Abstract
Software-defined networking (SDN) technology is expected to offer higher flexibility and programmability and enhanced transmission performance by decoupling control plane from data plane and enabling centralized network management. In SDN, switches may cache a certain number of flow forwarding rules, so that user flows can be forwarded accordingly. In this article, stressing the limited caching space of switches and the heterogeneous transmission performance of switches and links, we jointly design rule caching and flow forwarding strategy for multiple user flows in SDN. To emphasize the importance of the end-to-end delay caused by the transmission and processing of user flows in both the data plane and control plane, we formulate the joint optimization problem as an end-to-end delay minimization problem. As the original optimization problem is a non-deterministic polynomial hard (NP-hard) problem, which cannot be solved directly, we propose a heuristic algorithm which successively solves three subproblems, i.e., flow forwarding subproblem, rule caching and candidate path selection subproblem, and resource sharing subproblem. By applying the K-shortest path algorithm, a priority-based rule caching algorithm, and Lagrangian dual method, respectively, the three subproblems are solved and the joint rule caching and flow forwarding strategy is obtained. Simulation experiments are conducted to examine the effectiveness of the proposed algorithm, and the results indicate that our proposed algorithm is capable of improving system performance by about 20% compared with the previous solutions.
Highlights
T HE vertically integrated network architecture of traditional Internet and distributed packet forwarding schemes result in complicated network control andconfiguration, and highly limited packet transmission performance [1]
We investigate the problem of joint rule caching and flow forwarding for multiple user flows in software-defined networking (SDN), and formulate the joint optimization problem as an end-to-end delay minimization problem
This is mainly because the algorithm proposed in this paper jointly considers rule caching and flow forwarding issues and focuses on minimizing the total data plane and control plane delay of user flows, while the scheme proposed in [20] only investigates flow forwarding strategy and fails to consider rule caching strategy, may lead to undesired control plane latency, and the algorithm proposed in [28] aims to reduce the space occupation of flow forwarding rules, which might lead to excessively long delay
Summary
T HE vertically integrated network architecture of traditional Internet and distributed packet forwarding schemes result in complicated network control and (re)configuration, and highly limited packet transmission performance [1]. 2) While rule caching and flow forwarding schemes have been studied for SDN in previous work independently [9]–[27], it can be demonstrated that the two issues are closely related and the joint design is highly desired. To emphasize the importance of the endto-end delay caused by the transmission and processing delay of user flows in both data plane and control plane, we take into account flow forwarding requirements, rule caching and resource allocation constraints, and formulate the joint rule caching and flow forwarding problem as an end-to-end delay minimization problem.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
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
