الفهرس 
THE BASIC PRINCIPLES OF NETWORKSOnly 14 pages are availabe for public view
 |  | from | 233 |  |  |
| | | from | 233 | | |
Abstract
Introduction: A computer network is a collection of interconnected computers and devices by channels to facilitate communications among users to share resources and information. Failures have negative effects on computer networks, these effects differ according to the number and the distribution of faults on the network, and the network requires certain time to be recovered, this time is called time of convergence. The research problem: Faults: Failures causes increased time of convergence of the networks and also causes more traffic delay. Aim of the study: Reduce the time of convergence of routers and the traffic delay due to the separate and nested faults. Study procedures: The thesis presents three different mechanisms for tolerating multiple network failures, the first mechanism uses Route Information Protocol (RIP), while the second mechanism uses Open Shortest Path First (OSPF) and the third mechanism is a proposed networking area technology. In these three described mechanisms, links are re-employed on the basis of graph theory which states that there is always a spanning tree in any connected graph. In response to each network failure, these mechanisms will be able to find a unique spanning tree embedded in the resulting network. The spanning tree mechanism is able to deal with different topologies, either regular or irregular. Two examples of each case are studied under different number of failures with different distributions on the topologies. Separate failures and nested failures are applied. The thesis Conclusions: This mechanism is suitable for combating large number of faults and introduces less time of convergence than the normal structure of the network regardless their distribution inside the network. This proposal exhibits a number of characteristics that make it extremely effective. This mechanism is able to carry out an automatic reconfiguration. Reconfiguration time is short enough to be tolerated by most applications. The introduced mechanism self-stabilizes after the finite time. Measurements of reconfiguration times and performance are both excellent.