LocationAided Routing LAR in Mobile Ad Hoc Network

“Location-Aided Routing (LAR) in Mobile Ad Hoc Network” by Young-bae ko Nitin H. Validya presented by Mark Miyashita

Organization • Introduction • Related Work • Location-Aided Routing (LAR) protocol – – – Route Discovery using Flooding Location information Expected Zone and Request Zone LAR Scheme 1 LAR Scheme 2 Error in location estimate • Simulation Model and Results • Variations and Optimizations

Introduction Mobile Ad hoc Network(MANET) • Node mobility which is the cause of frequent and unpredictable topology changes leads to difficult task of route maintenance in MANET • Many protocols have been proposed for MANET to achieve efficient routing • This paper suggest an approach to decrease overhead of route discovery by utilizing location information (GPS or other method to obtain location information) • Two LAR protocols for route discovery presented in this paper uses location information(may be out of date) to limit search space which results in fewer route discovery messages

Related Work • Many protocols have been proposed for MANNET such as DSR, AODV, TORA, ZRP • Existing MANET routing algorithm mentioned do not utilize physical location of a destination node • Similar idea (utilizing location information) have been applied and developed called “selective paging for cellular PCS (personal communication service) networks • In selective paging, the system pages a selected subset of cells close to the last known location of mobile host which decrease location tracking cost • This paper propose and evaluate an analogous approach for routing in MANET

Route Discovery Using Flooding • This paper discuss the basic flooding algorithm and location-aided route discovery based on limited “flooding” Basic Flooding Algorithm – A source node S needs to find a route to destination node D, node S broadcasts a route request to all its neighbors – Intermediate node X receives a route request and compares the destination with its own identity – If it does not match, then node X broadcast the request to its neighbors(sequence numbers used to detect duplicate and eliminate/avoid redundant transmissions) – Node D responds by route reply messages to sender which traverse the path in reverse of the path received by D (route request packet contains path of all nodes traversed starting S)

Route Discovery Using Flooding Basic Flooding Algorithm – Timeout scheme is also used to re-initiate route request with new sequence number due to transmission error or node D is unreachable from S

Route Discovery Using Flooding • In this paper, implementation assumes that node S can know that route is broken only if it attempts to use the route by sending data packet and receiving route error messages – it initiates route discovery for D • Note that route request may reach every node in the network that is reachable from S (potentially all nodes in the MANET) • This paper claims that by using location information reduces the number of nodes to whom route request is propagated (limit the scope of route request propagation)

Location Information • Location information can be obtained by the use of Global Positioning System (GPS) • With use of GPS, mobile host can know its physical location – note that GPS includes some degree of error compared to the real coordinates and GPS-calculated – NAVSTAR GPS has positional accuracy of 50 -100 meters – Differential GPS has positional accuracy of few meters • This paper assumes that each node knows its current location precisely – possibility of error in location are discussed separately in the performance evaluation • Also assume that the mobile nodes are moving in a twodimensional plane

Expected Zone • The Expected Zone is the region where source node S thinks that the destination node D may contained at some time t – only an estimate made by S – Assume that node S knows that the node D was at location L at time t 0 and current time is t 1 – From the viewpoint of S, expected zone of node D is the region that node S expects to contain node D at time t 1 based on the knowledge that node D was at location L at time t 0 • If S knows that D travels with average speed v, then S assumes that the expected zone is the circular region of radius v(t 1 - t 0) centered at location L • Note that if actual speed is faster than the average, then the destination may be outside the expected zone at t 1

Expected Zone • Without the previous knowledge of the location of D, S will assume that the entire region is the expected zone and implementation uses the basic flooding algorithm • The size of expected zone can be reduced if node has more information about the mobility of a destination D

Request Zone • Node S defines (implicitly or explicitly) a request zone for the route request • Node forwards a route request only if it belongs to the request zone (it does not forward a route request to its neighbor if outside of the request zone) • Two LAR scheme differ in determining the membership of request zone • The request zone includes expected zone in addition to (possibly) other surrounding zone around the request zone • If a route is not discovered within the timeout period, S initiates a new route discovery with expanded request zone – all paths from S to D include nodes that are outside the request zone • Note that the probability of finding path can increase as size of request zone increases (route discovery overhead also increases with the size of the request zone

Request Zone

LAR Scheme 1 • The request zone is rectangular in shape • Assume S knows that the node D was at location (Xd, Yd) at time t 0 • Assume S knows the average speed v with which D can move • From above two, S defines the expected zone at time t 1 with radius R = v(t 1 - t 0) centered at location (Xd, Yd) • The request zone is the smallest rectangle that includes current location S and the expected zone such that the sides of the rectangle are parallel to the X and Y axes • Node D sends route reply message with its current location and time (may include average speed but simulation assumes all nodes knows each other’s average speed)

LAR Scheme 1

LAR Scheme 1 • Size of the request Zone is proportional to (i)average speed of movement v and (ii)elapsed time since recorded last location of the destination • Recall that R = v(t 1 - t 0) is used to determine the size of request zone • In general, a smaller request zone may be formed at speed that are neither too small nor too large • For instance, at low speed, factor (i) is small but route discovery occur after long intervals making (ii) larger (t 1 - t 0 is large)

LAR Scheme 2 • Node S includes two pieces of information with its route request – Assume that S knows the location (Xd, Yd) of D at some time t 0 which route discovery is initiated by S at t 1 where t 1 to – S calculates its distance from location (Xd, Yd) denoted DISTs and included with the route request – The coordinate (Xd, Yd) are also included with the route request • When node I receives the route request from S, node I calculates its distance from (Xd, Yd) denoted DISTi and: – For some parameter , if DISTs + DISTi, then I forwards request to its neighbors – this request includes (Xd, Yd) and DISTi replacing original DISTs and (Xd, Yd) from S – Else DISTs + DISTi, node I discards the route request • Each intermediate nodes repeat the process above

Comparison of the two LAR Schemes

Error in Location Estimate • Both LAR schemes assume that each node knows its own location accurately. However, just like GPS, there may be some error in the estimated location • Let e (location error) denote maximum error in the coordinates estimated by a node • If a node N believes that it is at location (Xn, Yn), then the actual location of node N may be anywhere in the circle of radius e centered at (Xn, Yn) • If LAR Scheme 1 is modified to take e into account, then the expected zone is a circle of radius e + v (Xn, Yn) which makes request zone larger since it includes larger expected zone • No modification is made to the LAR Scheme 2

Performance Evaluation • The simulation is performed using modified version of Ma. RS (Maryland Routing Simulator) • Ma. RS is discrete-event driven simulator providing a flexible platform for the evaluation and comparisons of network routing algorithm • Simulations were performed on flooding, LAR scheme 1, and LAR scheme 2 • Simulations are conduct by varying the number of nodes, transmission range of each node, and moving speed

Simulation Model • Number of nodes in the network was chosen to be 15, 30, and 50 for different simulation runs • The size of ad hoc network is 1000 unit x 1000 unit square region • Initial locations of the nodes are obtained using a uniform distribution • Each node moves continuously without pausing at any location – move with average speed v in the range 1. 5 to 32. 5 units/sec • The actual speed is uniformly distributed in the range v – and v + units/sec where =1. 5 when v < 10 and = 2. 5 when v 10 • A node travels distance d where d is exponentially distributed with mean 20 units • The direction of movement for a given move is chosen randomly • All nodes have the same transmission range

Simulation Model • Transmission range of 200, 300, 400, and 500 units were used with all wireless links having the same bandwidth 100 Kbytes per second • Transmission errors are not considered • The simulation time is inversely proportional to the average speed – as the average increased for given simulation, so does the number of moves simulated • A sender and a destination are chosen randomly • Any data packet that cannot be delivered to the destination due to a broken route is dropped • The source generates 10 data packet per second on average with the time between two packets being exponentially distributed • Assume each node knows its location accurately

Simulation Model

Optimization • Accuracy of a request zone can be improved by adapting the request zone determined by the source node S initially • Idea is to use location information at some intermediate node which may know more recent location for destination node than the source node (assuming source information is out of date compare to intermediate node) • Thus, using this up-to-date information at the intermediate node with its expected zone and adopting the request zone of source node • In Scheme 2, intermediate node may calculate distance from the more recent location of destination D, and use this distance in making decision whether or not to discard a route request

Optimization