Broadband Wireless Networking Lab

School of Electrical and Computer Engineering Georgia Institute of Technology

• Project Overview
• Personnel
• Project Description
• Publications
• Reading Material
• BWN Projects

TECS-SENNA: Traffic Engineering Control System for SDN/OpenFlow Network

Overview

Software Defined Networking (SDN) is an emerging networking paradigm that separates the network control plane from the data forwarding plane. In a conventional network, data flow is controlled by routers and switches. Thus, SDN promises to dramatically simplify the network management, reduce operating costs, and promote innovation and evolution by the following two basic concepts: The control plane – where the SDN controller resides, globally regulates the network states via network polices in either a centralized or distributed manner. The data plane – including OpenFlow switches, routers and other infrastructure elements, data packets are forwarded by following network policy rules such as forwarding flow tables that are programmed into OpenFlow switches through OpenFlow protocols. Using these key concepts, SDN provides a powerful control platform with a unified and global view of complex networks for the network management with traffic flow.

In recently research works, the main focus is on developing the SDN architecture rather than developing Traffic Engineering (TE) tools. However, TE is an important problem  to optimize the network performance, especially for a centralized controlled network such as the SDN paradigm, by dynamically analyzing, predicting, and regulating the behavior of data transmitted over that network. TE has been widely exploited in past and current data communication networks, such as ATM and IP/MPLS networks. However, their corresponding TE mechanisms are not useful for the next generation networking such as the SDN paradigm due to the following two main reasons: First, today’s Internet applications require the underlying network architecture to react in real time and to be scalable for a large amount of traffic types from different applications, and to provide a suitable and specific service for each traffic type in a very short time period. Second, facing the rapid growth in cloud computing and thus the demand for massive-scale data centers, a suitable network management tool should be able to improve resource utilization for optimum network  performance. Thus, an  intelligent and efficient TE tool is urgently needed at this stage.

Therefore, we develop a new TE management tool called TECS-SENNA, i.e., a traffic engineering control system for SDN/OpenFlow networks. There are four main components of TECS-SENNA namely OpenFlow interface, performance evaluation, network policy management, and network control update with an objective to realize an efficient and practical network management. Thus, the connection of TEs in TECS-SENNA provides a dynamically and globally optimized network resource allocation in such a way that the overall performance can be improved, including throughput, latency, stability, and load balancing, while satisfying the per-flow QoS requirements. Moreover, the network resilience is more strengthened by exploiting the global status and centralized control, and the new service deployment is more facilitated in large scale and within a short time.