Smart workload migration on external cloud service providers to minimize delay, running time, and transfer cost

With limited resources, overutilization is a challenging issue in the emerging system of smart devices, resulting in higher costs, poor performance and low prices due to service level agreement (SLA) violations. These factors extremely annoy the customers as well as providers. External resources are hired to overcome the resources scalability challenges; however, external resources cause higher delay, running time, and transfer cost. The lift in delay explicitly means more cost and more customers dissatisfaction. The delay and transfer cost increases with geographical distance and overutilization. In this article, a mechanism is proposed to make efficient migration decisions to external cloud service providers (CSPs) to minimize the delay, running time, and transfer cost by searching an optimum data center (DC), where the resources may be taken with optimum conditions. The Cloud Analyst is extended to simulate the proposed framework. Results were calculated for three different phenomena, where hiring the resources from underutilized DC decreases the running time to 0.237 ms; however, the response time and transfer cost increases to 499 ms and 0.228 $, respectively; similarly, getting the resources of nearest DC drops the delay time and transfer cost to 80 ms and 0.065 $ respectively; however, the execution time increases to 0.500 ms. The proposed framework optimized the delay time, execution time and transfer cost to 50 ms, 0.237 ms, and 0.065 $, respectively.

Automated summary

This article proposes a mechanism for making efficient migration decisions to external cloud service providers to minimize delay, running time, and transfer cost. Results show that using resources from different data centers in various scenarios can reduce running time, but may also increase response time and cost.