A Framework for the Modeling and Evaluation of the Mobile Agent in Network Fault Management

Abstract

Network fault management is one of the most important functions of network management. Good network fault management can enhance network availability and reliability, increase network uptime. With the increasing of network size, complexity and heterogeneity, the traditional centralized configured SNMP (Simple Network Management Protocol)-Based C/S(Client/Server) network management paradigms, the network fault management of them in particular, are facing a lot of problems, such as availability, flexibility and scalability of network, traffic bottleneck and so on. MAs (Mobile Agents), as a reasonable new potential technology, can improve these deficiencies and work better to achieve distributed network management, and several researchers have studied these areas. In this paper, we build up a general analytical framework to analysis and quantitatively evaluate the performances of the MAs (Mobile Agents) and the traditional SNMP-Based C/S approaches under different scenarios. We introduce how to make an analytical model in detail. We present two combination models of SNMP protocol and Mobile Agent technology, and introduce some formulas to assess the related performance and verify their performance by experiments. The results show that our proposed models are reasonable and available. Our work provides a favorable reference that analyses and evaluates the performance of a network fault management model based on mobile agent technology and the traditional SNMP protocol in the future.