Performance-aware cost-effective brokering and load balancing algorithms for data center in large scale cloud computing

The cloud computing transforms computing services into "as a service" form. It helps organization to reduce computing infrastructure cost. Incloud computing concept, cloud users can use computing resources according to their needs and requirements. Customers are able to scale hardware, software and application platform through Service Level Agreements (SLAs) in the cloud. In recent years, there has been an increasing interest in cloud computing among service providers and cloud users. To provide services, different cloud service providers build their own computing platform differently due to the lack of a common standard. From Day to day it becomes very challenging to select an appropriate provider considering the specific user requirements. Besides the standardization service, one of the most significant current discussions in cloud computing is the cloud brokering service. Cloud brokering is having a positive impact in choosing an appropriate provider along with the capability to handle cloud-to-cloud communication. Cloud broke ring is an intermediate negotiator between users and service providers. This negotiator helps users to select appropriate provider as their requests. . Questions have been raised about the efficiency of cloud brokering in various aspects of user requirements, such as cost, timeliness, or service performance. However, there has been little discussion about efficient cloud brokering services. We studied how brokering algorithms improves brokering performance. Through our research we found that cloud brokering algorithms and service load balancing algorithms able to improve brokering performance. The aim of this research is to propose a load balancing algorithm and propose cloud brokering algorithms in order to improve brokering performance. Proposed cloud brokering algorithms works with different types of cloud provider and deal with various user requirements. Proposed Cost Aware algorithm minimizes approximately 5.5% cost compared with closest data center algorithm. However, data center processing time response time was greatly increased. Further we developed Load Aware algorithm which minimizes average DC processing time and maximum DC processing time by 73% and 49% respectively. In order to make our proposed method cost efficient, we developed Load Aware Over Cost algorithm which is 8% cost effective comparing with Load Aware algorithm. This algorithm improves average response time by 43%. For all algorithm combinations our proposed State Based Load Balancing algorithm minimizes both processing time and response time.

Text FSKTM 2015 46 - IR.pdf Download (1MB)

Actions (login required)

View Item