Dynamic Capacity Planning with Comprehensive Formation of SLAs in Clouds

Abstract

The provision of service enabled connectivity is the significant art of clouds. The long-held dream of computing as a utility has become reality in the era of cloud computing. Cloud users are now able to run and access their applications from anywhere in the world on demand. The proposed research considers dynamic capacity planning for cloud systems. The aim is to dynamically adapt computing capacity such that Service Level Agreements (SLAs) are continuously met while minimizing the total costs incurred in running the cloud services. However, research in this regard is still at its infancy. Scalability, resource heterogeneity, workload dynamicity, resource sharing and virtualization are the main challenges that need to be overcome to have effective and trustworthy schemes for capacity management, that play a vital role in cloud computing. The work in this approach is based on developing a capacity planning scheme to ensure that high-level performance targets (SLAs) are continuously met. The scheme applies threshold-based techniques to ensure meeting the SLAs while minimizing the total incurred costs. The approach is designed to work on cloud environments and hence must address these environments specific challenges.

Computationally efficient approximate dynamic programming for multi-site production capacity planning with uncertain demands

“…(Pehlivan et al 2014) proposed to integrate a service level constraint to dynamic capacity planning and location decisions. Dynamic capacity planning has also seen some development for cloud services (Alasaly et al 2015). Dynamic capacity planning using simulation models was the first time investigated by (Atherton 1989).…”