S3 TRUNK ARRANGEMENTS

Radio School
S3 Trunk arrangements

Modulator

Detector

Channel decoder

Channel coder

Speech coder

Speech decoder

RCUR
Core Unit Radio Systems and Technology
1

S3

S3 Trunk Arrangements
Index
busy hour
collision
DSI
Engset distribution
erlang
Erlang B
Erlang C
idle marking
Offered traffic, o
loss system, Erlang B
multiple access
packet transmission
paging channel
Poisson distribution
Poisson model
queuing system
queuing time
radio exchange
radio trunk system
single access
trunk traffic, t
trunk utilization

S3 TRUNK ARRANGEMENTS

S3 Trunk arrangements
Contents
Page
3
7
13
15
17

1. Introduction
2. Queuing models
3. Signalling in radio switching systems
4. Summary
5. References
Appendix

18

1. Introduction
A user of a communications network generally needs service only a small
proportion of the total time. A telephone subscriber typically uses the phone for
1-3 minutes during the busy hour (the clock hour during which most traffic
occurs).
Furthermore, during a telephone call, each party will be active on average for
less than half of the time. This uneven loading due to bursty traffic results in
poor utilization of the transmission channels. The situation can be even more
extreme in respect of interactive data traffic between data terminals and central
computers (Fig. 1.1).
Level 1: Trunk system

Terminal (T) busy only

small proportion of the time

T
Terminal

Telenetwork

t1

t2

t3
t

T
Traffic (intensity): t1+t2+t3 0.01 erlang
T

Level 2: Speech interpolation. Packet.

Activity factor::
ta+tb+tc
0.4 speech

t1
0.01 interactive
data traffic

10

T is not active all the time during a call

ta
tb
tc
kb/s
t
t1

Fig. 1.1.
3

S3 TRUNK ARRANGEMENTS

Despite uneven traffic loads, good utilization of the transmission channels is

possible through the trunking of the traffic generated by a large number of
subscribers whose service requests are largely random and uncorrelated. The
greater the number of subscribers sharing a transmission resource, the greater
will be the trunking gain. An asymptotic limit is that the capacity of a network
need not be higher than the average traffic load during the busy hour. This
means that fluctuations in traffic flow over time are completely balanced out in
a perfect trunking system.
In a practical situation, however, the available peak transmission capacity must
exceed the mean traffic intensity. Besides the actual number of subscribers
sharing the trunking resource, the difference between the top capacity and the
mean traffic intensity is also determined by the level of inconvenience that
subscribers are prepared to tolerate. Because the capacity of the transmission
channel (trunk group) is lower than the number of subscribers sharing the trunk
group, there is a chance that the available capacity will be insufficient. Thus,
excess traffic must either be blocked (loss system) or placed in a queue to wait
for free capacity (delay system).
A general picture of these relationships is shown in Figs. 1.2 and 1.3. The
figures apply to a loss system based on the Erlang B formula. The traffic (intensity) is measured in erlangs. One erlang corresponds to 100% traffic load on
one line. One subscriber who during the busy hour has two calls, for instance,
each having a duration of 1.5 min, generates A = 3/60 = 0.05 erlang. If N
terminals are connected, then the total offered traffic Ao = N.A erlang. The
most important measure of the performance of a loss system is the relationship
between the offered traffic o per trunk (o = Ao/n) and the corresponding
blocking or loss probability, pb, in other words, the probability of a call being
blocked owing to there being no available trunk. Because of the blocking, a
proportion, pb, of the offered traffic will be lost, which means that the carried
traffic, t, per trunk will be:
t = (1 - pb)o
As shown in Fig. 1.2, a substantial trunking gain is obtained if a number of
subscribers share a trunk group comprising several lines, instead of a smaller
number of subscribers having access to just one common line.
Improved spectrum efficiency from trunking gain
N radio terminals use a common trunk group comprising n traffic channels.
Traffic per terminal is 3 min (0.05 erlang) during the busy hour.
Permitted blocking, pb, = 10%

pb
blocking, %
100
Loss system with
lost calls cleared

Telenetwork
Signalling
channel

n traffic channels
Radio switching
Idle traffic
channels
n=1
50

Radio channels

n=10

N terminals

n=36
Permitted blocking

n
1
10
36

Fig. 1.2

Call minutes
per hour

Number of terminals
in network

Per channel

Total

Per channel

6 (9)
45
57

6 (9)
400
2,052

3
15
19

Offered traffic/channel, o =

Trunk
utilization

10

10%

Total

3
150
684

0.5

68%

85%

If N>>1, the system can carry

60 . o . n minutes of call time per busy hour
60 . o . n
terminals
System capacity:
terminal traffic (call minutes/busy hour)

t
(1 pb)

or t n erlangs carried traffic

S3 TRUNK ARRANGEMENTS

If one radio channel is shared between a relatively large number, N, of terminals, the permitted blocking will determine the trunk utilization (t = pb). If 10%
blocking is allowed, then there will be a 90% probability that none of the other
terminals (N - 1) will be using the channel at the arbitrary moment at which the
nth terminal want to be connected. Thus, the offered traffic from (N - 1) terminals will be:

A 0 = 0 =

t
1 p b

pb
1 p b

= 0,11 erlang

(n=1)

If N >> 1, we can disregard the difference between N and (N-1). In other words,
the number of terminals that can share the channel is given by the equation:
Ao = 0.11 N.A. However, at low values of N, N must be replaced by (N - 1)
in the above equation. In this case (n = 1 in figure 1.2) 10% blocking is obtained
if (N - 1) terminals generate 6.6 min of traffic per busy hour, i.e. (N - 1) = 2,
since each terminal generates three minutes. Thus, one channel can only serve
three terminals.
For a trunk group consisting of more (n) traffic channels, the offered traffic per
channel, o, will be substantially higher at 10% blocking. For n = 10 we get
o = 0.75 (see the chart for Erlang B in Fig. 2.2 which gives t = 0.68). The
maximum number of terminals that can be connected is determined by the total
offered traffic: (0.75 x 10 x 60) = 450 call minutes per busy hour. The trunk
group can therefore serve (450/3) = 150 terminals. Similarly, for a trunk group
of 36 radio channels, 10% blocking gives o = 0.95, which means that the
number of terminals it can serve is (0.95 x 36 x 60/3) = 684.
The Erlang B model applies to a lost call cleared situation; in other words, once
a call has been blocked, the subscriber will no longer wish to establish the
desired connection. A more plausible case is that the subscriber whose call has
been blocked will try again in the hope that all trunks will no longer be busy. As
can be seen from Fig. 1.3, retrials increase the blocking probability (see also
section 2.2).

Offered traffic (o erlang) per trunk in a loss system

o
1
Lost calls cleared
Lost calls returning

0%

=2
pb
ing

ck

Blo

Lost calls cleared

Lost calls returning

=5
g pb

0.5

in
ck

Blo

Offered traffic = tsmN

ts = Average holding time
m = Average number of calls per
busy hour per subscriber
N = Number of subscribers (terminals)
(N>>1)

n = Number of trunks

Example:
Capacity with lost calls returning
How many subscribers can be accommodated
per radio channel with 20% blocking without
trunking (n = 1) and with a trunk group of four?
The total holding time per terminal during busy hour is 2 min.

Fig. 1.3.

Solution:
At n= 1 the capacity will be 0.16 erlang offered traffic;
that is, 60 x 0.16 = 9.6 10 call minutes per hour allowing
5 subscribers per channel.
At n = 4, the capacity will be 0.63 erlang per channel, that is 60 x 0.63 = 38 call minutes
per hour per radio channel.
Each channel can accommodate 19 subscribers, i.e. a trunk group of four channels can
accomodate 76 subscribers.

S3 TRUNK ARRANGEMENTS

The effective sharing of transmission channels is particularly important in radio

systems where the available traffic capacity is limited not only by cost but also
by frequency shortage. The acute frequency shortage in mobile radio has also
forced higher utilization of the channels than what is usual in the public telephone network. Subscribers often experience a relatively high level of blocking or considerable delays.
The transmission procedures described above are of the single-access type. This
means that the shared transmission channel connects two nodes (exchanges) to
which the subscribers put their service requests and feed the information to be
sent. The node therefore has full knowledge of the demands for service and
available trunks instant by instant, and can pack the calls densely. This achieves
optimum utilization of transmission capacity through trunking of the calls.
Single access is often used in radio networks, such as in multiplexed point-topoint communications and for outgoing traffic from radio base stations to
connected terminals. However, sharing of radio channels is also often based on
multiple access (MA). In MA, there is no common traffic-controlling node that
has total knowledge of the instantaneous traffic demand and thus able to provide
optimum sharing of the transmission resource between the subscribers. Consequently, multiple-access systems do not utilize the transmission channels as
well as single-access systems do.
This module deals with traffic models based on single-access. The usual arrangement consists of a loss system, which means that when all the channels are
busy, additional incoming calls are blocked (and not returned). In traffic theory,
this corresponds to the Erlang B model, which describes the relationship between blocking probability, the size of the trunk group and the total offered
traffic.
Another possibility is to have a queuing or delay system, whereby, when all the
channels are busy, additional demands for service are placed in a queue and
served when trunks are released. This corresponds to the Erlang C model. As
well as the delay probability, the Erlang C model also gives the average delay
(waiting) time as a function of the size of the trunk group and the total traffic
load.

S3 TRUNK ARRANGEMENTS

2. Queuing models
2.1. Overview
In this chapter we shall be looking at the overload characteristics of an exchange or concentrator that combines the traffic from a relatively large number
(N) of subscribers to a smaller number (n) of trunks (see Fig. 2.1). An important
difference between a loss system (Erlang B) and a delay system (Erlang C) is
that a delay system (with infinite capacity) must accomodate all offered traffic
(At = Ao), whereas a loss system rejects the excess traffic. In the latter case, we
need to distinguish between trunk utilization (trunking efficiency) t = At/n and
o = Ao/n, that is, the offered traffic load normalized to the number of trunks.

N
subscriber
lines

n trunks

Aa
A
(Trunk traffic, t )
)
n
n
The following is true during the busy hour:
Offered traffic per subscriber:
Call arrival rate
in the same unit
Average holding time, t m

a =
(Offered traffic, Ao or

Fig. 2.1

Traffic intensity per subscriber = tm erlang

Total offered traffic = Ao = N tm = tm (=N )
Normalized to number of trunks, o
Carried traffic per trunk,
t
Total carried traffic, A t = n
t
With delay system: A o = At = A
o = t =
With loss system: A t = Ao (1p b) t = o (1p b)
p b = Ao At : (probability of blocking)
Ao
Example for loss system:
= 2 calls per hour
tm = 3 min
Ao = 2 3 50 =
60
N = 50
Aa = 5 erlang
n = 8
Ppbb = 8%
a0 = 85
= 5(1 0.08) = 0.575
t
8
(from Fig 2.2)

Both the Erlang B and Erlang C models are based on the assumption that the
call arrivals are completely random, having no mutual correlation. The number
of call attempts per unit time, , is not time dependent. In the corresponding
mathematical model, the interarrival times have a negative exponential distribution, and the number of arrivals in a given period has Poisson distribution
(see the Appendix). A further assumption is that the number of terminals is
large (N ).
Another important traffic model is the Poisson model which is used in the USA
and discussed in the Appendix. The Poisson model gives somewhat higher
blocking than Erlang B, which roughly takes into account the retrials, which are
disregarded in the Erlang B model.

S3 TRUNK ARRANGEMENTS

2.2. Erlang B loss system

The Erlang B traffic model is based on the assumption that the calls blocked
during traffic overload will be cleared from the system, i.e. no calls are returned. It is also assumed that incoming calls have a totally random time distribution and that N = . This case is described in Fig. 2.2 (see also Table 1).

Loss system

Erlang B

Blocking probability

Ab
Ao

At = Ao (1-p b)

N=

n trunks

p = Ab
b A
o

1. Lost calls cleared

2. Call arrivals have Poisson
distribution
3. Number of connected
subscribers, N, =

Trunk utilization:
A A Ab Ao (1p ) = (1p )
t = t = o
b
o
b
=
n
n
n

0.5
n=1

0.3
n=2

0.2

Example:
Permitted blocking is 3%.
What traffic load can be offered to a trunk
0.1
group with
a) n=4?
b) n=16? What will be the corresponding
trunk utilization, t?
0.05

Solution:
Ao = 4 0.31 = 1.24
a) = 0.31
t = 0.31 0.97 = 0.30
b) = 0.67
Ao = 16 0.67 = 10.7
t = 0.67 0.97 = 0.65

16

00

50

n=1

n=

n=

30

n=

n=

n=

0.03
0.02
o = Ao
n

0.01
0.2

0.5

0.8

Fig. 2.2
Retrials resulting from blocking means that the traffic load consists of both new
traffic and return traffic. Mathematically this can be treated as an increase in the
offered traffic, provided that the return traffic is random with an average delay
longer than the average holding time. (If the interval between re-tries is too
short, the probability of blocking will increase as the overload situation will not
have been resolved. If several subscribers whose calls have been blocked try
again immediately, there will be an acute traffic peak, with a renewed risk of
system overload.)
Since several retries may be necessary, the following expression gives the total
number of call attempts per unit time:
= + pr pb + (pr pb ) + =
2

1 pr pb

pr is the probability of a blocked subscriber wanting to retry a call. This produces a corresponding increase in the offered traffic:
A o =

Ao
1 prpb

Return of lost calls increases the probability of blocking compared with the
basic Erlang B model.

S3 TRUNK ARRANGEMENTS

Table 1. Erlang B loss system

Attempts have Poisson distribution. Number of terminals, N, = . No retry.
Blocking probability = pb. Number of trunk lines = n.
Ao
n

Offered subscriber traffic = Ao

Normalized = o =

Carried traffic = At

Trunk utilization t =
t = (1pb)o

pb = 1%

pb = 2%

At
n

pb = 5%

pb = 10%

Ao

Ao

Ao

Ao

1
2
3
4
5

0.01
0.15
0.46
0.87
1.36

0.010
0.075
0.153
0.218
0.272

0.010
0.076
0.150
0.215
0.270

0.02
0.22
0.60
1.09
1.66

0.020
0.110
0.200
0.273
0.332

0.020
0.109
0.197
0.268
0.325

0.05
0.38
0.90
1.52
2.22

0.050
0.190
0.300
0.380
0.444

0.050
0.181
0.285
0.362
0.422

0.11
0.60
1.27
2.04
2.88

0.110
0.300
0.423
0.501
0.576

0.100
0.268
0.381
0.460
0.519

1
2
3
4
5

6
7
8
9
10

1.91
2.50
3.13
3.78
4.46

0.318
0.357
0.391
0.420
0.446

0.315
0.354
0.387
0. 416
0.442

2.28
2.94
3.63
4.34
5.08

0.380
0.420
0.454
0.482
0.508

0.372
0.411
0.444
0.473
0.498

2.96
3.74
4.54
5.37
6.22

0.493
0.534
0.568
0.597
0.622

0.469
0.507
0.539
0.567
0.591

3.76
4.67
5.60
6.55
7.51

0.627
0.667
0.700
0.728
0.751

0.564
0.600
0.630
0.655
0.676

6
7
8
9
10

12
14
16
18
20

5.88
7.35
8.88
10.44
12.03

0.490
0.525
0.555
0.580
0.602

0.485
0.520
0.549
0.574
0.596

6.62
8.20
9.83
11.49
13.18

0.552
0.586
0.614
0.638
0.659

0.540
0.574
0.682
0.626
0.646

7.95
9.73
11.54
13.39
15.25

0.663
0.695
0.721
0.744
0.763

0.629
0.660
0.685
0.706
0.724

9.47
11.47
13.50
15.55
17.62

0.789
0.819
0.844
0.864
0.881

0.711
0.737
0.759
0.777
0.793

12
14
16
18
20

22
24
26
28
30

13.65
15.30
16.96
18.64
20.34

0.620
0.638
0.652
0.666
0.678

0.614 14.89
0.631 16.63
0.646 18.38
0.659 20. 15
0.671 21.93

0.677
0.693
0.707
0.720
0.731

0.663
0.679
0.693
0.705
0.716

17.13
19.03
20.94
22.87
24.60

0.779
0.793
0.805
0.817
0.820

0. 740
0.753
0.765
0.776
0.785

19.69
21.79
23.68
25.99
28.11

0.895
0.908
0.911
0.928
0.937

0.806
0.817
0.82
0.835
0.843

22
24
26
28
30

32
34
36
38
40

22.05
23.77
25.51
27.25
29.01

0.682
0.699
0.709
0.717
0.725

0.682
0.692
0.701
0.710
0.718

0.742
0.751
0.759
0.768
0.777

0.727
0.736
0.744
0.752
0.759

26.74
28.70
30.66
32.62
34.60

0.836
0.844
0.852
0.858
0.865

0.794
0.802
0.809
0.815
0.822

30.23
32.37
34.51
36.65
38.79

0.945
0.952
0.959
0.964
0.970

0.850
0.857
0.863
0.868
0.873

32
34
36
38
40

23.73
25.53
27.34
29.17
31.09

Erlang B assumes that N = . Cases with low values of N and n can be analysed
using Engset distribution, which is described in published tables (e.g. see
reference 2). One example given in Reference 2 applies to the case with offered
traffic, Ao = 12 erlang and n = 12. For Erlang B (N = ), pb 1 % is obtained.
For N = 200, the Engset model gives blocking of approx. 0.8% and, for N = 50,
blocking of about 0.4%.
A related case is that where N and different subscribers generate different
traffic loads. This is discussed in references 4 and 5. The probability of blocking
here is lower for subscribers generating a high traffic load than for those generating little traffic.

S3 TRUNK ARRANGEMENTS

2.3. Erlang C delay system

In the Erlang B model, the distribution of holding times was not important.
Only the average holding time, tm, was used in the model. As regards queuing
systems, however, the distribution of the delays for calls held in a queue and the
average delay are influenced by the distribution of the holding times. The average delay will be shorter, for instance, if the service times are constant. In normal voice traffic, the distribution of service times, to a reasonable level of accuracy, is exponential. This corresponds to a constant probability of a given call
being terminated within a short period (the probability is not dependent on how
long the call has been in progress, and the average remaining duration is not
dependent on how long the call has been in progress).
The assumption of exponentially distributed service times greatly simplifies the
model and this case is therefore generally assumed to apply. The Erlang C
model for a delay system is based on this assumption and also that the queuing
capacity is unlimited. In addition the assumptions stated earlier (random incoming calls and N = ) are introduced. Another condition is that the queuing calls
are cleared in order. These assumptions together result in an exponential distribution of queuing delays. Thus, if the average delay is known, the distribution
of delays can easily be calculated.
The parameters of greatest interest to a delay system are the blocking probability
(delay probability, pd), for calls placed in a queue, and the average queuing delay. This can either be given in respect of all calls ( W ) or only for those delayed ( Wd). The first of these values will be pd times the second (see Figs. 2.3,
2.4 and Table 2). The table shows the average delay time, Wd , for delayed calls
for an average call duration, tm, of 100 seconds. The corresponding values of Wd
for other values of tm are obtained through proportioning. (For instance, if
tm = 200 s, the values of Wd given in the table must be doubled.)
Queuing system (Erlang C). Delay probability
Queuing system
A
Probability of delay, pd, = d
A

(p from Erlang B)
b

p
pd =

Infinite queuing capacity

1 (1pb)

Unlimited patience in waiting subscribers

0.5

Poisson distribution of incoming calls

n=1

Exponential distribution of service times

0.3

Number of subscriber lines, N,

n=2
0.2

Delayed traffic Ad

n=4
0.1

n trunks

n=8
0.05

n=16
n=30

Ad
0.03

At

Ao

0.2

A t =A o = A

0.5

0.8

A
n

Fig. 2.3
The average delay is proportional to the average service time and inversely
proportional to the number of trunks (n). If acceptable absolute delays are
determined by the overriding system requirements, better trunk utilization will
be obtained for short service times and large trunk groups. Trunk utilization
here will approach 100%.
10

S3 TRUNK ARRANGEMENTS
Delays in queuing systems
(Erlang C)
Average delay for queuing subscribers, W d =

Special case: n=1

n =1
pd =

tm

n(1 )

Average delay for all subscribers, W = p d . W d

W d=

t .
W= m
(1)
If constant t s :

tm = Average service time

= Offered traffic per trunk
pd = Probability of delay (as per Fig. 2.3)

tm
(1)
(ts service time
has exponential distribution)

t
W= s
2(1)

When calls are serviced in order, the delay distribution is exponential

P [Wd Wd*] = exp

W *

( W dd ) = exp (Wd*

n (1)
m

p
Pd

W
d
=
tm n(1 )

Example:
Mean service time tm =
a) 3 min and b) 30 s
= 0.8 n=4
What is the probability of a call being delayed?
If it is, what is the probability that the delay exceeds 30 s?

10
3

n=

Solution: a and b p d = 0,45

0.5
a. Wd = 3 = 3.75 min
P[Wd 30s] = exp
3.75
4.0.2
= exp (0.133) = 88%

4
n=

n=

16

n=
0.1
0.2

0.5

b. W d = 0.5 = 0.625 min

4.0.2

n=

0.3

0.5
(0.625
)=

= exp (0.8) = 45%

0.9

0.7

P[Wd 30s] = exp

)=

Fig. 2.4
Table 2. Erlang C delay system
Poisson distributed arrivals. Number of terminals, N, = .
Infinite queue capacity.
Exponential distribution of service times.
Probability of queuing = pd
Number of trunks = n
Traffic = A
Trunk utilization, = An
Wd: = Average call delay in seconds (for calls in queu) if average service time
is 100 s.
pd = 2%

pd = 5%

pd = 10%

pd = 20%

Wd

Wd

Wd

Wd

1
2
3
4
5

0.02
0.21
0.55
0.99
1.50

0.020
0.105
0.185
0.249
0.299

102
56
40.9
33.3
28.6

0.05
0.34
0.79
1.32
1.91

0.050
0.171
0.262
0.330
0.381

105
60
45.2
37.3
32.3

0.10
0.50
1.04
1.65
2.31

0.10
0.25
0.347
0.413
0.463

111
61
51
42.6
37.2

0.20
0.74
1.39
2.10
2.85

0.200
0.370
0.464
0.525
0.569

125
79
62
53
46.4

1
2
3
4
5

6
7
8
9
10

2.05
2.63
3.25
3.88
4.54

0.341
0.376
0.406
0.432
0.454

25.3
22.9
21.0
19.6
18.3

2.53
3.19
3.87
4.57
5.29

0.422
0.455
0.484
0.508
0.529

28.8
26.2
24.2
22.6
21.2

3.01
3.73
4.46
5.22
5.99

0.501
0.532
0.558
0.580
0.599

33.4
30.5
28.3
26.5
24.9

3.62
4.41
5.21
6.03
6.85

0.603
0.629
0.651
0.670
0.685

42.0
38.6
35.9
33.6
31.8

6
7
8
9
10

12
14
16
18
20

5.90
7.31
8.77
10.25
11.77

0.492
0.522
0.548
0.570
0.588

16.4
15.0
13.8
12.9
12.1

6.76
8.27
9.82
11.40
13.00

0.563
0.591
0.614
0.633
0.650

19.1
17.5
16.2
15.2
14.3

7.6
9.2
10.8
12.4
14.1

0.630
0.654
0.674
0.691
0.706

22.5
20.7
19.2
18.0
17.0

8.53
10.23
11.96
13.70
15.45

0.711
0.731
0.747
0.761
0.773

28.8
26.6
24.7
23.2
22.0

12
14
16
18
20

22
24
26
28
30

13.30
14.86
16.44
18.03
19.64

0.605
0.619
0.632
0.644
0.655

11.5
10.9
10.4
10.0
9.7

14.62
16.25
17.91
19.57
21.25

0.664
0.677
0.689
0.699
0.708

13.6
12.9
12.4
11.9
11.4

15.8
17.5
19.2
21.0
22.7

0.719
0.730
0.739
0.748
0.756

16.1
15.4
14.8
14.2
13.7

17.22
19.00
20.79
22.58
24.38

9.783
0.792
0.800
0.806
0.813

20.9
20.0
19.2
18.4
17.8

22
24
26
28
30

32
34
36
38
40

21.26
22.89
24.53
26.18
27.84

0.644
0.673
0.681
0.689
0.696

9.3
9.0
8.7
8.5
8.2

22.93
24.63
26.34
28.05
29.77

0.717
0.724
0.732
0.738
0.744

11.0
10.7
10.4
10.1
9.8

24.4
26.2
27.9
29.7
31.5

0.763
0.770
0.776
0.782
0.787

13.2
12.8
12.4
12.1
11.7

26.19
28.01
29.83
31.65
33.48

0.818
0.824
0.829
0.833
0.837

17.2
16.7
16.2
15.8
15.3

32
34
36
38
40

11

S3 TRUNK ARRANGEMENTS

A considerable complication for a heavily loaded queuing system is its vulnerability to overloading if the offered traffic during certain periods is substantially higher than the design value (e.g. mean traffic load during the busy hour).
When the traffic intensity reaches a value corresponding to 100% trunk utilization, the system collapses because of rapidly increasing delay times and number
of calls in the queue. When designing a system, we may therefore be forced to
introduce a safety margin in respect of the permissible trunk utilization. A lost
called clear system, such as Erlang B, is less sensitive to overloading, owing to
the greatly increased level of blocking when trunk usage approaches 100%; in
other words, an increasing proportion of the offered traffic is blocked. Even at
high values of o, t will never attain a value of 1.
On a radio trunking system for dispatch applications, call durations are
generally much shorter than on a mobile telephone network. Consequently, any
inconvenience to users will be fairly small, even if the average queuing delay is
as long or even longer than the average call duration. Delay systems are therefore often preferable to loss systems. However, the Erlang C model has limited
application, since the effective number of users is mainly determined by the
relatively small number of dispatchers and, in addition, the number of trunk
lines due to practical limitations seldom exceeds eight.
The relationships are so complex that we usually have to resort to simulation. A
good summary of this, complete with bibliography, is given in the CCIR Green
Book, Vol. VIII (Ref. 7). The report draws the conclusion that in certain conditions the maximum trunk utilization is achieved with a trunk group as small as
five. This is because the system must be able to handle a reasonable peak in
incoming traffic. In large trunking systems, trunk utilization is usually so high
that unacceptable delays will result if a moderate safety margin is added to the
design value of offered traffic load.
In packet transmission using single-access, a large number of packets share
a single wideband transmission channel. The simplest case is where all the
packets are of the same length and with random arrivals. This case has many
similarities to Erlang C with n=1, but one essential distinction is that the average delay here is halved. (As mentioned above, the average delay is lower with
constant service times than with varying times. The probability of delay, on the
other hand, is not affected.) In this case the average normalized delay is

w
=
tp 1

tp is the packet length and the channel utilization.

12

S3 TRUNK ARRANGEMENTS

3. Signalling in radio switching systems

An additional complication to a radio trunking system (radio switching system)
is the transmission requirement to enable terminals to make requests for assignment of traffic channels and to enable base stations to page the terminals. There
is no corresponding requirement when the terminals are connected by wire to
the exchange, since these lines are available all the time for signalling to and
from idle subscribers.
The easiest way to allocate traffic channels in a radio switching system is, for
the base station to transmit a radio signal with a special identifier on the idle
channel to be used for the next call from a terminal (see Fig. 3.1). This outward
channel is used for paging of terminals and the return channel for call requests.

Channel allocation through idle marking of an available channel

Base station

Terminals

S1

T1
T2

Terminal number
(acknowledgement)

M1
Kanal 1

S2

M2

S3

M3

T5

Acknowledgement

TB
B subscriber
number

Terminal
number

T3
T4

= Idle tone

Channel 1 BT

Call request from terminal user

Channel 2 BT
Terminal
number

T6

Call request from

fixed network

Channel 2 TB

All idle terminals search for

the tagged radio channel

Terminal number
(acknowledgement)

Channel 3 BT

Fig. 3.1

Risk of collision between call requests from

terminals

All idle terminals scan the group of radio channels that have been allocated to
the base station and look for the allocated paging channel. Once the paging
channel has been found, the search ceases and the terminals monitor this channel continuously. When the next call shall be set up, the call is assigned to the
paging channel, and the idle marking is cleared. Provided that all the available
radio channels are not busy, one of those still free is designated as the next
paging channel. The idle terminals leave the previous paging channel as soon as
idle marking is cleared, whereupon they start searching for the new paging
channel.
The procedure is simple, as no signalling other than the idle marking is required
to assign traffic channels. However, there are limitations. In large radio systems
with comprehensive signalling needs, it may therefore be necessary to introduce
a separate signalling and paging channel. This principle is shown in Fig. 3.2.

13

S3 TRUNK ARRANGEMENTS

Allocation of traffic channels using a separate signalling channel

Paging & traffic channel allocation B T

Terminals

Base stations

(Acknowledgement)
Terminal
number

Code

Traffic
channel

BT

Terminal
number

Code
t

TB
Request for channel allocation TB
No traffic channel free

Traffic
management
System signalling

Terminal
number
Terminal
number

All idle terminals monitor the dedicated

signalling channel. Assigned traffic channel
communicated via data message B T

TB

Code
t

Queuing status
Code (if any)
t

Allocation of traffic channel

Terminal
number

Code

Traffic
channel

BT
Risk of collision between data messages from terminals to base
(multiple access)

Fig. 3.2
Signalling consists of data messages, which include the identity of the terminal
in question. A wide variety of messages can be sent: paging, allocation of traffic
channel, acknowledgement of calls placed in queue to await traffic channel
becoming free. The different types of messages are identified by different
codes.
A complication to both signalling arrangements shown in Figs. 3.1 and 3.2 is
that when signalling takes place (request for channel assignment) from terminals to the base station, several terminals may transmit nearly simultaneously.
This can cause the signalling packets to collide or overlap in time, with the
result that the base station receiver can neither decode any of them nor identify
which terminals were attempting to make a call. This signalling from the terminals to the base really constitute a multiple-access situation (Aloha). This is
discussed in the packet-radio module. Protocols must be introduced so that, in
the event of collisions occurring, the risk will be minimized of further collisions
occurring when the packets are repeated.

14

S3 TRUNK ARRANGEMENTS

4. Summary
Fig. 4.1 contains a chart showing the probability of blocking, pb, or the probability of delay, pd, for Erlang B, Poisson and Erlang C, for two values of n. The
Poisson model clearly gives rise to a higher probability of blocking than Erlang
B at given values of n and . As mentioned earlier, the probability of blocking
in the Erlang B model is lower than in reality, because the effect of retries of
blocked calls has been disregarded. Since the Poisson model produces slightly
higher blocking values than in Erlang B, it might be a better model for a lost
call returning system. The Poisson model is generally used in the USA.
POISSON and ERLANG models. Overview
pb , pd
1

Erlang C
Poisson
Erlang B

0.5

0.2
n=1

0.1

0.05
n=16
0.03

Fig. 4.1

0.2

0.8

0.5

o = Ao
n

The principal parameters in the Erlang C queuing system, i.e. the probability of
v , and the
delay, pd, the normalized average delay for queuing subscribers, W
t
m

normalized average delay for all subscribers, tm , are shown In Fig. 4.2 for two
values of n. An important characteristic is the behaviour when = 1, whereupon
pd will be 1 and the delay time will be infinite. A queuing system can break
down due to overload.
Erlang C delay system: performance overview
10
n=1
5

n=8

3
2

Wv
tm

1
pd
W
tm

0.5
0.3
0.2

Wv
tm

0.1

pd

0.05
0.03

W
tm

0.02

=
0.01

Fig. 4.2

0.2

0.5

15

0.7

A
n

S3 TRUNK ARRANGEMENTS

The Erlang and Poisson distributions are based on certain assumptions; these
are shown in Fig. 4.3. The fundamental assumption is that arrivals are totally
random. This is reasonably valid for most of the time. However, situations can
arise where a large number of subscribers may attempt to initiate calls simultaneously, thus giving rise to traffic peaks that overload the system. Such situations may be caused by extraneous events, e.g. a large number of subscribers
experiencing mains power cuts or the like. Traffic peaks induced by offers on
TV programmes have also been known to jam the telephone network. Future
personal telephone systems based on microcells may have difficulty in coping
with concentrated peak loads, such as can occur immediately after a major conference or public event. (The major part of the coverage area of a microcell may
be a congress centre or sports arena/complex. In this case the assumption that
there is no correlation between service request are evidently nor valid).

Common assumptions
1. Random call arrivals
No correlation between service requests
Constant arrival rate (independent of time)
Exponential distribution of interarrival times
Poisson distribution of the number of service requests per unit time
2. Exponential distribution of holding (service) times
Probability of a call being terminated during the period, t to t + t,
is independent of t.
3. Infinite number of subscribers (N )
Offered traffic not dependent on the number of busy subscribers
(N )
Fig. 4.3

16

S3 TRUNK ARRANGEMENTS

5. References
1. Kleinrock: Queuing systems
Volume 1: Theory, Wiley 1975
Volume 2: Computer Applications, Wiley 1976
2. Bear: Principles of Telecommunication - Traffic Engineering
P. Pergrinus Ltd., 1976
3. J. Bellamy: Digital Telephone
Wiley 1982
4. Davis, Mitchell: Studies of Small Trunking Systems for Mobile Radio
Communications 78, Birmingham
5. Davis, Mitchell: Traffic Handling Capacity of Trunked Land Mobile Radio
Systems
IEEE/ICC 79
6. Dartois: Lost Call Cleared Systems with Unbalanced Traffic Sources
6th Int. Teletraffic Congress Mnchen
7. CCIR Green Book 1986 Vol VIII-1, sid 110 - 124
8. Descloux: Delay Tables for Finite and Infinite Source Systems
McGraw Hill, 1962

17

S3 TRUNK ARRANGEMENTS

Appendix
The Poisson model. TASI/DSI
Mathematical analysis of Erlang B and C is complicated as the traffic flow is
affected by the overload characteristics of the system. In the loss system, lost
calls are cleared; in the delay system, calls are placed in a queue. This is a
significant complication and it would take too long to analyse these cases here.
However, it is relatively easy to analyse the case whereby the traffic to be
handled is not influenced by calls being definitively or temporarily lost when all
the trunks are busy. The influence of overloading is disregarded in the Poisson
model (blocked calls are assumed to wait during the originally planned service
time).
Another important case in which the same applies is that in TASI/DSI, dealt
with in Module S2. In this case, the first part of the speech segment is cut off if
a trunk cannot be allocated immediately. The probability of blocking is obtained
by calculating the percentage time (the probability) that the number of incoming
active lines is greater than the number of outgoing trunks.
Poisson distribution
The basic assumption in the analysis below is that arrivals are totally random
and that the traffic intensity is not dependent on time. This means that the
number of arrivals during a given period of time will be a stochastic variable
determined by the Poisson distribution (see Fig. A-1).

Poisson and exponential distributions

Probability of event occurring during the period, t t + t = c . t

P [t t*] = X (t*) = Probability that the interval between two events is t*.
[l X (t*)] = Probability of no event occurring during the period t*.
Probability that the time between two events is within the interval of t* t* + t*:

X (0) = 0

.
X (t*) = l e c t*
P (t t*) = l X (t*) = e
dX
= c . e ct*
dt*

dX(t*) . t*
[lX (t*)] . c . t* =
dt*
X() = 1

c . t*

t* = l
c

The probability of k events occurring during

period T is given by Poisson distribution:
T
(T/ t*)k . e t*
PK (T) =
k!

( )

t*
P(tt*) = exp
t*

Example: T = t*

0.5
1 = 37%
e

Probability of the time between

two events exceeding t*

Pk (t*)

Events

0.4

t*
t*

t*

0.2
0

0 1 2 3 4 5 6

tl t2

tn
T

tn + 1

Fig. A-1
The Poisson distribution may also be regarded as an extreme case of the binomial distribution with N .
Poisson model
If, to start with, we make the simplifying assumption that all calls have the same
duration (tm), the probability of overload (blocking) can easily be calculated. At
18

S3 TRUNK ARRANGEMENTS

any given moment blocking will occur if more than n call requests have been
received during the previous time interval of length tm.
If the average arrival rate is , the probability of there being k incoming calls
during the interval, tm will be:
Pk =

(t m )k e t m
k!

Since the total traffic intensity is A = tm, the probability of blocking, Pb, may
be written as:
Pb = e A

Ak

k=n+1 k!

This traffic model is known as Poisson distribution. The relationship between

the blocking probability and the incoming traffic normalized to the number of
trunk lines is shown in Fig. A-2.
Poisson
1. Poisson distribution of arrivals
2. Exponential distribution of holding times
3. Number of connected subscriber lines, N, =

pPbb
0.6
n=1

0.3

n=2

0.2

n=4
n=8

0.1

n=16
n=30

0.05

50
100

0.03
0.02

o =

Ao
n

0.01
0.2

0.5

0.8

Example for a TASI/DSI arrangement

Traffic intensity on incoming lines = 0.25 erlang/line

Fig. A-2.

a) N = 8
n=4

A = 2 erlang
= 0.5

p b = 15%

b) N = 20
n = 10

A = 5 erlang
= 0.5

p b = 4%

A more complicated calculation for an arbitrary statistical distribution of holding times yields the same relationship if tm represents the average holding time.
19

S3 TRUNK ARRANGEMENTS

Analysis of overload characteristics in TASI/DSI

Let us assume that a concentrator has N incoming lines and n outgoing lines.
The probability that an incoming line will be active at a given instant is designated p, the value of which is determined by the incoming traffic on the line. If,
for example, this is 0.25 erlang, then p = 25%. (In this case the traffic from a
terminal is determined by the active speech intervals - level 2 in fig. 1.1).
Blocking will occur if at least n of the other (N - 1) lines are active whilst the
studied incoming line is active. This is the same mathematical problem that we
analysed in conjunction with channel coding. There, the problem was to find the
probability that a code word comprising (N - 1) bits would contain at least n
incorrect bits if bit errors occur randomly and have a probability of p. Here, the
probability that | incoming lines will be active is given by:
N1
P(l) = pl (1 p)(Nl)1
l

The probability of blocking, pb, is obtained by adding the probabilities of | = n

up to | = (N - 1) incoming lines being active:

ps =

N1

l=n

N1 pl (1 p)(Nl)1
l

Another interesting parameter in a TASI/DSI system is the average length of the

cut-off parts of the speech segment in question. If only a small part of the segments is cut off, the resulting inconvenience will be relatively insignificant. We
can estimate this if we first calculate the probability of blocking, pb, that is the
instantaneous probability that a given incoming call will be subjected to blocking. If k incoming lines are active at a given instant (k > n), then (k - n) of these
will be blocked. Thus, the number of blocked lines will be:

Px =

(k n)Pk

k =n+1

N
where
dr Pk = pk (1 p)Nk
k

is the probability that k incoming lines will be active at a given instant.

The probability that a given line will be blocked is obtained by dividing Px by
the average number of active lines, N p. From this it follows that the probability
of blocking will be:
N
1
N
p b = ( k n ) p k (1 p)N k
k
k=n+1
Np

If the average duration of an active speech segment is T, the average time cut
off will be pbT. Note that this average is valid for all speech segments, even
those that have not been affected by overloading. However, our interest is confined to those segments affected by overloading. For these, the average cut-off
time, Tbf, is given by the following equation:
Tbf =

pb T
ps

The above function becomes relatively complex at high values of N and n.

However, it is generally an acceptable approximation to use Poisson distribution
instead of binomial distribution.

20

S3 TRUNK ARRANGEMENTS

Author professor Sven-Olof hrvik

in cooperation with Ericsson Radio Systems AB
unit ERA/T, Core Unit Radio System and Technology
Publisher Ericsson Radio Systems AB
T/Z Ragnar Lodn
Ericsson Radio Systems AB
Torshamnsgatan 23, Kista
S-164 80 Stockholm, Sweden
Telephone: +46 8 757 00 00
Fax +46 8 757 36 00

EN/LZT 123 1245/3 R3

Ericsson Radio Systems AB,1997
21