Optimized Availability-Aware Component Scheduler for Applications in Container-Based Cloud

Cloud Computing is a popular platform for providing online computing services to the tenants. With migrating large number of the applications to the cloud, application availability is raised as a concern issue. Availability is a non-functional requirement that refers to the percentage of time the application service is available for the use. Many factors can impact the application availability especially in the cloud environment. In this paper, we show the impact of scheduling (placement) of the application's components inside the Data Center of the cloud on the application availability. We propose an optimized Availability- Aware strategy to schedule the components with the goal to maximize the availability of the application while satisfying several constraints. We map the scheduling problem to Integer Linear Programming (ILP) optimization model. We compare the Availability-Aware strategy with other scheduling strategies. The results show that the Availability-Aware strategy achieves higher application availability compared to the other strategies.

I. Introduction

Recently, the Cloud Computing (CC) is considered the most attractive platform for many of the Application Service Providers (ASPs). Cloud allows ASPs to get online access for virtual infrastructure resources such as Virtual Machines (VMs) and containers to deploy their applications at any time from anywhere. ASP gets resources from the Could Service Provider (CSPs) according to Service Level Agreement (SLA) between them. Although the cloud is a suitable environment to host many of the applications services, the Quality of Service (QoS) such as service availability is raised as a real concern issue. Many of ASPs expect their applications services to be available during their lifetimes. Application availability refers to the duration of the time the application is fault tolerant and available for the end user [1].