Abstract
Network Function Virtualization (NFV) decouples network functions from hardware, improves the flexibility of resource allocation, and enhances network scalability. Any failures of software and hardware in an NFV environment will result in the interruption of Service Function Chains (SFCs). As one of the key technologies of 5G, NFV has more stringent delay and reliability requirements for services. This paper takes software and hardware failures into account, and proposes a reliability-aware service function chain backup protection (RABP) method to meet SFCs' high reliability and low latency demands. First, we formulate a mixed-integer linear programming model to maximize the revenue-to-cost ratio. Then, we propose a heuristic algorithm for suboptimal solutions. SFC deployment is divided into two stages: primary virtual network function (VNF) deployment and backup VNF deployment. The reliability enhancement and delay optimization deployment method is presented to deploy primary VNFs. A breadth-first search method is used to consolidate primary VNFs as much as possible to improve reliability of SFCs and reduce transmission delay. When deploying backup VNFs, a resource-efficient backup selection method is presented to reduce backup resource consumption while meeting reliability demands. Finally, experiment results show that the proposed RABP method not only has the best acceptance ratio and long-term average revenue to cost ratio, but also reduces backup resource consumption and transmission delay.
Highlights
In recent years, network service demands have increased dramatically
We take the reliability of virtual network function (VNF) and server nodes into account, and propose a reliability-aware service function chain backup protection (RABP) method
It improves the reliability of Service Function Chains (SFCs), and reduces resource consumption and transmission delay by consolidating VNFs and hop constraints
Summary
Network service demands have increased dramatically. For traditional networks, software and hardware are tightly coupled, and specific network functions (e.g., firewalls, intrusion detection systems and network address translation) run on dedicated hardware, which significantly increases capital and operating expenditures. Most current works assume that software and hardware are completely reliable, and they study the ability of efficient SFC deployments to improve the revenue of network operators. Wang et al [16] take software and hardware failures into account and propose a reliability model Since both types of failures can cause service interruptions, this paper proposes new backup methods for software failures and a resource-efficient backup selection method considering software and server node failures. We take the reliability of VNFs and server nodes into account, and propose a reliability-aware service function chain backup protection (RABP) method. (ii) We present a reliability enhancement and delay optimization (REDO) deployment method to deploy primary VNFs It improves the reliability of SFCs, and reduces resource consumption and transmission delay by consolidating VNFs and hop constraints.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
