Design and analysis of distributed load management: Mobile agent based probabilistic model and fuzzy integrated model

Abstract

In large-scale distributed systems, performing load monitoring and load balancing is a challenging task in terms of load management. In order to enhance the overall system performance, we have developed and implemented two different models for large-scale distributed load management. The mobile agent-based system is based on a probabilistic normed estimation model. This model uses mobile agents for collecting the instantaneous status of currently available node resources autonomously. The mobile agent is goal oriented and consumes less network and system resources, which is ideal for load monitoring for large-scale distributed systems. Moreover, we have proposed an integrated load balancing and monitoring model for distributed computing systems employing type-1 fuzzy logic. Furthermore, we have proposed a smooth and composite fuzzy membership function in order to model fine-grained load information in a system. In this paper, a detailed software architectural design for mobile agent based load monitoring system as well as the fuzzy-based load balancing approach are presented. The experimental evaluation is presented to compare the behavior and performance of the mobile agent-based probabilistic model and fuzzy integrated model under different load conditions. A detail comparative analysis is presented for the mobile agent-based probabilistic model and fuzzy integrated model to show the performance and efficiency of each model. In this paper, we have computed cross-correlation to find the relation between our proposed models (FIM and MABMS).