RSFP: Reliable Service Function Placement with Optimized Resource Consumption in Data Center Networks
Hong Tang, Ao Ku, Wei Liu, Haojun Huang · 2020
With the development of Network Function Virtualization (NFV) in Data Center Networks (DCN), network functions can be virtually realized and decoupled from the traditional physical middleboxes. By proper the resource management of NFV placement, it can provide flexible Service Function Chain (SFC). Among the existing NFV placement polices, the negative effects of temporal error or failure of NFVs are not fully considered, which leads to unreliable SFC services. In this paper, we focus on the reliability guarantee, and propose a new policy, namely Reliable Service Function Placement (RSFP). It not only takes the reliability of service chain path into account, but also find the minimized resource consumption along the path. Simulation results show that, compared with the classical greedy policy, RSFP increases the reliability of SFC path and reduces the average resource utilization ratio.