Reliability of Mobile Agents for Reliable Service Discovery Protocol in MANET

TL;DR

This paper presents a Monte Carlo simulation-based algorithm to estimate the reliability of mobile agent systems in MANETs, considering node mobility, link failures, and radio propagation effects, to improve service discovery dependability.

Contribution

It introduces a novel reliability estimation algorithm for mobile agent systems in MANETs that accounts for mobility patterns and link failures, enhancing service discovery dependability.

Findings

The proposed algorithm effectively estimates route reliability.

Mobility patterns significantly impact MAS performance.

Optimal network bandwidth for agent support is identified.

Abstract

Recently mobile agents are used to discover services in mobile ad-hoc network (MANET) where agents travel through the network, collecting and sometimes spreading the dynamically changing service information. But it is important to investigate how reliable the agents are for this application as the dependability issues(reliability and availability) of MANET are highly affected by its dynamic nature.The complexity of underlying MANET makes it hard to obtain the route reliability of the mobile agent systems (MAS); instead we estimate it using Monte Carlo simulation. Thus an algorithm for estimating the task route reliability of MAS (deployed for discovering services) is proposed, that takes into account the effect of node mobility in MANET. That mobility pattern of the nodes affects the MAS performance is also shown by considering different mobility models. Multipath propagation effect of radio signal is considered to decide link existence. Transient link errors are also considered. Finally we propose a metric to calculate the reliability of service discovery protocol and see how MAS performance affects the protocol reliability. The experimental results show the robustness of the proposed algorithm. Here the optimum value of network bandwidth (needed to support the agents) is calculated for our application. However the reliability of MAS is highly dependent on link failure probability.