A Quantitative and Qualitative Comparison of Distributed Information Processing Using Mobile Agents Realised in Rmi and Voyager

Abstract

We review mobile agents in the context of distributed object computing and parallel processing. We compare these three paradigms qualitatively. For a quantitative comparison of RMI and Voyager as mobile agent platforms, we identify distributed information processing with flexible load balancing as a convincing application to evaluate the two platforms.

Analysis on the performance of mobile agents for query retrieval

“…However, our proposed query retrieval approach is different compared with the previous approaches [15][16][17]19,20,23,18] object cloning, dispatching, and disposing. This is done by combining a hierarchical and a parallel dispatching of mobile agents to the LP servers in order to retrieve the query results which has been applied in the EHQR approach as described in Section 4.6.…”

“…We try to minimize the number of agents used to fetch the query results from the remote hosts in order to reduce network overload. It is different compared with the previous approaches [15][16][17][18][19][20]23,24,[26][27][28] by utilizing the mobile agent properties such as object cloning, dispatching, and disposing in the EHQR approach. This is done by combining a hierarchical and a parallel dispatching of mobile agents to the local prefetch servers in order to retrieve the query results in the MaSS.…”

“…However, the mobility patterns of mobile agents are only based on the simple approaches such as a sequential client-server approach, a sequential mobile agent approach, a parallel client-server approach and a parallel mobile agent approach. Also, Cogan et al [20] have analyzed the performance of mobile agents based on different platforms that are Voyager mobile agent platform [21] and Java Remote Method Invocation (Java-RMI) platform [22]. However, from the experiments, only the mobility and remote method invocation on mobile agents applications have been compared.…”