Master Thesis - 2013

In conventional service platforms, service control function and network control function work independently. However, it is desirable that these control functions are integrated and work in cooperatively in order to satisfy the required service quality. To achieve this, we have proposed a service platform which integrates service control and network control by componentization of them. In this platform, a provided service is realized as a combination of service/network components. Therefore, load balancing of these components is important in order to guarantee the quality of provided services.

In this study, we aim to decrease both the call blocking probability of services which require constant bandwidth and the processing time of services which get computational resources based on Processor Sharing. To achieve this load balancing, I propose a load balancing method considering both computational and network resources. In this method, the platform estimates the future call blocking probability and the future processing time from the access frequency of service requests. Moreover, the platform assigns computational/ network resources to each service in order that these estimated values keep the predetermined values by dynamic rearranging of components.

In this thesis, the effectiveness of the proposed method is evaluated by simulation experiments.