39

1 Introduction

A basic definition of Traffic Engineering (TE) is

performance optimisation of operational net-

works, including measurement, modelling,

characterisation and control.

This basically means that running networks are

in focus; however, also longer-term planning has

to be considered for a running network to be

able to cope with the future traffic flows and

their characteristics.

Key performance objectives of Traffic Engineer-

ing (TE) can be categorised, ref. Figure 1, as

• traffic oriented; or


• resource oriented.

The first includes means undertaken to improve

the provision of services, having objectives like

reduced delay, reduced packet loss, increased

throughput. Resource oriented objectives opti-

mise the resource utilisation, resulting in less

installed network capacity. Trade-offs are com-

monly seen between these two perspectives.

The traffic flows to be served by the network

will likely have a range of requirements. Two

types of requirements are delay and loss. To a

certain extent, these may also face a trade-off

as shown in Figure 2.

Hence, for a given traffic load there may be an

option to “trade” between delay and loss by

adapting the buffer size, as long buffers give

higher delays but lower losses. By reducing the

traffic load, both loss and delay decrease. Typi-

cally, real-time traffic flows ask for fairly low

losses and low delays (and low delay variations).

Hence, shorter buffers and lower traffic loads

can be used as thresholds for such flows. On the

other hand, when more elastic traffic flows are

Traffic Engineering Principles, Activities

and Mechanisms

TERJE JENSEN

Moving beyond the single class, best-effort IP network, most operators introduce Traffic Engineering

(TE) mechanisms. These mechanisms are fairly crucial for further operation, supporting the portfolio of

services requested.

This article gives an overview of TE concepts and mechanisms by describing taxonomy and organisa-

tion of TE activities. It is mainly drawing on results presented in an Internet draft (ref. [ID_tepri]).

Figure 1 Balancing traffic and resource concerns

Terje Jensen (39) is Research
Manager at Telenor R&D,
Kjeller. He earned his PhD
degree in 1995 from the Norwe-
gian University of Science and
Technology. Activities include
performance modelling and
analysis, dimensioning and net-
work evolution studies.

[email protected]

Resources

Traffic Engineering:

• measurements


• characterisation


• modelling


• control

Traffic

High resource

utilisation

Low delay

Low loss

High throughput

Telektronikk 2/3.2001