3Com 4210G Series, Configuration Manual, Page: 432 / 1133

1-1
1
QoS Overview
This chapter covers the following topics:
z
Introduction to QoS
z
Introduction to QoS Service Models
z
QoS Techniques Overview
Introduction to QoS
For network traffic, the Quality of Service (QoS) involves bandwidth, delay, and packet loss rate during
traffic forwarding process. In a network, you can improve the QoS by guaranteeing the bandwidth, and
reducing the delay, jitter, and packet loss rate.
The network resources are always scarce. QoS requirements exist on any occasion where traffic flows
contend for network resources. QoS is a relative concept for traffic flows, that is, guaranteeing QoS for
a certain traffic flow may damage QoS of other traffic flows. For example, in the case of fixed
bandwidth, if a traffic flow gets more bandwidth, the other traffic flows will get less bandwidth and may
be affected. Therefore, the network administrator should reasonably plan and allocate network
resources based on the characteristics of various traffic flows, thus utilizing the network resources
effectively.
The following part introduces the QoS service models, and some mature QoS techniques used most
widely. Using these techniques reasonably in the specific environments, you can improve the QoS
effectively.
Introduction to QoS Service Models
This section covers three typical QoS service models:
z
Best-effort service
z
Integrated service (IntServ)
z
Differentiated service (DiffServ)
Best-Effort Service Model
Best effort is a single service model and also the simplest service model. In the best effort service
model, the network delivers the packets at its best effort but does not guarantee delay or reliability.
The best-effort service model is the default model in the Internet and is applicable to most network
applications. It is implemented through FIFO queuing.
IntServ Service Model
IntServ is a multiple services model that can accommodate multiple QoS requirements. In this model,
an application must request a specific kind of service from the network before it can send data. The
request is made by RSVP signaling. RSVP runs on each device from the source end to the destination
end, and monitors each data flow to prevent each data flow from consuming more resources than the