Traffic engineering in software-defined networking: Measurement and management
Zhaogang Shu,Jiafu Wan,Jiaxiang Lin,Shiyong Wang,Di Li,Seungmin Rho,Changcai Yang
DOI: https://doi.org/10.1109/access.2016.2582748
IF: 3.9
2016-01-01
IEEE Access
Abstract:As the next generation network architecture, software-defined networking (SDN) has exciting application prospects. Its core idea is to separate the forwarding layer and control layer of network system, where network operators can program packet forwarding behavior to significantly improve the innovation capability of network applications. Traffic engineering (TE) is an important network application, which studies measurement and management of network traffic, and designs reasonable routing mechanisms to guide network traffic to improve utilization of network resources, and better meet requirements of the network quality of service (QoS). Compared with the traditional networks, the SDN has many advantages to support TE due to its distinguish characteristics, such as isolation of control and forwarding, global centralized control, and programmability of network behavior. This paper focuses on the traffic engineering technology based on the SDN. First, we propose a reference framework for TE in the SDN, which consists of two parts, traffic measurement and traffic management. Traffic measurement is responsible for monitoring and analyzing real-time network traffic, as a prerequisite for traffic management. In the proposed framework, technologies related to traffic measurement include network parameters measurement, a general measurement framework, and traffic analysis and prediction; technologies related to traffic management include traffic load balancing, QoS-guarantee scheduling, energy-saving scheduling, and traffic management for the hybrid IP/SDN. Current existing technologies are discussed in detail, and our insights into future development of TE in the SDN are offered.
computer science, information systems,telecommunications,engineering, electrical & electronic