay are major parts of the overall handover delay.

The delay of IP address renewal (> 2s) has significant effect on the overall handover performance.

The application layer authentication and registration delay is harder to reduce than the DHCP delay and cannot be ignored due to security consideration.

Seamless Handover Architecture

Solution Contd..

Layer 2 Tunneling Protocol (L2TP) VPN tunnels are constructed between the L2TP Network Server (LNS) and all L2TP Access Concentrators (LACs).

Service and authentication requests and data packets are protected under IPSec tunnels while transmitted between the MN and LNS.

They are further encapsulated into L2TP VPN tunnels during transmission between the LNS and LAC.

contd..

The LNS function as a service proxy to forward the service requests from the MN to the application server.

To minimize the DHCP delay, IP binding delay and application layer authentication delay there are three techniques

VPN with a private static IP address. Multi-homing. Mobile Agent.

contd.. L2TP VPN can be implemented as an Intranet.

It can have the static private IP addresses assigned to its private MNs regardless of their location.

The fast handover for Mobile IPv6 tries to minimize the period of service disruption by the packet tunneling mechanisms while performing network layer handover.

The multi-homing concept is used to minimize the disruption time and packet loss ratio.

Message flows During Handover

Conclusion The issues discussed above deal with the various VoIP protocols and

various standard both at the network layer and application layer.

In order to achieve transmission during roaming is a challenge and this can be achieved with proper hand over of the signal to the next BS.

References Fathi, Chakraborty, Prasad .”Mobility management for VoIP:

Evaluation of Mobile IP-based protocols”.IEEE ,2005.

Misra ,Das,Anthony.”Hierarchical Mobility Management for VoIP Traffic”.IEEE 2001.

Lin ,Shun Yang.” Mobile Intelligent Agent Technologies to Support VoIP Seamless Mobility”.IEEE 2005 .

T. T. Kwon, M. Gerla, and S. Das, “Mobility Management for VoIP service: Mobile IP vs. SIP,” IEEE Wireless Communications, vol. 9, no. 5,pp. 66–75, October 2002.