CHAPTER ONE

INTRODUCTION

1.1 Background of study

Quality of Service (QoS) is the ability of a network element to have some level of assurance so that its traffic and service requirements can be met. QoS reflects the performance that an application may require and its experience in a network. It can be considered subjective, as users may have different views on quality. QoS is especially important for applications with stringent requirements such as telephony and media streaming. Nevertheless, QoS control has never been an easy task. End-to-end QoS such as Integrated Service (IntServ) is considered too complex and not scalable (Kuipers, 2004). Unfortunately, since most of the time the network does not operate at full capacity, this results in low network utilization. This is all the more evident in recent years as the emergence of high-speed applications forces service providers to move to Gigabit networks.

Without user application adaptation and advanced traffic control, insufficient network resources could lead to network congestion. (Bianzino, et al., 2012). This can cause significant traffic degradation that manifests as frame drops or delays. This is particularly important in unmanaged networks where there are more and more applications communicating from a variety of different users.

Moreover, with technologies like virtual reality (VR) and robotics, we can expect many new applications and services. In the not-too-distant future, an employee might accomplish their task using a live video feed from a set of camera eyes of a robot moving towards a pair of VR glasses (Bianzino, et al., 2012). The interactivity of the remote control and operating applications makes it very sensitive to frame delay and jitter. They demand high standards to the point where they impose real-time constraints on the network.

For these new applications to meet their demands, next-generation networks should become application aware and allow applications that are highly sensitive and have the highest requirements to negotiate with a network controller to allocate resources and receive guarantee on the network resources received, for example, a guarantee on the maximum bandwidth used.

When an application has negotiated with the network controller, network resources are allocated to a dedicated channel for that specific application. In practice, this is done by automatically configuring flows and priorities to meet the promised network guarantee.

The Software Defined Network (SDN), as a new networking paradigm, offers the possibility of reintroducing QoS control on networks. The centralized nature of SDN greatly reduces the complexity typically associated with QoS guarantees. With its adoption and the support of leading companies in the technology industry, SDN is well on its way to being adopted as a baseline standard. By having strong QoS control included in SDN, networks of the future could have native QoS support.

1.2 Problem Statement

Networks today are experiencing much larger consumption of data via an array of devices. With a continuous influx of data, and an increased demand for more resources, bandwidth, and accounts, an information overload is always a big concern in the networking industry. The ability to reduce complexity by automation is required to cope with real time changes at the application and user level, which the existing IP protocol network infrastructure is unable to provide. The current model is also not the most efficient use of resources, time, or capital to address tailored a specific business’ needs. Rather IP Networks provide a general pool of resources setup with a predefined structure and limited configurations that confine users, which also leads to poor network efficiency. In order to align to each business and optimize business process, redundant network resources need to be removed from the business process. A software-defined network can provide singular control over an overwhelming data load and multifaceted network infrastructure. Software-defined networks are a strategic user-friendly approach to resolving this problem and give customers more opportunity to craft their data streams today. There is significant growth opportunity and value in the software infrastructure model; however, the networking community has not picked up the SDN solution as quickly as it should, considering the value this model could bring to both its providers and its users. The new software infrastructure model is reported as being more efficient while less likely to experience technical difficulties. Awareness of software-defined networking and its growth potential in the networking community is vital to pushing the optimized process to market. Along with defining the elements, mechanics, deployment models, protocols, and advantages of software-defined networks in comparison to IP network infrastructure, we seek to define the market opportunities for implementation within a model large enterprise as well as a small business environment in today’s operating environment. We are aware that software-defined networks are becoming available in the marketplace however our research project will enable us to identify cost and productivity improvements within varying size businesses which would make this project and the associated implementation valuable to the business.

1.3 Objectives

The main objective of our work is to find a way that can help companies to have a good quality of service by using SDN paradigms.

To achieve this, the following characteristics must be applied:

• Provide bandwidth guarantees per stream;
• Correctly prioritize traffic according to the guarantee, thus respecting the limits fixed;
• Allow excessive traffic when resources are available.