Abstract
IP networks only offer best-effort service to all kinds of traffic. This means that IP tries to deliver each packet as quickly as possible, but makes no service guarantees. However, as the diversity of applications increase, this simple model with no service guarantees cannot satisfy all of them. For example, novel interactive applications such as Internet telephony, video conferencing, or networked games expect some performance guarantees to operate right. The growing importance of these recent applications with stringent constraints behooves network service providers to differentiate among various types of traffic and provide a new range of service models able to accommodate heterogeneous application requirements and user expectations.
Key Terms in this Chapter
Expedited Forwarding (EF): A PHB intended to provide low loss, low delay, and low jitter services.
Integrated Services (IntServ): An IETF proposal to provide a framework that enables per-flow QoS guarantees to individual application sessions in IP networks.
Internet Engineering Task Force (IETF): A large open international community of network designers, operators, vendors, and researchers concerned with the development and promotion of Internet standards. It is organized into a large number of working groups, each dealing with a specific topic (e.g., routing, transport, security, etc.).
Per Hop Behavior (PHB): The forwarding treatment applied to a particular collection of packets with the same codepoint at a DiffServ node.
Traffic Conditioning: Control functions typically performed on boundary nodes to enforce rules specified in a SLA. These functions may include measuring the temporal properties of a traffic stream, marking packets with appropriate codepoints, shaping traffic streams, and discarding packets.
Differentiated Services (DiffServ): An IETF proposal to provide a framework that enables deployment of scalable service distinction in IP networks.
Service Level Agreement (SLA): A service contract between a customer and a service provider that specifies the forwarding service a customer should receive.
Active Queue Management: Queue management algorithms that control the length of packet queues by dropping packets when necessary or appropriate, usually before the queue becomes full so that end nodes can respond to congestion before buffers overflow.
Assured Forwarding (AF): A PHB group intended to offer different levels of forwarding assurances for IP packets.
Scheduling Discipline: An algorithm intended to distribute network resources among parties which simultaneously and asynchronously request them.