Task Oriented Load Balancing Strategy for Service Resource Allocation in Cloud Environment

Load balancing strategy is one of the most important issues for service resource allocation in order to balance tasks for different service resources. However, cloud computing has brought about many great changes to the traditional information service pro

  • PDF / 421,392 Bytes
  • 10 Pages / 439.37 x 666.142 pts Page_size
  • 100 Downloads / 194 Views

DOWNLOAD

REPORT


Abstract. Load balancing strategy is one of the most important issues for service resource allocation in order to balance tasks for different service resources. However, cloud computing has brought about many great changes to the traditional information service process when adjusting tasks among different service resources. In this paper, service alliance is introduced into the new information service model, which can provide a new communication mechanism for service providers and service users. Then, a task-oriented load balancing strategy, named double weighted least connection, is proposed. This strategy not only considers the usage of the service resources, but also takes account of the size of different tasks. Furthermore, a set of simulation experiments is discussed in order to evaluate the performance of different load balancing strategies in different situations. Keywords: Cloud computing  Resource allocation  Load balancing  Service alliance  Double weighted least connection

1 Introduction Cloud computing, a new type of information service environment, has been established on the Internet in the recent years [1] and developed by the core technologies including data center, Internet, mobile terminal and so on. Cloud computing also enables outside users to use service resources in a convenient and flexible way that individuals can be allowed to provide resources on-demand on a pay-as-you-go basis [2], like Amazon’s Elastic Compute Cloud (EC2) [3], Google APP Engine [4] and Salesforce CRM [5]. The users in cloud computing can then make use of service resources like using water, power, gas, and other social public service resources. It is thus clear that cloud computing has greatly changed the traditional information service model and information service process. However, information service in cloud computing is a very complicated process [6] which involves the design of service model, the matching of service demands, the allocation of service resources, the evaluation of service quality, as well as the analysis, design and implementation of service system and so on. According to the continuous change of user’s needs and the unceasing increase of service resources, the service resource allocation in cloud computing has become an important issue.

© Springer International Publishing Switzerland 2016 Y. Tan et al. (Eds.): ICSI 2016, Part II, LNCS 9713, pp. 37–46, 2016. DOI: 10.1007/978-3-319-41009-8_5

38

H. Luo et al.

In order to improve the utilization of service resources, cloud computing puts the dispersed service resources together into the resource pool, making a single service application providing services for different users in the multi-tenant way [7]. But when faced with the magnitude of service demands, the unreasonable situation of service resource allocation often arises. Some service resources may be idle and others overloaded. This situation will cause the decreasing of service capacity, even system robustness. Therefore, the load balancing strategy is necessary for service resource allo