Virtualization of network functions combined with the cloud paradigm enables the derealization of mobile network functionalities in data-centers (DC). Such virtualized mobile network functions (MNF) can be then provisioned dynamically and flexibly, leading to more scalable and cost-effective mobile services. Although advantageous, the provisioning of mobile network functions imposes a significant burden on the DCs and on the transport network. Furthermore, derealization is expected to increase the latency. This paper focuses on the problem of provisioning on-demand resources in the DCs (storage/computational resources) and in the flexible optical network (network resources) interconnecting the DCs, in order to support MNFs. An orchestrator is proposed for handling the DC and network resources in a unified way. Two orchestration strategies targeting either the minimization of the latency or the blocking probability are presented and assessed through simulation. High performance improvement is achieved through the use of the unified orchestrator.
An Orchestrator of Network and Cloud Resources for Dynamic Provisioning of Mobile Virtual Network Functions
Martini B;
2016-01-01
Abstract
Virtualization of network functions combined with the cloud paradigm enables the derealization of mobile network functionalities in data-centers (DC). Such virtualized mobile network functions (MNF) can be then provisioned dynamically and flexibly, leading to more scalable and cost-effective mobile services. Although advantageous, the provisioning of mobile network functions imposes a significant burden on the DCs and on the transport network. Furthermore, derealization is expected to increase the latency. This paper focuses on the problem of provisioning on-demand resources in the DCs (storage/computational resources) and in the flexible optical network (network resources) interconnecting the DCs, in order to support MNFs. An orchestrator is proposed for handling the DC and network resources in a unified way. Two orchestration strategies targeting either the minimization of the latency or the blocking probability are presented and assessed through simulation. High performance improvement is achieved through the use of the unified orchestrator.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.