Software Defined Network-Based Edge Cloud Resource Allocation Framework

Abstract

A growing number of user applications rely on the computing resources of edge cloud data centers (ECDCs). The efficient management of ECDC resources has a positive impact on the cloud service providers' profitability and on the stringent quality-of-service (QoS) requirements of user applications. This paper proposes a resource allocation framework of interconnected ECDCs using software-defined networking (SDN). SDN technology is used to ensure QoS and to efficiently embed user applications into the ECDCs. The proposed framework, called infrastructure as a service provisioning using SDN (IaaSP-SDN), includes the two-phase coordinated IaaS requests. Provisioning approach and a set of SDN management modules to set up accepted IaaS. The performance of IaaSP-SDN is evaluated in two steps: 1) a prototype is compared with generalized multi-protocol label switching (GMPLS) and 2) the impact of SDN controller physical location on the performance of the IaaSP-SDN framework is evaluated. The results illustrate that the SDN framework is more stable and scalable than GMPLS. We also show that SDN controller location attributes have a significant effect on the acceptance ratio of IaaS requests.

Keywords

• Controller placement
• edge data-centers
• IaaS provisioning
• metro network
• software defined networking