Optimized hybrid service brokering for multi-cloud architectures

Abstract

A cloud computing platform provides access to shared resources along with diverse services including computation and storage to its users. The ubiquitous access to resources requires the service providers to ensure an efficient, reliable, and a fault-tolerant infrastructure. In this context, the cloud brokering mechanisms enable the cloud computing platforms to manage cloud resources through mediation between cloud service providers and cloud users. Corresponding to user requests, the cloud service brokering and load balancing aim at efficient real-time provision of services with minimal monetary cost through selection of appropriate data centers and virtual machines. This paper proposes a normalization-based hybrid service brokering approach integrated with throttled round-robin load balancing to improve resource management through cost- and performance-aware provision of cloud services. The proposed approach incorporates a hybrid (static and dynamic) evaluation criteria using normalization for determining the impact of cost and performance-oriented parameters in a multi-cloud environment. The subsequent selection of the most appropriate service provider along with throttled round-robin load balancing optimizes cloud resource management. The experiments performed with diverse number of user bases and data centers show that the proposed cloud service brokering approach outperforms other well-known approaches by improving response time, data center processing time, and monetary cost up to 17.39%, 31.35%, and 7.06%, respectively.