Software-defined network (SDN) has become one of the most important architectures for the management of largescale complex networks, which may require repolicing or reconfigurations from time to time. SDN achieves easy repolicing by decoupling the control plane from data plane. Thus, the network routers/switches just simply forward packets by following the flow table rules set by the control plane. Currently, OpenFlow is the most popular SDN protocol/standard and has a set of design specifications. Although SDN/OpenFlow is a relatively new area, it has attracted much attention from both academia and industry. In this paper, we will conduct a comprehensive survey of the important topics in SDN/OpenFlow implementation, including the basic concept, applications, language abstraction, controller, virtualization, quality of service, security, and its integration with wireless and optical networks. We will compare the pros and cons of different schemes and discuss the future research trends in this exciting area. This survey can help both industry and academia R&D people to understand the latest progress of SDN/OpenFlow designs.

/pdf/a-survey-on-software-defined-network-and-openflow-from-2w15cg0qir.pdf

A Survey on Software-Defined Network and OpenFlow: From Concept to Implementation

Software-Defined Networking

Software defined networks

Open Networking Foundation(ONF)の標準化動向

A new adaptive traffic engineering method for telesurgery using ACO algorithm over Software Defined Networks

The overhead portion especially for a small packet is one of the most significant parameters that can lower the efficiency of the link utilization. Therefore, many traditional hop-by-hop header compression techniques were introduced to reduce the amount of packet header before transmission. However, those techniques still cannot overcome the time performance issue as compression mechanism requires longer processing time. In the time-sensitive applications such as VoIP, high network latency can lead to voice quality problems. To improve the time performance, instead of using hop-by-hop basis, the authors propose a new header compression mechanism which can be deployed in end-to-end nodes using the Software-Defined Network concept. This mechanism can reduce both packet size and time delay. This research will show the results of round-trip time between our end-to-end header compression mechanism, multiple hop-by-hop header compression mechanism, and normal IP routing algorithm.

End-to-End Header Compression over Software-Defined Networks: A Low Latency Network Architecture

A small link bandwidth can lead to insufficient capacity and delay problems. Header compression can reduce the amount of packet header to make link utilization more efficient, but this compression technique introduces a longer processing time and may increase end-to-end delay. In time-sensitive applications such as VoIP, this additional delay is unacceptable since high network latency can lead to many voice quality problems. Therefore, we propose a new header compression mechanism that can be deployed in end-to-end nodes using the software-defined networking concept in contrast to conventional header compression where the compression procedure works on a hop-by-hop basis. This new architecture allows both the packet size and the time delay to be appropriately reduced. Here, we compare end-to-end packet delay results for our end-to-end header compression mechanism, the multiple hop-by-hop header compression mechanism, and the normal IP forwarding method.

Better network latency with end-to-end header compression in SDN architecture

Since the TCP transmission rate is low during the startup phase, the QoS of various applications is seriously degraded. To improve the TCP transmission rate in the startup phase, many fast startup algorithms have been proposed. In this paper, we evaluate the fundamental performance of fast startup algorithms. Specifically, we focus on the initial window size and Rate-Based Pacing (RBP) start method, which is employed by many of fast startup algorithms. Through the simulation experiments, we first compare the performance of the RBP start method with that of the normal TCP. We then derive a way to set the initial window size for fast startup TCP algorithms.

Initial CWND determination method for fast startup TCP algorithms

To accommodate a wide variety of application services, the network virtualization environment (NVE) has received much attention. In an NVE, infrastructure providers (InPs) provide service providers (SPs) with virtual networks (VNs) using the InPs' substrate resources for the SPs' application services. Service providers have flexible requirements of topology, quality of service (QoS), and their dynamic changes on VNs. To meet all these requirements, traditional approaches based on performance guarantee relying only on InPs are inefficient and sometimes infeasible because they result in significant operational cost. We propose an architecture in which SPs can select appropriate resources from available resources and dynamically use the allocated resources depending on their application characteristics and user situations to meet their demands. In the proposed architecture, SPs can improve VN performance through resource control, and InPs are not required to have any performance guaranteed networks. Therefore, the proposed architecture can balance QoS and resource efficiency. We also present the basic design of the proposed architecture and show use cases and a prototype implementation using OpenFlow technologies.

Architecture for resource controllable NVE to meet service providers' dynamic QoS demands

The explosive growth of its usage means that the Internet has to become more flexible in order to handle a huge traffic volume efficiently and a variety of users requirement adaptively. OpenFlow architecture is a promising technology for flexible switching and routing to achieve this. In OpenFlow networks, OpenFlow switches can change their action on the basis of header information of an incoming packet. This allows to control traffic flows on the basis of various routing granularities such as destination IP address, TCP connection, etc. While a fine-grained control can increase routing optimality, its control overhead may decrease the communication performance. In this paper, therefore, the impact of routing granularity on communication performance is investigated through simulation evaluations. The results suggest dynamic changes in the routing granularity be considered for efficient traffic control.

Masayoshi Shimamura

Papers

End-to-End Header Compression over Software-Defined Networks: A Low Latency Network Architecture

Better network latency with end-to-end header compression in SDN architecture

Initial CWND determination method for fast startup TCP algorithms

Architecture for resource controllable NVE to meet service providers' dynamic QoS demands

Effects of routing granularity on communication performance in OpenFlow networks