ABSTRACT
Multimedia applications also bring new challenges to current generalpurpose cloud computing. Nowadays, the general-purpose cloud employs a utility-based resource management to allocate computation resources (e.g., CPU, memory, storage, etc.). In utility-based resource allocations, cloud resources are packaged into virtual machines (VMs) as a metered service. By using VMs, cloud resources can be provisioned or released with minimal eorts. As cloud providers, such as Amazon Elastic Compute Cloud (Amazon EC2) [5], the only guaranteed parameter in the service-level agreement (SLA) is the resource availability, that is, users can access rented resources at any time. However, for multimedia applications,
in addition to the computation resources, another important factor is the stringent quality-of-service (QoS) requirements in terms of service response time, jitter, and packet loss. If the general-purpose cloud is used to deal with multimedia applications without considering the QoS requirements, the media experience may become unacceptable to users. Hence, an eective resource allocation scheme is strongly needed for multimedia cloud computing. Recently, a lot of research eorts have been made on the resource allocation and QoS provisioning for cloud-based multimedia applications [6-26].
