Title :
Performance based reliability optimization for computer networks
Author :
Ahuja, Sanjay P.
Author_Institution :
Dept. of Comput. Sci. & Inf. Syst., Grand Valley State Univ., Allendale, MI, USA
Abstract :
This paper addresses the important problem of allocating capacities to the links of a computer network such that the performance based reliability is maximized and a cost constraint is satisfied. The performance and reliability of a computer network depends both on the reliabilities and capacities of the communication links and the computing elements. Performance based reliability oriented capacity allocation scheme is developed for computer networks. In this scheme capacities are allocated to the links of the network such that the performance based reliability is maximized using the genetic algorithm (GA). The overall effectiveness of a computer network can be measured in terms of the previously developed measure, the average network throughput. This measure computes performance based reliability from the network connectivity point of view. To demonstrate the effectiveness of the GA based approach, the results obtained using the GA are compared with optimal solutions obtained by an exhaustive search of the problem state space
Keywords :
channel capacity; computer network reliability; genetic algorithms; performance evaluation; search problems; state-space methods; telecommunication links; average network throughput; capacity allocation; communication link capacity; communication link reliability; computer networks; computing elements; cost constraint; exhaustive search; genetic algorithm; network connectivity; optimal solutions; performance based reliability optimization; state space; Channel capacity; Computer network reliability; Computer networks; Cost function; Genetic algorithms; Network topology; Telecommunication network reliability; Telecommunication traffic; Throughput; Upper bound;
Conference_Titel :
Southeastcon '97. Engineering new New Century., Proceedings. IEEE
Conference_Location :
Blacksburg, VA
Print_ISBN :
0-7803-3844-8
DOI :
10.1109/SECON.1997.598624