Scheduling of Containerized Resources for Microservices in Cloud

Most developers consider that microservice-based application design and development can improve scalability and maintainability. The microservices are developed as small independent modules and deployed in containers. The containers are deployed in virtual machines (VMs), which in turn run in hosts. Effective consolidation of the service requests to the containers may reduce the number of active hosts in a cloud environment, resulting in lesser power consumption of the cloud data centers. This research aims to maximize the resource utilization of the hosts by effectively allocating the containers to the VMs and VMs to the hosts. In this scheduling, a few additional containers and VMs are kept in the available resource pool so that during peak demand for services, the users get their service at the earliest (preferably without any delay). This paper presents a heuristic algorithm for microservice allocation in a containerized cloud environment to achieve these objectives. The performance of the proposed algorithm is validated and justified through the extensive experimental results. We have compared the performance of the proposed technique with the existing state-of-the-art. The number of container deployments in the proposed policy is reduced by 12.2–17.36% compared to the Spread policy and 6.13–10.57% compared to First-Fit and Best-Fit policies.