Side_1_360

87

activities has to be conducted to prepare for pro-

vision of services.

These inputs are categorised as:

• Demand characteristics. The demand patterns
  for the different applications have to be speci-
  fied. Besides the demand patterns for the
  applications, characteristics of each applica-
  tion have to be more closely specified, e.g. in
  terms of required bit rates, delay requirements,
  etc. This also includes the set of traffic matri-
  ces referring to certain instances of time.
  Commonly a number of traffic matrices would
  be needed, each related to a set of applica-
  tions. A traffic matrix gives the amount of
  traffic requested from an originating location
  area towards a destination. When the traffic
  sources are moving, the spatial impacts have
  to be taken into account more closely.


• Network element characteristics. The network
  elements belong to the set of building blocks
  that the network may be composed of. These
  elements have to be described in ways rele-
  vant for traffic and resource handling. That is,
  information like capacity per unit, hierarchy
  of units, dependability and load sharing fea-
  tures, queueing management principles,
  admission control mechanisms, and so forth,
  have to be specified. When a cost model is
  used, some of the network element character-
  istics will also be included in the cost calcula-
  tions. Which elements to consider depends on
  the scope of the study. In some cases only
  routers and transmission link capacity are
  looked at, while other types (e.g. call servers,
  management systems, etc.) may also be rele-
  vant for other studies.


• Management policy. The main principles of
  handling traffic and network resources belong
  to the management policy. That is, principles
  like which routing policy to apply, which
  dependability principle to use, and so forth,
  are placed in this category. Ways of integrat-
  ing and segregating types of traffic flows may
  be candidates to be decided on.
  
  
  • External factors. Other phenomena to be taken
    into account are put into a common group.
    Examples of such factors are ways of inter-
    connecting, regulatory directives, competitors’
    actions, etc.
  
  
  • Charging and accounting policy. The charging
    principles may be identified as flat rate charg-
    ing, volume based charging, time based charg-
    ing and congestion-based charging, or a com-
    bination of these. In addition to the charging
    principles the tariffs may vary over certain
    time periods, i.e. over the day and specific
    days during a week, and so forth.
  
  
  
  

Deciding upon an efficient network design

implies that an objective function has to be spec-

ified in order to settle which designs are the bet-

ter ones. That is, the objective function would be

a measure on how good the solutions is, imply-

ing that any improvement in the design results in

a better objective value. Thus, this can be seen

as an optimisation problem using the objective

function and a set of constraints. One set of clas-

sical objective functions used contains a measure

of cost. This will reflect the capacities needed of

the different equipment types. The constraints

would then state the demands to be served as

well as other requirements, e.g. those belonging

to the management policy group of inputs.

After finishing the deployment investigations,

an appropriate way of configuring the network

resources in order to handle the traffic flows is

found. In addition to the technical solution, other

aspects might also be relevant, like some eco-

nomic measures.

2.2 Relations with Economic

Considerations

Usually, a management team considers the eco-

nomic variables when deciding upon steps for

changing a network. This means that these vari-

ables should be estimated. In order to carry out

such studies, a combination of technical and eco-

nomic considerations is needed. That is, the

technical aspects may involve demand estima-

tion, description of network building blocks,

Figure 1 Groups of input

data for service deployment

investigations

Demand

characteristics

Network element

characteristics External

factors

Management

policy

Charging and

accounting policy

• Technical configuration


• Economics


• Performance


• ....

Deployment

investigations

Rima Venturin (32) is Research
Scientist at Telenor R&D. She is
involved in projects related to
broadband network planning.
She is currently working with
techno-economics of IP-based
networks. Other activities in-
clude demands and application
modelling. She received her
MScEE degree from the Univer-
sity of Zagreb in 1995.

[email protected]

Inge Svinnset (46) is Senior
Research Scientist at Telenor
R&D. His research interests
include teletraffic, network per-
formance, Quality of Service
and dimensioning. During the
last ten years he has been
involved in several European
research projects related to
ATM network performance and
dimensioning. He is currently
working with Quality of Service
and Traffic Engineering in IP
networks.

[email protected]

Telektronikk 2/3.2001