Title :
Evaluating recovery protocols for mobile agents in network management applications
Author :
Bruce, S. ; Linton, D.G.
Author_Institution :
Sch. of Electr. Eng. & Comput. Sci., Central Florida Univ., Orlando, FL, USA
Abstract :
Reliable mobile agents (MAs) should be able to handle events that impede their normal operation, such as host or server failures. In this paper, we investigate four recovery protocols which support reliable mobile agents and have applications in network management. For each protocol, analytical models are developed for average response time (the time from when an MA is launched until it has visited or attempted to visit all of the hosts on its itinerary) and expected coverage (the number of hosts from the agent´s itinerary it is able to visit and complete its task). Simulations are used to verify the analytical models for response time and coverage. The results from the verified analytical models are used to recommend the recovery protocols that best support a low response time and a high coverage under different network management applications. Simulation output using non-exponential repair laws lead to the same conclusions gleaned from the exponential repair law models
Keywords :
computer network management; computer network reliability; protocols; software agents; average response time; exponential repair law models; mobile agents; network management applications; nonexponential repair laws; recovery protocols; Analytical models; Application software; Computer network management; Computer science; Delay; Engineering management; Intelligent networks; Mobile agents; Network servers; Protocols;
Conference_Titel :
Integrated Network Management Proceedings, 2001 IEEE/IFIP International Symposium on
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-6719-7
DOI :
10.1109/INM.2001.918010
