A Fuzzy Logic Approach to Performance Analysis and Grading Through Benchmarking of Load Balancers

It is found that evaluation of performance of various load balancing algorithms is done by many researchers on the basis of some individual metrics by simulation. But it is found that a load balancing algorithm which is better on the basis of one metric may be worse on the basis of another metric. The research paper introduces a grading mechanism rooted in Fuzzy Logic to evaluate various load balancing algorithms in cloud computing settings. This can rank the prevailing Load Balancing Algorithms. This study aims to assess the efficacy of several widely-used load balancing algorithms through select metrics, subsequently assigning them a composite grade based on these metrics. Utilizing simulations, the performance of different algorithms is gauged under diverse load scenarios. These simulation outcomes are then analyzed, culminating in a performance-based grade for each algorithm. The technique can assess the composite efficacy of any Load Balancing Algorithm on the basis of given selected matrices, when all the metrics may or may not be equally important. The analysis provides insights into the algorithms' efficiency, aiding in the identification of the most proficient algorithms for cloud computing. Five algorithms - Weighted Round Robin (WRR), Round Robin (RR), Equal Load, Least Connection, and Random - are evaluated using five metrics: Average Response Time, Throughput, Transaction Rate, and concurrency. The study evaluates various load balancing algorithms, treating all metrics as equally important, although acknowledging that they can be weighted differently. The results, sourced from Software-Based Data Centre Networks (SDN), show that Weighted Round Robin is the top performer with a grade of 7 out of a possible 10. Round Robin and Equal Load both score 6, Least Connection scores 5, and Random has the lowest score of 4. The paper highlights gaps in current research and suggests the need for developing a load balancing algorithm that can achieve a perfect grade.