Side_1_360

Figure 6 Feedbacks on two
levels – technical and
economic

Figure 7 Multiplying and adding contributions for estimating the overall demands
when conducting techno-economic studies

Figure 8 Referring traffic load to network level

How to efficiently capture these effects in a

techno-economic model is a non-trivial chal-

lenge. Naturally, the feedback could be included

by carrying out iterative algorithms; however, a

main question is still how the relations from the

factors on the demands really are. For example,

how does an increase in the tariff influence the

demand for a certain service.

2.3 Estimating Demands – Time

Scales and Network Levels

As described above, estimating the demands is

conditional for carrying out the network deploy-

ment studies. Arriving at a tractable model advo-

cates that several approximations are introduced,

leaving the finer details of traffic flow character-

istics aside.

For a techno-economic study, average values

are likely to suffice. Then a multiplication and

addition as illustrated in Figure 7 may apply. It

should be noted that this may be carried out in

several ways. Moreover, a number of traffic

classes may be supported, such that the opera-

tions should be done for each traffic class.

The parameters considered are:

• Arrival intensity: giving the number of started
  sessions per time unit;


• Holding time: giving the duration of the ses-
  sion;


• Effective rate: stating the bit rate for the ses-
  sion (note that a session may contain a number
  of flows, each with individual holding time
  and bitrate);


• Reference period factor: reflecting the spread-
  ing of the session during the day (see illustra-
  tion in lower left corner of Figure 7);

Interest

Tariffing

policy

Demand

other factors (e.g. profit

level, revenue sharing,

competition)

network cost

performance

Network

design

other factors traffic model

At* =

c a

λc,a

b

ha,b,tBa,b,t Rc,a ρc,a Nc

arrival

intensity

holding

time

effective

bitrate

reference

period factor

penetration

number of

sources

a - number of applications

b - number of flows

c - number of classes

time and space

variations traffic

time

Session type 1

Session type 2

traffic

time

Centre

Edge

Access

traffic matrices

traffic flow parameters

reference traffic load

traffic volume

“smoothing”

multiple flows

“peak performance”

single user/ flows

peak

load levels