MODIFICATION AND COMPARISON OF DSDV AND DSR PROTOCOLS

MODIFICATION AND COMPARISON OF DSDV AND DSR PROTOCOLS Department of computer science Thadomal Shahani engineering college Mumbai-400051, ABSTRACT: In this era of mobile devices, mobile ad-hoc network (MANET) plays very crucial role. Mobile ad-hoc network is formed collectively by two or more mobile nodes enabled to communicate. These networks don t have static infrastructure. In this paper improvements are proposed in two important MANET protocols which are DSDV and DSR protocols and focus of paper will be on their comparison with existing traditional protocols and with each other on the basis of packet delivery ratio (PDR), overhead, packet dropping ratio, jitter, and throughput using Network Simulator 2 (NS2). Improved protocols are renamed as MDSDV (Modified DSDV) and MDSR (Modified DSR). Improvements in DSDV protocols are in delay and it also reduces congestion. Improvements in DSR are in terms of number of paths and load balancing. Unlike traditional DSR protocols MDSR builds and uses multiple path increasing reliability and security. Keywords: Multipath, Weighted Settling Time (WST), Packet Delivery Ratio (PDR), throughput, Network Simulator 2 (NS2). [1] INTRODUCTION Wireless networking is the technology using which user can access information data and services from anywhere without having physically connection with network. Wireless networks are broadly categorized as: - a. Infrastructured networks Infrastructured network is a network which has fixed access points or base station with all the clients connected with it. All the clients or devices communicate through the access points with each other. [1] b. Ad-hoc networks Ad-hoc networks which also known as peer-to-peer networks don t have any centralized server or access points. All nodes or devices of these networks are connected arbitrarily without any fixed structures. Each node in these networks are responsible for route discovery and update. Ad-hoc routing protocols basically classified into two categories as: 1. Table driven protocols 61

MODIFICATION AND COMPARISON OF DSDV AND DSR PROTOCOLS In this protocols up-to-date routing information is maintained at each node at each moment. Thus routes are always available 2. On demand protocols In this protocols routes are created when demanded. Whenever source wants to send packets, routes are invoked from source to destination. In last few years various routing protocol are developed specifically for this environment. There are mainly four mobile ad hoc networks (MANET) routing protocols. They are also called baseline protocols. They are Destination sequenced distance vector (DSDV), Dynamic source routing (DSR), Optimized link source routing (OLSR), Ad-hoc on demand distance vector routing (AODV) protocols. These protocols are termed as baseline protocols as they are considered as guidelines for implementation of new protocols. While DSDV and OLSR are table-driven routing protocol and DSR, AODV, fall under the On-demand routing protocols category. These traditional protocols have some limitations that can be resolved. In this paper some improvements are presented in DSR and DSDV protocols in order to eliminate some of their limitations and improve performance and efficiency. [2] [2] OVERVIEW On the basis of process of route establishment and updating it, protocols are categorized as proactive and reactive protocols where proactive are table driven protocols and reactive means by on demand protocols. DSDV is proactive and DSR is reactive protocol. These traditional protocols follows different procedures for route establishment. [2.1] DSR This protocol uses source routing. It means source or sender knows the route. Routes for different destinations are stored in route cache. If sender does not know route already then it invokes route discovery process. This dynamically finds routes to required destination hop by hop. It floods network route request (RREQ) packets which are rebroadcasted by subsequent nodes until the route to destination node found out. Such a node send route reply (RREP) packet back to source via route it got RREQ packet. For any errors in route RERR packet is sent back. [5] As shown follow S is source and D is destination and travelling of RREQ and RREP packet is shown. [2] Figure 1. Route request and route reply in DSR 62

[2.2] DSDV Each node maintains a routing table that contains lists of destinations that are available, total number of hops to reach destination from selected node and the sequence number. There might be formation of loops due to using same old routes. To resolve this problem sequence number is used to distinguish new routes from old ones which avoid the formation of loops. Routing tables are transmitted periodically by station to adjacent nodes and also in events like if any node is added or deleted also if path crashes. Node uses odd sequence number for link break and even number for normal table update. [1] Node B broadcasts its table to adjacent nodes. Figure 2. Routing table update in DSDV Criteria for route updates: 1. Preference is given to routes of newer sequence number. 2. In case of route with same sequence number, it is preferred if it offers better metrics. Older routes are discarded. But updates may be send by many nodes in networks independently in different intervals. So it might happen that some node may keep changing its routes continuously. This is known as damping fluctuation and caused due to second criteria of better metric preference. [4] Concept of weighted settling time (WST) is used to solve this. It is calculated for each node by using weighted average of recent route update message. Criteria for using WST 1. For unreachable node immediately forward update as soon as it arrives. For previously known reachable node, it has to wait for time referred from WST table before forwarding update [3] LIMITATIONS OF TRADITIONAL DSDV AND DSR PROTOCOLS [3.1] PROBLEMS WITH DSDV: If we consider a scenario where packet B want to send packet to Z and it has two routes available having 6 and 8 hops each. If packet reaches destination via route having 8 hops first than the route having 6 hops then according to metric criteria route having 6 hops is selected. [8] Another problem is In this case node has to wait more time hence increasing WST. This is in contrast with concept of WST for reducing delay as delay in network increases with WST. The reason behind the packet reached node later via shorter route is must be congestion. Congestion might be caused by different reasons like capacity of processing node 63

MODIFICATION AND COMPARISON OF DSDV AND DSR PROTOCOLS or collision of packets. Shorter route with congestion is selected based on metrics then it makes the matter worst as packets are sent through route which is already congested. For benefit of few milliseconds few seconds are wasted. [9] [3.2] PROBLEMS WITH DSR Traditional DSR protocol invokes the process to find routes to destination on demand. From working of the DSR protocol it is clear that when RREQ packet find the route to destination first, RREP packet is sent via same route back to destination and all packet sending is done by same route in future. This method does well but when it comes to security and reliability concern then it don t look that convincing. If the route breaks or fails due to any reason then source has to again invoke the route discovery process. Problem of traditional DSR is about security, reliability and compulsion. [6] [4] PROPOSED WORK [4.1] PROPOSED METHOD FOR MDSDV: The root of the mentioned limitations of DSDV protocols as delay and congestion in network is its policy of better metric which favors the route with minimum hop counts. This requires minor change in the route update algorithm. The new algorithm for MDSDV consist of following modifications in the process of route updates. [4] 1. Route advertisement message arriving should be advertised as FCFS basis without considering the metric. Message arriving later should be discarded or stored as it might be useful in future 2. Message with newer sequence number are preferred in any condition. [4.2] PROPOSED METHOD FOR MDSR Though there are not much problems with traditional DSR protocols, it still can be updated. This protocol will work better with Multipath routing. Multipath routing works with alternative paths means it gives more than one path so they are available as alternatives. [7] For using this with DSR algorithm there are two data structures introduced: One is Route cache which consist of the alternative routes available. It consist of three fields which are Route_number to differentiate between the routes. Route which consist of route from source to destination and then validity field is to assign validity period before which the route should be considered. Another data structure is Load Table which keeps track of loads on the alternative routes. It consist of two fields Route_number and traffic. Traffic filed consist of the count of number of packets transferred on each route means load per route. [5] Working of MDSR: Each type of node has different part to play. Source node maintains up to 5 paths in its route cache unlike traditional DSR protocol which refers route cache for sending packets to next hop as mentioned in cache, it first has to refer route cache and then it has to look in Load table for balancing loads between routes. Route with least load is preferred. 64

Intermediate node can only pass on RREQ to next node. Destination node however has different job than it had in traditional protocol. It has to reply to first 5 RREQ with RREP and rejects others. [5] RESULTS AND DISCUSSION For simulation environment Network simulator 2.34 (NS-2.34) is used and network animator nam is used for visualization and comparison graphs are plotted in xygraph. Simulation is run for existing original DSR and DSDV protocols as well as new proposed MDSR and MDSDV protocols. Simulation is done under same environment for each protocols mentioned and compared them in identical conditions. Simulation is done for different number of nodes for different time period to get accurate results. Following scenario used for simulation. [3] Table 1 Simulation setup and parameters Number of nodes 60,80,100,120,140 Area 1000 1000 Traffic type Constant bit rate Packet size 512 bytes Mobility Random Velocity 10 Transmission range 250m Sensing range 250m Pause time 5 sec WST 6 sec Simulation time 300, 500, 1000 sec Periodic route update 15 sec interval Number of sources 15 [5.1] PDR W.R.T. NODES It is a ratio of number of packets received at destination to total number of packets sent. From figure 3 it is clear that for any number of nodes PDR shows no uniform behavior but MDSR and MDSDV gives better PDR than original protocols. [5.2] CONTROL OVERHEAD W.R.T. NODES This overhead consist of the information exchanged over network for network control tasks such as scheduling routing. It contain information as source and destination address, queue length, etc. From figure 4 it can be seen that control overhead increases significantly with 65

MODIFICATION AND COMPARISON OF DSDV AND DSR PROTOCOLS number of nodes in DSDV whereas it shows slight decrement in case of DSR. MDSDV and MDSR has edge over original protocols. [5.3] NORMALIZED ROUTING OVERHEAD W.R.T. NODES Normalized routing overhead is total number of routing packets transmitted per data packet. As in the figure 5 result shows that normalized routing overhead increases with number of nodes in DSDV and there is no significant change is shown in case of DSR. Both modified protocols has better results over original. MDSDV lot lesser overhead than DSDV. [5.4] PACKET DROPPING RATIO W.R.T. NODES It is the ratio of total number of packets dropped to total number of packets sent. From figure 6 it is clear that PDR is very high in case of DSDV and greater in case of DSR as compared to MDSDV and MDSR. [5.5] JITTER W.R.T. NODES Each packet does not reach to destination in equal intervals there can be irregularities in delay. Jitter is change in latency from packet to packet. This may be caused due to various reasons as packet queue size or buffer size. From results shown in figure 7 it is clear that DSDV gives very high jitter. Jitter in DSR are lesser than DSDV but MDSDV and MDSR outperforms both original protocols. [5.6] THROUGHPUT W.R.T. NODES It is calculated as packets delivered successfully to destination per second. From figure 8 it can be seen that DSDV gives very poor throughput where DSR gives better throughput but MDSR and MDSDV gives better throughput than both original protocols. Figure 3. Control Overhead w.r.t. Nodes 66

Figure 4. Normalized Routing Overhead w.r.t. Nodes Figure 5. Packet Dropping Ratio w.r.t. Nodes Figure 6. Jitter w.r.t. Nodes 67

MODIFICATION AND COMPARISON OF DSDV AND DSR PROTOCOLS [6] CONCLUSION Figure 7. Throughput w.r.t. nodes Comparison of the modified MDSDV and MDSR protocols done with each other and with original DSR and DSDV protocols. From simulation result based on various criteria for different number of nodes it is clear that the DSR works better if load balancing multipath routing is used instead of original single path routing. Multipath routing increases reliability and thrives well in case of failures as well as its load balancing feature reduces chance of congestion. Also it is clear from results that DSDV is better with proposed new rules of routing table updates, as it performs better by not sending packets to already congested path also it reduces delay for data transmission. REFERENCES [1] Hemanth Narra, Yufei Cheng, Egemen K. Çetinkaya, Justin P. Rohrer and James P.G. Sterbenz Destination sequenced distance vector (DSDV) routing protocol implementation on ns-3 WNS-3 2011, March 21, Barcelona, Spain Copyright 2011 ICST 978-1-936968-00-8 DOI 10.4108/icst.simutools.2011.245588. [2] Yufei Cheng, Egemen K. Çetinkaya, and James P.G. Sterbenz Dynamic source routing protocol implementation in ns-3 WNS-3 2012, March 23,Italy Spain Copyright 2011 ICST978-1-936968-47-3 DOI 10.4108/icst.simutools.2012.247749. [3] Understanding the basics involved in ns-2 simulations: http://www.isi.edu/nsnam/ns/tutorial/index.html [4] Mohd. Nasee, Chiranjeev kumar, EDSDV: Efficient DSDV Routing Protocol for MANET, 978-1-4799-1597-2/13 2013 IEEE [5] Lokesh Kumar Malviya, Damodar Tiwari, LMP-DSR: Load Balanced Multi-Path Dynamic Source Routing Protocol for Mobile Ad-HOC Network, 4th ICCCNT 2013 July 4-6, 2013, Tiruchengode, India [6] Ehsan Khosrowshahi Asl, Morteza Damanafshan, Maghsoud Abbaspour, EMP-DSR: An Enhanced Multi-Path Dynamic Source Routing Algorithm for MANETs Based on Ant Colony Optimization,Third Asia International Conference on Modelling & Simulation, IEEE 2009 [7] D. Jagadeesan, S.K. Srivatsa, Multipath Routing Protocol for Effective Local Route Recovery in Mobile Ad hoc Network, Journal of Computer Science,Science Publications 2012. 68

[8] T.S. Kumaran, V. Sankaranarayanan, "Early detection congestion and control routing in MANET," in WOCN, vol. 1, no. 5, Sept. 2010, pp. 6-8. [9] Hasna Abboud. A Study of Congestion Control Mechanism Using route failure detection in MANET American Journal of Applied systems, vol. 3, 2013. 69