Cloud-based architecture for deploying ultra-high-definition media over intelligent optical networks

Abstract

A key driver for the evolution of Future Media Networks (FMNs) is the emergence of beyond High Definition (HD) media formats. These formats impose far greater demands on networks for high-capacity, low latency and stringent Quality-of-Service (QoS) compared to other existing formats. In addition, their data-intensiveness will require real-time interconnection of multiple, possibly distributed, high performance media processing and storage resources. To cater to this, novel networked architectures are required. In this paper, we propose a cloud-based architecture for the deployment of FMNs over intelligent optical networks. The key requirements of FMNs are identified and we discuss the features of our proposed architecture which meet these requirements. We formulate Integer Linear Programs (ILP) which model Non-adaptive and Adaptive Media- and Network-aware schedulers for the implementation of media streaming applications over our architecture. These models aim to maximize the number of jobs provisioned while ensuring the serving network is never overloaded. Results obtained from evaluations show the adaptive service implementation outperforms its non-adaptive counterpart in terms of number of jobs provisioned and overall Quality-of-Experience (QoE) offered by the infrastructure.