الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 179 |  |  |
| | | from | 179 | | |
Abstract
Mobile ad hoc networks (MANETs) are wireless networks characterized by the dynamic nature of their members. The topology of a MANET can rapidly change and, therefore, these networks require mechanisms which allow participant devices to communicate with each other in spite of their mobility.
Routing is a critical issue in a MANETs. Each node acts both as host and router and forwards packets for nodes that are out of transmission range. Generally, routing protocols in MANETs can be classified into Proactive and Reactive routing protocols. The pro-active routing protocols (Table-driven) are the same as current Internet routing protocols such as the Routing Information Protocol, Distance-Vector, Open Shortest Path First and Link-State. They attempt to maintain consistent, up-to-date routing information of the whole network. Some of the existing pro-active ad hoc routing protocols are: Destination Sequenced Distance- Vector (DSDV), Optimized link state routing (OLSR), Cluster Head Gateway Switch Routing, Global State Routing (GSR), Fisheye State Routing (FSR), Hierarchical State Routing (HSR), Zone based Hierarchical Link State and Source Tree Adaptive Routing. The Reactive routing protocols (On-demand) maintain only the routes that are currently in use, thereby trying to maintain low control overhead, reducing the load on the network when only a small subset of all available routes is in use at any time. Some of the existing reactive routing protocols are Ad hoc On-demand Distance Vector (AODV) routing protocol, Ad hoc On-demand Multipath Distance Vector (AOMDV) Dynamic Source Routing (DSR) protocol, Associativity Based Routing (ABR) protocol and Signal Stability Routing (SSR) protocol.
Since the proactive and reactive routing protocols in MANETs have relative advantage and disadvantage, then evaluating and comparing them is a very critical issue. Significant works has been carried out to evaluate and compare these routing protocols under various traffic types. However, very few did the simulation with Transmission Control Protocol (TCP) traffic sources for Constant Bit Rate (CBR) applications. CBR traffic pattern is very well known traffic model for MANETs which generates data packets at a constant rate, whereas, TCP provides reliability to data transferring in all end-to-end data stream services on the MANETs. There are several TCP traffic patterns such as TCP Reno, TCP New Reno, TCP Vegas, and TCP selective Acknowledgment (Sack).
Congestion is another critical issue in MANETs. The congestion can occur in any intermediate node, often due to limitation in resources, when data packets are being transmitted from the source to the destination. Congestion will lead to high packet loss, long delay and waste of resource utilization time.
This thesis studies the performance of AODV, DSDV, DSR and AOMDV routing protocols in MANETs using TCP traffic types such as TCP-Reno, TCP- Newreno, TCP-Vegas and TCP-Sack. The performance analysis is done in terms of packet delivery ratio, average end-to-end delay and average throughput using the NS2 simulator. The simulation results show that TCP-Vegas performs bet- ter compared with others in the case of end-to-end delay, and has higher packet delivery ratio. The TCP-Reno has highest throughput in the case of low data connections compared with TCP-Newreno, TCP-Vegas and TCP-Sack. In case of high data connections the TCP-Vegas have the higher throughput compared with the others.
This thesis considers also the effect of mobility models (Reference Point group Mobility (RPGM), Gauss Markov (GM) and Manhattan Grid (MG)) and traffic patterns TCP and CBR on the behavior of Reactive AODV and Proactive DSDV and OLSR routing protocols. The results of conducted experiments us- ing NS2 simulator show that the relative ranking of routing protocols may vary depending on both mobility models and traffic patterns.
In this thesis, an enhanced Partitioned Queue Management (PQM) scheme is proposed to avoid the congestion. The proposed scheme utilizes both a packet type and number of hops that a packet traversed from the source through the intermediate nodes to the destination node. The results of conducted experiments using NS2 simulator show that the proposed scheme outperforms the DropTail queue management scheme in terms of packet loss ratio, packet delivery ratio and end-to-end delay.
Finally, in the thesis a modification of the well-known AODV routing pro- tocol is proposed. The modification (Queue Congestion Adaptive QCA-AODV) is based on queue congestion status of MANET nodes. The results of conducted experiments show improvement in throughput and reduction in both packet loss and end-to-end delay compared with AODV routing protocol.