Reliability-Aware Multi-Source Multicast Hybrid Routing in Softwarized Networks

Abstract

Undoubtedly, these days our telecommunication networks are witnessing not only a major spike is data volumes, but also a shift in the mode of communications. Employees, news anchor and students are conducting their daily business and learning activities through online platforms as they shelter homes during this pandemic and this is expected to continue for some time. An overwhelming shift to one-to-many and many-to-many communications is observed and end users expect from their providers efficient, secure and reliable services. Operators of digital platforms are challenged to respond quickly to the rising demand, by enhancing deployability and manageability of their service. Virtualization is a key enabler for enhanced deployability and manageability where virtual functions can be automatically deployed on demand. Another challenge that providers deal with is the individualized requirements by services offered to users which may vary between high reliabilities, low latency, robust security and any combination thereof. This paper considers the problem of provisioning multi-source multicast services where each service consists of a set of in-network virtual functions that must be chained in a particular order to meet the quality of service demanded by end users. We deal with a reliable service where reliability is attained by provisioning backup functions for the service. We first calculate the requirements of VNF backups which account for fewer computing resource consumption. Next, we formulate the multi-source multicast hybrid routing as a Mixed Integer Linear Programming (MILP) and find a solution with optimal VNF placement and traffic routing. We also proposed a K-shortest path-based greedy algorithm to reduce the complexity for solving MILP. Numerical analysis and simulations are conducted to validate the proposed algorithms. Our results show multi-source multicast has a better routing selection compared to single-source multicast due to the more options of multicast sources for providing a reliable network service.

Keywords

• Network function virtualization
• reliability
• delay
• multi-source multicast
• resource optimization