Cost-aware VNF placement in cloud infrastructures
Nadir Ghrada · Espace ÉTS (ETS) · 2018
The concept of Network Function Virtualization (NFV) has been introduced as a new paradigm in the recent few years. NFV offers a number of benefits including increased maintainability and reduced deployment overhead. Several works have been done to optimize deployment (also called embedding) of virtual network functions (VNFs). However, no work to date has looked into optimizing the selection of cloud instances for a given VNF and its specific requirements. Furthermore, none of the existing proposals have addressed VNF placement when multiple instances are required to be provisioned. In this thesis, we evaluate the performance of VNFs when embedded on different Amazon EC2 cloud instances. Specifically, we evaluate three VNFs (firewall, IDS, and NAT) in terms of arrival packet rate, resources utilization, and packet loss. We then formulate the VNF placement and chaining using integer linear program (ILP) and propose two heuristics aiming at minimizing the operational and synchronization costs taking into account the discrepancy of energy costs between the different hosting locations. Our results indicate that the performance of the VNFs varies across different instance types and that CPU is the critical resource for the tested VNFs. We also show, through extensive simulations, that the two proposed service chain embedding heuristics succeed to efficiently minimize the cloud provider’s operational cost as well as the synchronization cost between VNF instances.