Availability-Aware Strategy for Service Upgrade in Cloud Computing Platform

Cloud computing has become the main provider for modern IT computing infrastructure, offering economical and scalable on-demand services. Service availability is a non-functional requirement, which refers to the success rate of providing services to end-users over a specific period. However, service upgrades pose significant challenges regarding service availability, resource utilization, and dependency management. A poorly planned upgrade can have severe consequences, including service outages and violations of Service Level Agreements (SLAs). In this paper, we propose an Availability-Aware Service Upgrade Strategy that balances service availability with resource utilization while considering dependencies between services. We modeled the service upgrade scheduling problem as an Integer Linear Programming (ILP) optimization problem, aiming to minimize the service upgrade execution time. Given the complexity of ILP model, we introduce a greedy heuristic approach that dynamically schedules service upgrades to optimize execution time while ensuring the availability of the services. We validate our proposed strategy through simulations of different scenarios, comparing its performance with the traditional sequential approach. The results show that our strategy achieves substantial improvements in the execution time of the service upgrade and reduces overall costs.