Effects of Applying Mobility Localization on Source Routing Algorithms for Mobile Ad Hoc Network Hridesh Rajan presented by Metin Tekkalmaz
2 Outline Introduction Source Routing (SR) Problem Related Work SR w/ Mobility Localization (ML) An Example Scenario for ML Evaluation of ML Conclusion
3 Introduction Routing is one of the problems in Mobile Ad Hoc Networks (MANETs) Source Routing (SR) is one of the solutions SR is easily affected by topology changes Aim is to minimize possible overheads due to toplogy changes
4 Source Routing Problem (1/2) SR is a problem of finding maintaining a complete ordered list of nodes through which the packet should pass to reach the destination.
5 Source Routing Problem (2/2) SR consists of Route discovery Route maintenance Discovery procedure is done before any data packet can be send The only overhead associated with SR is discovery (or maintenance) procedure But this overhead can be significant due to rapid topology changes
6 Related Work Works to reduce overhead in SR can be grouped as two catagories Decrease the number of nodes receiving route queries Decrease the number of route requests generated by the source node Location Aided Routing is an example of the first group and Dynamic Source Routing is an example of the second one
7 SR w/ Mobility Localization Fundamental principle behind the ML approach is Even if a mobile node has moved at a time t, there is a fair possiblity that it might be found in the neighborhood until time t+δt δ depends on Velocity vector V Definitiaon of neighborhood
8 SR w/ Mobility Localization So in ML approach a node is attempted to be found in the vicinity instead of restarting the discovery protocol Use of ML affects maintenance procedure, not the discovery procedure
9 SR w/ Mobility Localization When the data link layer of a node reports a transmission error, it analyzes Source route Localization value (α) for the source route Its position in the source route
10 SR w/ Mobility Localization If the node is within α hops from the node encountering next hop error it tries to repair the source route on its own If it is successful traffic is forwarded to the new route Else it checks the value of localization again to see whether any other node just before itself in the source route is eligible to repair the route If such a node exists a route error message is sent to it
11 SR w/ Mobility Localization A node can receive a route error only if it is Within α hops to a node encountering next hop error in a source route the source node In the first case route repair and in the second case route request procedures are executed
12 SR w/ Mobility Localization A node receiving the route error sends a route repair message to every other one- hop neighbor except the node route error is received the node just before itself Route repair is sent in a unicast manner If there is no node to send route repair as unicast message, it is broadcasted Route repair messages have TTL-like info
13 SR w/ Mobility Localization When a node repair message is received It sends route repair reply to originator if it knows the requested node Otherwise it checks whether it can forward route repair message If it can forward, it forwards the message every node except the one the route repair is received from
14 An Example Scenario for ML
15 An Example Scenario for ML
16 An Example Scenario for ML
17 An Example Scenario for ML
18 An Example Scenario for ML
19 Evaluation Evaluation environment is NS Since prominent optimization techniques are already implemented effect of ML on top of others is experimented The maintenance cost is the number of control packets until network is stable Experiment networks have 10 to 20 and 25 nodes Value of localization is between 1 and 3
20 Evaluation
21 Evaluation
22 Evaluation
23 Conclusion ML is an enhancement over existing SR techniques in ad hoc wireless networks ML approach decreases the overhead by significant amount Overhead is not exponential with increase in nuber of nodes Value of localization should be arranged according to average source route length Probabilty of finding the moving node in the neighborhood is important
24 Conclusion Dynamic nature of MANETs strongly affects the effectiveness of ML Lcalization value can be arranged according to the changing dynamics of MANET