Optimized Scalable SFC Traffic Steering Scheme for Cloud Native based Applications

Abstract

Network Function Virtualization (NFV) has already proven its efficiency to deploy networking services in large-scale. Recent advances of cloud-native applications may bring new advantage by deploying and implementing Virtual Network Function (VNFs) as cloud-native Containers rather than virtual machines. Beside remarkable advantages such as lower overhead and faster running, microservices (cloud-native containers) intend to save costs while increasing the service agility. To this end, in this paper we extend consolidated state-of-the-art tools and technologies developed in two domains cloud-native applications and Network Function Virtualization (NFV). The proposed framework chains services provisioned across Kubernetes and Contiv/VPP domains and using containers. Our orchestration framework chain services across distributed CNFs. Furthermore, we propose K -TS scheme to load balance the traffic over services replicas. K -TS is based on Ketama Consistent hashing algorithm. Experimental simulations show very good results for both the service chaining framework in term of QoS satisfaction such as: packet error rate, throughput satisfaction and jitter.