Posted on

UBITECH co-authors a scientific paper on performance evaluation of service chains placement algorithms for Edge Computing at ITC30

A scientific paper entitled “Performance Evaluation of Service Functions Chain Placement Algorithms in Edge Cloud” has been co-authored by University of Würzburg and UBITECH, and is presented at the International Teletraffic Congress ITC 30, that is held September 4-7, 2018, at the University of Vienna, Austria, and constitutes the 30th edition of the international flagship congress in the field of networking science and practice. In this paper, Dr Anastasios Zafeiropoulos and his co-authors propose four placement algorithms that aim to efficiently place the Service Functions Chains (SFCs) in servers with regard to minimizing service response time and resource utilization. Herein, heuristic approaches are evaluated against optimal solutions for the placement problems, which are formulated by using Integer Linear Programming (ILP). They evaluate and compare these placement strategies in a simulator. The result shows that the optimized solutions produce lowest service response time and least server utilization in all types of simulated SFCs.

In particular, the co-authors use EdgeNetworkCloudSim to simulate a fixed network topology in an edge cloud. Users randomly request service chains consisting of three VNFs that can be placed on different servers. The performance of all placement algorithms is evaluated with respect to QoE in terms of service response time, and with respect to resource consumption in terms of number of utilized servers. The simulation result shows that the optimized solutions obtained by using ILP model achieve lowest service response time and least service utilization rate among the others. However, the heuristic algorithms are able to come close to the optimum by simple placing rules. Such insights may help network operators and the research community to quickly compute good SFC placements for NFV infrastructures in an edge cloud context.