Multi‐objective container scheduling and multi‐path routing for elastic business process management in autonomic multi‐tenant cloud

Abstract

SummaryCloud multi‐tenancy has a variant requirement, due to its resource sharing nature, satisfying such requirements and maintaining a balance between the resources and the business workloads of multiple tenants is a challenging task, and also the communication between scheduled containers leads to high power consumption. To address these issues, this article proposes an autonomic approach to ensure the elasticity of BPM in multi‐tenant cloud. Where it employs the autonomic computing capabilities for scheduling the containers into the available servers then regulates the communication between the containers using multi‐path routing. For the container scheduling, a multi‐objective crow search optimization algorithm is proposed to schedule the containers into appropriate servers. Then, the discrete wolf search algorithm based multipath routing is proposed to route the communication flows between the containers by finding the optimal path with an objective to minimize the energy consumption. The optimal path is constructed as a multi‐tenancy graph with bandwidths determining the shortest distance between the servers and containers. The overall simulations shows that the proposed algorithm outperformed the other compared approaches in terms of make‐span, resource utilization, execution cost, execution time, and energy consumption.