TELECOMMUNICATION

S ENGINEERING

Traffic Theory
&
Traffic Analysis
 Teletraffic Theory

 Traffic engineering provides the basis for the analysis and design of
telecommunication networks.

 It is not only the switching elements but also many other common shared sub­
systems in a telecommunication network that contribute to the blocking of a
subscriber call.

 In a telephone network, these include digit receivers, inter-stage switching links, call
processors and trunks between exchanges.

 The load or the traffic pattern on the network varies during the day with heavy traffic
at certain times and low traffic at other times.

 The task of designing cost effective networks that provide the required quality of
service under varied traffic conditions demands a formal scientific basis.

2
 Teletraffic Theory

 Such a basis is provided by traffic engineering or teletraffic theory.

 Traffic engineering analysis enables one to determine the ability of a

telecommunication network to carry a given traffic at a particular loss probability.

 It provides a means to determine the quantum of common equipments required to

provide a particular level of service for a given traffic pattern and volume.

3
 Network Traffic Load
and Parameters

 The traffic load on a typical working day during the 24 hours is shown in Fig.1.

Fig. 1: Typical telephone traffic pattern on a working day

4
 Network Traffic Load
and Parameters

 The originating call frequencies (no. of calls / second) are relative and the actual
values depend on the area where the statistics is collected.

 There is a little use of the network during 00:00 and 06:00 hours when most of the
population is asleep.

 There is a large peak around mid-noon and mid-afternoon signifying busy office
activities.

 The afternoon peak is, however, slightly smaller.

 The load is low during the lunch-hour period, i.e. 12.00-14.00 hours.
 The period 17.00-18.00 hours is characterized by low traffic signifying that the
people are on the move from offices to their residences.

 The peak of domestic calls occurs after 18.00 hours when persons reach home
and reduced tariff applies.
5
 Busy Hour

 The period during which the reduced tariff applies has been changed to begin later than
18.00 hours and one may expect the domestic call patterns also to change accordingly.
 During holidays and festival days the traffic pattern is different from that shown in the
Figure 1.
 Generally, there is a peak of calls around 10:00 hours just before people leave their
homes on outings and another peak occurs again in the evening.
 In a day, the 60-minute interval in which the traffic is the highest is called the Busy Hour
(BH).

 The busy hour may vary from exchange to exchange depending on the location and the
community interest of the subscribers.
 The busy hour may also show seasonal, weekly and in some places even daily variations.
 There are also unpredictable peaks caused by some “social/natural events” such as:
 stock market or money market activity,
 weather,
 natural disaster,
 international events, sporting events
 Eid and other festivals etc.
6
 Busy Hour

 To take into account such fluctuations while designing switching networks, three
types of busy hours are defined by ITU in its recommendations E.600:

 Busy Hour: Continuous 1-hour period lying wholly in the time interval concerned,
for which the traffic volume or the number of call attempts/connected calls is
greatest.

 Peak Busy Hour: The busy hour each day; it usually varies from day to day, or
over a number of days.

 Time Consistent Busy Hour: The 1-hour period starting at the same time each day
for which the average traffic volume or the number of all attempts is greatest over
the days under consideration.

7
 Example

 Example 1: An exchange serves 2000 subscribers. If the average BHCA (Busy

Hour Call Attempt) is 10,000 and the CCR is (Call Completion Rate) 60%, calculate
the busy hour calling rate.

 Solution:

Average busy hour calls = BHCA X CCR

= 10,000 X 0.6 = 6000 calls

Busy hour calling rate = Average busy hour calls / Total number of subscribers
=6000 / 2000 =3.

8
 Busy Hour

 The busy hour calling rate is useful in sizing the exchange to handle the peak
traffic.

 In a rural exchange, the busy hour calling rate may be as low as 0.2, whereas in a
business city it may be as high as three or more.

 Another useful information is to know how much of the day's total traffic is carried
during the busy hour.

 This is measured in terms of day-to-busy hour traffic ratio which is the ratio of busy
hour calling rate to the average calling rate for the day.

 Typically, this ratio may be over 20 for a city business area and around 6 or 7 for a
rural area.

9
 Traffic Intensity

 The traffic load on a given network may be on the local switching unit, interoffice
trunk lines or other common subsystems.

 The traffic on the network may be measured in terms of the occupancy of the
servers in the network.

 Such a measure is called the traffic intensity which is defined as:

10
 Erlang

 Generally, the period of observation is taken as one hour.

 A0 is dimensionless. It is called Erlang (E) to honor the Danish telephone engineer A.K.
Erlang, who did pioneering work in traffic engineering. His paper on traffic theory
published in 1909 is now regarded as a classic.

 A server is said to have 1 erlang of traffic if it is occupied for the entire period of
observation.

 Traffic intensity may also be specified over a number of servers. 11

 Example

 Example 2: In a group of 10 servers, each is occupied for 30 minutes in an observation

interval of two hours. Calculate the traffic carried by the group.
 Solution:

Traffic carried per server = Occupied duration / Total duration

= 30/120 = 0.25 E
Total traffic carried by the group = 10 x 0.25 E = 2.5 E.

 Example 3: A group of 20 servers carry a traffic of 10 erlangs. If the average duration of a

call is three minutes, calculate the number of calls put through by a single server and the
group as a whole in a one-hour period.
 Solution:

Traffic per server = 10/20 = 0.5 E

i.e. a server is busy for 30 minutes in one hour.
Number of calls put through by one server = 30/3 = 10 calls
Total number of calls put through by the group = 10 X 20 = 200 calls

12
 CCS, CS and CM

 Traffic intensity is also measured in another way. This measure is known as

Centum Call Second (CCS) which represents a call-time product.

 One CCS may mean one call for 100 seconds duration or 100 calls for one second
duration each or any other combination.

 CCS as a measure of traffic intensity is valid only in telephone circuits.

 Sometimes, Call Seconds (CS) and Call Minutes (CM) are also used as a measure of
traffic intensity. We may note that
1 E = 36 CCS = 3600 CS = 60 CM

 For the present day networks which support voice, data and many other services,
erlang is a better measure to use for representing the traffic intensity.

13
 Example

 Example 4: A subscriber makes three phone calls of three minutes, four minutes
and two minutes duration in a one-hour period. Calculate the subscriber traffic in
erlangs, CCS and CM.

 Solution:

Subscriber traffic in erlangs = Busy period / Total period = (3+4+2) / 60 = 0.15 E.

Traffic in CCS = [ (3+4+2) x 60 ] / 100 = 540/100 = 5.4 CCS.

Traffic in CM = 3+4+2 = 9 CM.

14
 Call Arrival Rate and
Holding Time

 Whenever we use the terminology "subscriber traffic" or "trunk traffic", we mean the
traffic intensity contributed by a subscriber or the traffic intensity on a trunk.

 Traffic intensity is a call-time product.

 Hence two important parameters that are required to estimate the traffic intensity or
the network load are:
 Average call arrival rate, C or λ
 Average holding time per call, th or tm.

 We can then express the load offered to the network in terms of these parameters
as:
A = Cth = λtm.
 C and th must be expressed in like time units. For example, if C(λ) is in number of
calls per minute, th (tm ) must be in minutes per call.
15
 Example

 Example 5: Over a 20-minute observation interval, 40 subscribers initiate calls.

Total duration of the calls is 4800 seconds. Calculate the load offered to the
network by the subscribers and the average subscriber traffic.

 Solution:

16
 Loss System and Delay
system

 It is possible that the load generated by the subscribers sometimes exceeds the
network capacity.

 There are two ways in which this overload traffic may be handled:

1) The overload traffic may be rejected without being serviced. …….OR

2) The exceeding Calls are held in a queue until the network facilities become
available.
 In the first case, the calls are lost …………….and
 In the second case the calls are delayed.

 Correspondingly, two types of systems are thus designed:

 loss systems and
 delay systems
17
 Loss System and Delay
system

 Conventional automatic telephone exchanges behave like loss systems.

 Under overload traffic conditions a user call is blocked and is not serviced unless the user makes
a retry.

 On the other hand, operator-oriented manual exchanges can be considered as delay systems.

 A good operator registers the user request and establishes connection as soon as network
facilities become available without the user having to make another request.

 In modern Telecommunication networks,

 circuit-switched networks behave as loss systems whereas
 store-and-forward (S&F) message or packet networks behave as delay systems.

 In the limit, delay systems behave as loss systems.

 For example, in a call centre network if the queue buffers become full, then further requests have
to be rejected.
18
 Grade of Service and Blocking
Probability

 The basic performance parameters are:

 for a loss system are  the grade of service and the blocking probability
 for a delay system  the service delays.

 Average delays, or probability of delay exceeding a certain limit, or variance of delays may be
important under different circumstances.

 Traffic Model:

 The traffic models used for studying loss systems are known as blocking or congestion models and
 the ones used for studying delay systems are called queuing models.

 In loss systems, the traffic carried by the network is generally lower than the actual traffic offered to
the network by the subscribers.

 The overload traffic is rejected and hence is not carried by the network.

 The amount of traffic rejected by the network is an index of the quality of the service offered by the network.

 This is termed as grade of service (GOS) and is defined as the ratio of lost traffic to offered traffic.
19
 Grade of Service

 GOS is given by:

 Offered traffic is the product of the average number of calls generated by the users
and the average holding time per call.

 On the other hand, the actual traffic carried by the network is called the carried
traffic and is the average occupancy of the servers in the network.
 The smaller the value of grade of service, the better is the service.

 If the recommended value for GOS is 0.002 it means that two calls in every 1000
calls or one call in every 500 calls may be lost.
 Usually, every common subsystem in a network has an associated GOS value.
20
 Grade of Service

 The GOS of the full network is determined by the highest GOS value of the
subsystems in a simplistic sense.

 A better estimate takes into account the connectivity of the subsystems, such as
parallel units.

 Since the volume of traffic grows as the time passes by, the GOS value of a
network deteriorates* with time.

 In order to maintain the value within reasonable limits, initially the network is sized
to have a much smaller GOS value than the recommended one so that the GOS
value continues to be within limits as the network traffic grows.

* Become progressively worse

21
 Blocking Probability

 The blocking probability Pb is defined as the probability that all the servers in a system are
busy .

 When all the servers are busy, no further traffic can be carried by the system and the
arriving subscriber traffic is blocked.

 At the first instance, it may appear that the blocking probability is the same measure as the
GOS.

 The probability that all the servers are busy may well represent the fraction of the calls lost,
which is what the GOS is all about.

 However, this is generally not true.  For example, in a system with equal number of
servers and subscribers, the GOS is zero as there is always a server available to a
subscriber.

 On the other hand, there is a definite probability that all the servers are busy at a given
instant and hence the blocking probability is nonzero.
22
 Blocking Probability

 The fundamental difference is that the GOS is a measure from the subscriber point
of view, whereas the blocking probability is a measure from the network or
switching system point of view .

 GOS is arrived at by observing the number of rejected subscriber calls, whereas

the blocking probability is arrived at by observing the busy servers in the switching
system.

 GOS and Pb may have different values depending upon the traffic characterization
model used.

 In order to distinguish between these two terms clearly,

 GOS is called  call congestion or loss probability and
 the blocking probability is called  time congestion.

23
 Delay Probability

 In the case of delay systems, the traffic carried by the network is the same as the
load offered to the network by the subscribers.

 Since the overload traffic is queued, all calls are put through the network as and
when the network facilities become available.

 GOS is not meaningful in the case of delay systems since it has a value of zero
always.

 The probability that a call experiences a delay, termed “delay probability”, is a

useful measure, as far as the delay systems are concerned.

24
 Delay Probability

 If the offered load or the input rate of traffic far exceeds the network capacity, then
the queue lengths become very large and the calls experience undesirably long
delays.

 Under such circumstances, the delay systems are said to be unstable as they
would never be able to clear the offered load.

 In practice, it is possible that there are some spurts of traffic that tend to take the
delay systems to the unstable region of operation. An easy way of bringing the
system back to stable region of operation is to make it behave like a loss system
until the queued up traffic is cleared to an acceptable limit.

 This technique of maintaining the stable operation is called  “flow control.

25
 Quality of Service (QoS)

 In more recent times, a new term called quality of service (QoS) is being used and
it is considered to be more general than the GOS and includes other factors like
quality of speech, error-free transmission capability etc.

 However, it is essential to recognize that there are two different important

performance measures, one obtained by observing the subscriber traffic and the
other by observing the network behavior:

 Subscriber viewpoint:
GOS = call congestion = loss probability

 Network viewpoint:
Blocking probability = time congestion

26
Traffic Analysis

27
 Call Arrival Distributions

 Most fundamental assumption of classical traffic analysis is that  call arrivals are
independent

 It means that  an arrival from one source is unrelated to an arrival from any other
source

 This assumptions may be invalid in some cases..............but......it has general

usefulness for most applications

 This assumptions lead to some mathematical formulations ............which are very

useful to find approximate solutions to some real problems which are otherwise
mathematically intractable

28
 Negative Exponential Inter-arrival
Times

Negative Exponential Inter-arrival Times:

 To designate the average call arrival rate, λ, from a large group of independent sources
(subscriber lines), following assumptions are considered:

1) Only one arrival can occur in any sufficiently small interval.

2) The probability of an arrival in any sufficiently small interval is directly proportional to the length of
the interval.
 The probability of an arrival is  λ.∆t, where ∆t is the interval length.)

3) The probability of an arrival in any particular interval is independent of what has occurred in other
intervals.

 The first two assumptions made in deriving the negative exponential arrival distribution, can
be intuitively justified for most applications.

 The third assumption implies certain aspects of the sources that cannot always be supported.
29
 Negative Exponential Inter-
arrival Times

 Based on the assumptions, The probability distribution of inter-arrival times is:

 Where:
 “λ” is the Arrival rate, which is expressed as the average number of arrivals
during a unit of time.

 The equation defines the probability that no arrivals occur in a randomly selected
interval t.

 This is identical to the probability that t sec elapse from one arrival to the next.

30
 Example

 Example 6: Assuming each of 10,000 subscriber lines originate one call per hour,
how often do two calls arrive with less than 0.01 sec between them?

 Solution:
The average arrival rate λ =(10,000/3600) = 2.78 arrivals per second

The probability of no arrival in a 0.01-sec interval can be found by:

P0 (λ.t = 2.78x0.01 = 0.0278)

 So, the probability of arrivals within 0.01 sec is: 1-0.973 = 0.027 = 2.7%

 Since the arrival rate is 2.78 arrivals per second, the rate of occurrence of inter
arrival times less than 0.01 sec is: (2.78)(0.027) = 0.075 times a second.

31
 Negative Exponential Inter-
arrival Times

 Certain events, such as television commercial breaks, might stimulate the sources to place their
calls at nearly the same time. In this case the negative exponential distribution may still hold, but for
a much higher calling rate during the commercial.

 A more subtle implication of the independent arrival assumption involves the number of sources,
and not just their calling patterns.

 When the probability of an arrival in any small time interval is independent of other arrivals, it
implies that the number of sources available to generate requests is constant.

 If a number of arrivals occur immediately before any subinterval in question, some of the sources
become busy and cannot generate requests.

 The effect of busy sources is to reduce the average arrival rate.

 Thus the inter arrival times are always somewhat larger than what the equation predicts them to be.

 The only time when the arrival rate is truly independent of source activity is  when an infinite
number of sources exist.
32
 Negative Exponential Inter-
arrival Times

 If the number of sources is large and their average activity is relatively low , busy
sources do not appreciably reduce the arrival rate.

 For example, consider an end office that services 10.000 subscribers with 0.1
erlangs of activity each.

 Normally, there are 1000 active links and 9000 subscribers available to generate
new arrivals.
 If the number of active subscribers increases by an unlikely 50% to 1500 active
lines, the number of idle subscribers reduces to 8500, a change of only 5.6%.

 Thus the arrival rate is relatively constant over a wide range of source activity. 
Whenever the arrival rate is fairly constant for the entire range of normal source
activity, an infinite source assumption is justified.

33
 Poisson Arrival Distribution

 Exponential equation merely provides a means of determining the distribution of

inter arrival times.

 It does not, by itself, provide the generally more desirable information of how many
arrivals can be expected to occur in some arbitrary time interval.

 Using the same assumptions presented, however, the probability of “j” arrivals in an
interval t can be determined by:

 The equation is the well-known “Poisson probability law” . When j=0, the probability
of a no arrivals in an interval t is P0(t), as obtained in “negative exponential
equation”.

34
 Poisson Arrival Distribution

 Poisson equation assumes - arrivals are independent and occur at a given average
rate A, irrespective of the number of arrivals occurring just prior to an interval in
question.

 Thus the Poisson probability distribution should only be used for arrivals from a
large number of independent sources.

 The equation defines the probability of experiencing exactly j arrivals in t sec.

 Usually there is more interest in determining the probability of j or more arrivals in t

sec:

where Pi( λ t) is defined in Poisson equation.

35
 Example

 Example 7: Given a message-switching node that normally experiences four

arrivals per minute, what is the probability that eight or more arrivals occur in an
arbitrarily chosen 30-sec interval?

 Solution:
The average number of arrivals in a 30-sec interval is:

The probability of eight or more arrivals (when the average is 2) is

36
 Example

 Example 8: What is the probability that a 1000 bit data block experiences exactly
four errors while being transmitted over a transmission link with a bit error rate
(BER) = 10-5?

 Solution:

Assuming independent errors (a questionable assumption on many transmission

links), we can obtain the probability of exactly four errors directly from, the Poisson
distribution. The average number of errors (arrivals)

λt = (103)10-5 = 0.01.

Thus, probability (4 errors) = P4(0.01) = [(0.01)4/4!]e-0.01 = 4.125(10)-10

37
 Holding Time Distributions

 The second factor of traffic intensity is the average holding time, tm.

 In some cases the average of the holding times is all that needs to be known about
holding times to determine blocking probabilities in a loss system or delays in a
delay system.

 In other cases it is necessary to know the probability distribution of the holding

times to obtain the desired results.

 The two most commonly assumed holding time distributions are:

 constant holding times

 exponential holding times.

38
 Constant Holding Time

 Constant holding times cannot be assumed for conventional voice conversations.

 It is a reasonable assumption for such activities as:

 per-call call processing requirements,
 interoffice address signaling,
 operator assistance and
 recorded message playback.

 Constant holding times are obviously valid for transmission times in fixed length
packet networks.

 When constant holding time messages are in effect, it is straightforward to use

Poison equation to determine the probability distribution of active channels.

39
 Constant Holding Time

 Assume, for the time being, that all requests are serviced, then the probability of j
channels being busy at any particular time is merely the probability that j arrivals
occurred in the time interval of length tm immediately preceding the instant in
question.

 Since the average number of active circuits over a time is the traffic intensity A =
λtm, the probability of j circuits being busy is dependent only on the traffic intensity:

40
 Exponential Holding Time

 The most commonly assumed holding time distribution ‘for conventional telephone
conversations’ is the ‘exponential holding time’ distribution:

 where tm is the average holding time.

 The equation specifies the probability that a holding time exceeds the value t.

 This relationship can be derived from a few simple assumptions concerning the
nature of the call termination process.

 Its basic justification lies in the fact that observations of actual voice conversations
exhibit a remarkably close correspondence to an exponential distribution.

41
 Exponential Holding Time

 The exponential distribution possesses the curious property that:

 the probability of a termination is independent of how long a call has been in progress.
 No matter how long a call has been in existence, the probability of it lasting another t sec
is defined by the equation.

 In this sense exponential holding times represent the most random process that is
possible.

 Not even knowledge of how long a call has been in progress provides any
information as to when the call will terminate.

42
 Exponential Holding Time

 Combining a Poisson arrival process with an exponential holding time process to obtain the
probability distribution of active circuits is more complicated than it was for constant holding
times because calls can last indefinitely.

 The result proves to be dependent on only the average holding time.

 Thus the constant holding time equation is valid  for exponential holding times also (or
any holding time distribution).

 The probability of j circuits being busy at any particular instant, assuming a Poisson arrival
process and that all requests are serviced immediately is

where A is the traffic intensity in erlangs.

 This result is true for any distribution of holding times.

43
 Example

 Example 9: Assume that a trunk group has enough channels to immediately carry
all of the traffic offered to it by a Poisson process with an arrival rate of one call per
minute. Assume that the average holding time is 2 min. What percentage of the
total traffic is carried by the first five circuits, and how much traffic is carried by all
remaining circuits? (Assume that the traffic is always packed into the lowest-
numbered circuits.)
 Solution:

The traffic intensity (offered load) of the system is A= (1)(2) = 2 erlangs. The traffic
intensity carried by i active circuits is exactly i erlangs. Hence the traffic carried by the
first five circuits can be determined as follows:

All of the remaining circuits carry (2 - 1.89) = 0.11 erlangs.

44
 Example

 The result of the previous Example demonstrates the principle of diminishing

returns as the capacity of a system is increased to carry greater and greater
percentages of the offered traffic.

 The first five circuits in the example carry 94.5% of the traffic while all remaining
circuits carry only 5.5% of the traffic.

 If there are 100 sources. 95 extra circuits are needed to carry the 5.5%.

END

45