Abstract:
The problems of congestion and faults have been major menace in cloud computing. To solve these issues, the concept of load balancing and fault tolerance were introduced. Load balancing is essential for dynamic, fairly distribution of loads across all network nodes. However, fault can occur, a node failure can lead to denial of service. A fault tolerant cloud is one that has the ability to respond to an unexpected failure. Many researchers have worked on these issues and many algorithms have been proposed in order to find solution to these. This research is meant to improve load balancing and fault tolerance in cloud computing. To achieve these, Ant Colony optimization Algorithm (ACO) was used. ACO is an algorithm based on the behavior of an ant. The ants’ trail laying ability was modeled to achieve load balancing. For fault tolerance, job migration technique was used; this technique ensures that loads on a particular node are migrated to another if such node should fail. The design was simulated on CloudSim simulator. The result shows 73% processor utilization which is an upgrade on the 56% recorded when round robin algorithm was used. Also, the research algorithm was evaluated against the algorithm proposed by Nguyen and Tran (2017). The result shows that this algorithm is more efficient as it recorded a response time of 1.37 seconds compare to 1.69 seconds achieved by the Nguyen and Tran (2017) algorithm. Also, the algorithm was evaluated based on time complexity against the algorithm proposed by Divaya and Farhat (2017). Result shows that the proposed algorithm was more efficient as it was able to achieve a linear time complexity compared to quadratic time complexity achieved by Divya and Farhat (2017). In conclusion, the result obtained from this research shows that the model has proven to be an efficient means of providing load balancing and fault tolerant system for cloud computing.
