DEVELOPMENT OF A MOBILE AGENT SYSTEM FOR MONITORING SOFTWARE RESOURCES IN A NETWORK ENVIRONMENT

Abstract

Over the years, searching, detecting and monitoring on the network often involve the physical movement of the network administrator or user from one computer to another with sequential, serial or procedural diagnostics of the resources. This approach is not only stressful but introduces various degrees of delay in searching, detecting and monitoring of resources and events on the network. The introduction of diagnostic software packages has not in any way caused a change in this regard because they are static in nature. This research presents the use of autonomous, dynamic, intelligent and structured utility software as a monitor of software tools in a network environment. The architecture of the agent-based utility software is characterized by the relational database of the computer network software resources and procedure for dispatching and receiving of the mobile agent. The monitor of software tools on the network is modeled using C-sharp (C#) Programming language. A case study of a computer network environment of Federal Polytechnic, Ede, Osun State, Nigeria which comprises of Microsoft Windows NT as Platform, Microsoft Access database as back-end engine, C-sharp as front-end engine has been carried out. At the source, the mobile agent is dispatched with dedicated assignment (Monitor of software tools) to the target workstation where it interacts with the host operating system and its appendages or utility programs for network monitor of software tools. The results obtained by the mobile agent after each successful visit to a set of target workstations are analyzed. To demonstrate a superior scheme provided by the proposed research over the conventional scheme using Remote monitoring (RMON) scheme, simulation was adopted and the parameters used are: Cost of service against number of requests per service, query time against number of requests per service and service delay overhead against number of requests for service. The results from the simulation were analyzed and interpreted using tables and line graphs for decision making. The outcome shows that the agent based remote monitor of software tools proved a cost effective solution than RMON as the number of requests per service increases. Secondly, it optimizes query times as the number of requests per service increases because it is able to eliminate repeated requests and thereby reduces the number of queries at the remote workstation. Finally, it generates a lower service delay overhead than RMON scheme for about 18 out of 20 sample runs thereby performing better, resulting to 90% efficiency. It can therefore be concluded that mobile agent-based scheme for remote monitor of software tools is found to be an improved, efficient and dynamic management scheme over the conventional method using Remote Procedure Calls.