23-02-2013, 02:11 PM
NETWORK TRAFFIC MANAGEMENT
NETWORK TRAFFIC.docx (Size: 893.7 KB / Downloads: 27)
INTRODUCTION
Network resources are shared as needed by a community of users. Without effective traffic controls, networks are vulnerable to possible congestion when the offered traffic exceeds the network capacity, leading to serious deterioration of network performance. This chapter gives an overview of traffic control methods and their underlying principles. The main preventive controls are admission control and access regulation (policing). They seek to avoid congestion by limiting the amount of traffic entering the network. Within the network, packets may have different requirements for reliable or timely delivery. These requirements are recognized by packet scheduling and buffer management algorithms implemented in switches and routers.
Although an analogy between computer networks and automobile highways is simplistic, the view of networks as a type of infrastructure points out the need for traffic control. Highways have a limited capacity which can be exceeded when many people want to travel at the same time. Vehicles begin to slow down and ack up in a congested area. The backup spreads if traffic approaches the congested area faster than traffic can leave. Similarly, computer networks are designed to handle a certain amount of traffic with an acceptable level of network performance. Network performance will deteriorate if the offered traffic exceeds the given network capacity.
NETWORK TRAFFIC MANAGEMENT
Network Traffic Management (NTM) is a collection of techniques that may be used by Internet service providers to attain optimum performance for diverse classes of users. These techniques include the use of performance measures to define optional service levels tailored to different user needs, and to assure quality of service. Traffic management is already common in the portion of a network where it is possible to smooth traffic flow without affecting performance.
NTM is critically important to the proper functioning of the Internet, yet, NTM can also be misused to create unfair access or use of the Internet. The paper presumes that an objective exposition of NTM’s technical issues will help policy makers, regulators, and the industry develop fair and informed regulations and policies.
The Internet, as an international interworking of independently operated, autonomous networks, has, by definition, neither a central governing body nor a policy enforcer. However, the United States has historically been a leader in Internet governance. Any action taken by the U.S. government in Internet governance has a far-reaching impact on how other governments look to the Internet. Moreover, because of this impact, the U. S. government’s positions in Internet governance, even if in an area that is wholly within the purview of the United States and not the global Internet, come under intense scrutiny. Thus, it is of utmost importance that the debate on Internet availability and accessibility is based on technological facts, capabilities, and projected growth.
FUNDAMENTALS OF NETWORK TRAFFIC MONITORING AND ANALYSIS
Traffic measurement is a well-established field of telecommunications research. Early work in this field focused on the circuit-switched telephone network. In this environment, information about the duration of a call, its origin and destination points, and its route are usually well defined, and the centrally managed switching and signaling infrastructure provide a platform for collecting this network traffic data. In contrast, the Internet is a packet-based and highly decentralized network. The design of the Internet has aimed to minimize the amount of higher layer information and connection state data that needs to be kept within the network layer. When coupled with the highly decentralized structure of the Internet, this has created major challenges for network managers of IP networks. If users experience packet delay or loss, there is no intrinsic support to identify the route those packets took. This creates a challenge for effective performance and fault management. Similarly, it can be difficult to analyze patterns of customer usage because service information is kept in application clients or servers, rather than in the network.
CLASSIFICATION OF NETWORK TRAFIC
TYPES
Today’s networks must accommodate an increasingly complex set of data traffic. Network traffic management tools enable network administrators to identify traffic types, users and applications, and facilitate the optimization of the network so all three components interact appropriately.
The first step toward network optimization is to understand where inefficiencies and problems are located, and why they are occurring.
The integration of various standards and technologies, such as XRMON and sFlow, combined with effective traffic management software tools, such as ProCurve Manager Plus, enables companies to effectively support various traffic types with optimum fault resolution, performance management, and capacity planning on all ports.
The following examples represent several types of network traffic that must be considered and supported
MODELS FOR MANAGING TRAFFIC
Given the need for network traffic management noted above, two main approaches have evolved — managing traffic in the core and managing traffic at the edge. Though techniques in these areas are not mutually exclusive, service providers’ goals in applying these techniques differ between the core and the edge.
Comprehensive end-to-end traffic management across the Internet is not currently feasible. Large consumers of service provider resources often employ service level agreements, however, to guarantee QoS within a single provider’s network, illustrated in Figure 1. Given that much of today’s Internet traffic does not stay within a single service provider’s network, as illustrated in Figure 2, techniques may need to be developed to support end-to-end traffic management.