Performance Analysis of a Hybrid Approach to Enhance Load Balancing in a Heterogeneous Cloud Environment

Abstract

Cloud computing is rapidly becoming more advanced. It has become more attractive due to its cloud potentials of being easy to use and anywhere accessibility in comparison to other technologies. Load balancing is a very important component of efficient operations in the cloud computing environment. Several algorithms have been designed but yet unable to rectify the holes found in the performance area like minimizing response time, processing time and cost. In this work, it is proposed a hybrid approach that is devoted to enhancing performance for the cloud system user putting the concepts of enhanced Throttled and Equally spread current execution algorithms(ESCE) into use. The proposed hybrid algorithm was implemented in cloud analyst environment and the results obtained were analyzed. The performance of the proposed hybrid algorithm was evaluated with the existing algorithms ESCE and the improved throttled using the response time, data processing time, cost, physical memory and storage capacity. The results of the evaluation analysis show that with network delay, physical memory and the storage capacity, the Performance differences in various scenario vary, where scenario 1 is 10.5 (ms) average response time and processing time, $ 2.10 total cost. Scenario 2 ,5.6 (ms) response time and 8.5(ms)processing time $2.03 cost but the ESCE recoded best minimum response time of 40.30(ms). Scenario3,8.3 (ms) response time, 3.5 (ms)processing time,$2.00 cost. scenario 4 of 0.31(ms) response time, the processing time of 2.41(ms) and cost of $2.14 but the improved throttled is best in terms of minimum processing time with 2.00(ms).scenario 5,2.3 (ms) response time ,processing time 1.1 (ms)and cost of $2.00. Considering the physical memory, the fixed physical memory provided a better average response time of 0.50(ms) than the varied physical memory, and the varied physical memory provides a better maximum processing time of 5.03(ms). In terms of storage capacity, high storage capacity provides a better response time of 320.02 (ms). This work recommends the adaptation of hybridized improved Throttled and ESCE algorithms by cloud service providers so as to get better performances.