Realistic traffic generation capability for SAAM testbed

Abstract

Traffic modeling is an important component of the design of any communication network. This is even more crucial emerging networks, which are expected to operate in high speed and high bandwidth environments. As the design of a network depends to a great extent on the types of traffic it is expected to carry, it is essential to characterize the traffic that a network expected to carry. This is where traffic models are very important. They can be used to produce artificial traffic input that exhibits essential characteristics of real network loads. This thesis describes a design and implementation of a general- purpose traffic generator for a test bed of the Server a Agent Based Active Network Management (SAAM) architecture. The traffic generator is easy to use and implements the four most important traffic models (Constant Bit Rate (CBR), Poisson, Packet-Train, and Self-Similar). With this traffic generator, SAAM project now has the capability of simulating and testing the system components in more accurate and more realistic environments.