Traffic Engineering

• Telecom system to service the voice traffic and data traffic . The
traffic is defined as the occupancy of server .
• The basic purpose of the traffic engineering is to determine the
condition under which adequate service is provided to the
subscribers. While using economical use of resources providing
the service.
• The functions performed by the telecom network depend on
applications it handles. Some major functions are ,switching ,
routing , flow control, security, failure monitoring , traffic
monitoring , internetworking and network management.
• To perform the above functions a telephone network composed
of variety of communication equipment such as digit receiver,
call processing , inter-stage switching links
 Traffic Engineering
and inter office links etc.
• The telecom engineering provides basis for analysis and design of
telecom networks . It provides means to determine the quantum of
common equipment required.
• To provide particular level of service for given traffic pattern and
volume. The developed model is capable of providing best
accessibility and greater utilization of their lines and links.
• The traffic engineering also determine the ability of telecom
network to carry given traffic at particular loss probability.
Traffic theory and queuing theory is used to estimate the
probability of occurrence of call blocking.
• Earlier traffic analysis based purely on analytical approach that
involves advanced mathematical concepts .
 Traffic Engineering
present day approach combine the advent of powerful and
affordable software tools that aim to implement traffic
engineering concepts and automatic engineering tasks.
 Traffic Design Requirements
• In the study of traffic engineering , to model a system
and to analyze the change in traffic after designing , the
statistics of an exchange should be studied .
• The incoming traffic undergoes variation in many ways ,
due to peak hours , business hours ,seasons ,weekends ,
festival and location of exchange etc.
• The traffic is unpredictable and random in nature. The
traffic characteristics of an exchange should be analyzed
for system design .
• The grade of service and blocking probability are also
important parameters for traffic study.
 Traffic Design Requirements
1 Traffic Statistics
• Following are the traffic statistics which are important for
analysis and design of network.
A Calling rate
• This is average number of requests for connections that
are made per unit time . If the instant in time that a call
request arises is random variable.
• The calling rate is stated that as the probability that a call
request will occur in certain short interval of time . If n is
the average number of calls to and from a terminal
during T seconds. The calling rate is defined as
λ = n/T (1)
• In telecom system voice traffic and data traffic are two
types of traffics . The calling rate (λ) is also referred as
average arrival rate . The average calling rate is measure
in calls per hour.
 Traffic Design Requirements
B Holding Time
• The average holding time or service time h is average
duration of occupancy of traffic path by a call. For voice
traffic it is average holding time per call in hours.
• The reciprocal of average holding time referred to as
service rate (µ) is call per hour .
µ = 1/h (2)
• The probability of call lasting at least t sec is given by
P (t) = exp (- t/h ) (3)
• For mean holding time h = 100 sec the negative
exponential distribution function is give by figure 8.1 .
The fig shows that 50 % probability calls last longer
than 70 sec.
 Traffic Design Requirements

50 %

70
 Traffic Design Requirements
C The distribution of Destination
• Number of calls receiving at an exchange may be
destined to its own exchange or remote exchanges or
foreign exchange . The distribution is defined as
probability calls request for particular destination
• The hierarchical structure of telecom network includes
many intermediate exchanges , the knowledge of this
parameter helps in determining the number of trunks
needed between individual exchange.
D User Behavior
• The user may abandon a request if his 1st attempt to
make a call is failed. The user may make repeated
attempts to make a call . The user may wait for some
times to make next attempt to setup a call.
• The behavior varies person to person and also depend on
situation.
 Traffic Design Requirements
E Average Occupancy
• If the average number of calls to and from a terminal
during time t sec is n and average holding time is h . The
average occupancy of terminal is given by.
A = nh / T = λh = λ/µ (4)
• Average occupancy is ratio of average arrival rate to the
average service rate. It is measured in Erlangs . Average
occupancy is referred as traffic flow or traffic intensity
or traffic carried.
 Traffic Design Requirements
2 Traffic Pattern
• The telephone traffic and its distribution with respect to time
(traffic load) , which is normally 24 hours . It helps in determining
the amount of lines required to fulfill the subscribers needs.
• According to the need of subscribers the telephone traffic varies
greatly. The variations are not uniform and varies season to
season, month to month , day to day and hour to hour.
• The degree of hourly variation is greater than that of any other
period. Fig 8.2 shows the variation of calls from 0800 hrs to mid
night (00:00 hrs).
• If the behavior of the traffic shown in fig is symmetric for period
of time or season ,good judgment about the design of switching
system or lines or trunks or any shared equipment can be made.
Number of outgoing calls
Traffic Design Requirements

Fig. 8.2: Variations of calls from 8.0 A.M. to mid night

 Traffic Design Requirements
• The combination of historical records , experience ,
location of exchange (business area or residential area ) ,
vacations , holidays etc. decide the design of the of telecom
network . Various parameters related to traffic pattern are
below:
a. Busy Hour
• The telecom facilities are designed on the intensity of
traffic during the busy hour . The busy hour vary from
exchange to exchange , month to month and day to day etc.
• The busy hour is defined as the 60 minutes interval in a
day in which traffic is highest.
b. Call Completion Rate
• Based on the status of the subscribers or the design of the
switching system , the call attempt may be successful or
not.
 Traffic Design Requirements
• The call completion rate is defined as the ratio of number
of successful calls to the number of call attempts. A CCR
value of 75 % is considered as excellent and 70 % is
usually expected.
c. Busy Hour Call Attempts (BHCA)
• It is an important parameter in deciding the processing
capacity of an exchange . It is defined as number of calls
attempt in busy hour.
d. Busy Hour Calling Rate (BHCR)
• It is useful parameter in designing a local exchange to
handle the peak hour traffic . It is defined as the average
number of calls originated by a subscriber during busy
hour.
 Traffic Design Requirements
3 Unit of Telephone Traffic
• The traffic is measured in Erlangs , its name after Danish
Mathematician Agner Erlangs . Erlang is international
unit of traffic .
• A server is said to have traffic of 1 Erlang if it is occupied
for entire period of observation. Simply 1 Erlang
represents 1 circuit occupied for 1 complete hour . The
maximum capacity of a single server (or channel) is 1
Erlang (server is always busy). Thus the maximum
capacity in Erlangs of a group of servers is merely equal
to the number of servers. The traffic intensity is ratio of
period for which the server is occupied to the total period
of observation is measured in Erlangs.
 The Unit of Traffic
• The traffic intensity is defined as the average number of
calls in progress per unit time.
• The unit of traffic is Erlang (E) name after Agner Erlang
the pioneer of traffic theory.
• On a group of trunks , the average number of calls in
progress depend on both the number of calls which arrive
and their duration.
• The duration of call is always called holding time , because
it holds a trunk for that period of time.
• The Erlang of traffic carried by the group of trunks is
given by :
A = n h/ T (1)
Where A is traffic in Erlangs, n is average number of calls
arrival during time T and h is average call holding time.
 The Unit of Traffic
Example
On average ,during the busy hour a company makes
120 outgoing calls of average duration 2 minutes. It
receives 200 incoming calls of average duration 3 minutes.
Find :
1- Outgoing traffic
2- Incoming traffic
3- total traffic .
Solution :
1 Outgoing traffic = 120 * 2/60 = 4 E
2 Incoming traffic = 200 * 3/60 = 10 E
3 Total traffic = 4 E + 10 E = 14 E.
 The Unit of Traffic
Example
During the busy hour , on average a customer with a
single telephone link makes 3 calls and receive 3 calls. The
average duration is 2 minutes . What is the probability that
a caller will find the line engage .
Solution :
Occupancy of line = (3 + 3 ) * 2/60 = 0.1 E
0.1 E is the probability of finding the line engage.
Example :
If a group of 20 trunks carries 10 Erlangs and the average call
duration is 3 minutes . Calculate (a) Average number of calls in
progress (b) Total number of calls originated per hour.
Solution :
No. of trunks = 20
Traffic intensity (A) = 10 E
Holding time (h) = 03 minutes
Observation time (T) = 60 minutes
(a) Traffic intensity per trunk = 10 E/20 = 0.5 E/Trunk
Average No. of calls per trunk for 1 E for 60 minutes = 20
For 0.5 E average number of calls = 20 x 0.5 = 10
(b) Traffic intensity :
A = n h/T =10 E
Total number of calls originated per hour (n = A x T/h)
n = 10 x 60/3 = 200 calls
 Grade of Service (GOS)
• For non blocking service of an exchange , it is necessary to
provide as many lines as there are subscribers. But it is not
economical . So , some calls have to be rejected and have to
be retried when lines have been used by other subscribers.
• The grade of service refer to proportion of unsuccessful
calls . GOS is defined as ratio of lost traffic to offered
traffic.
GOS = Blocked busy hour call / Offered busy hour calls
GOS = A-A0/A
A0 = Carried traffic
A = Offered traffic
A-A0 = Lost traffic
 Grade of Service (GOS)
• The smaller the value of GOS ,the better is the service.
• Recommended GOS is 0.002 i.e. 2 calls per 1000 offered
may lost . In system with equal number of servers and
subscribers , GOS is equal to zero.
Example:
During a busy hour ,1400 calls were offered to
a group of trunks and 14 calls were lost . The average call
duration is 3 minutes . Find (a) Traffic offered (b) Traffic
carried ( c ) GOS (d) Total duration of congestion period .
Solution :
n = 1400 calls
h = 3 minutes
T = 60 minutes
( a) Traffic offered = A = n h/T = 1400 x 3 /60 = 70 E
 Grade of Service (GOS)
(b) Traffic carried A0 = 1386 x 3/60 = 69.3 E
(c ) GOS = A – A0/ A
A – A0 = 70 – 69.3 = 0.7 E (lost traffic )
GOS = 0.7 /70 = 0.01
(d ) Total duration of congestion = 0.01 x 3600 = 36 second.
 Congestion
• It is uneconomical to provide sufficient equipment to carry
all traffic that could possibly be offered to telecom system.
• In a telephone exchange it is theoretically possible for
every subscriber to make a call simultaneously . The cost
of meeting this demand is not possible ,but probability of it
happening is negligible .
• The situation arises that all the trunks in a group are busy,
so it can accept no further calls . This is known as
congestion.
• In message switching system , calls that arrives during
congestion wait in a queue until an outgoing trunks
becomes free. They are delayed but not lost. Such systems
are called queuing system or delay system.
 Congestion
• In circuit switching system such as telephone exchange , all
attempts to make calls over a congested trunks group are
unsuccessful such systems are called lost call systems.
• In a lost call system , the result of congestion is that the
traffic actually carried is less than the traffic offered to
system.
Traffic carried = Traffic offered – traffic lost.
• For lost call system the grade of service B is defined as :
B = No. of calls lost ÷ No. of calls offered
= Portion of time for which congestion exist
= Probability of congestion
= Probability that a call will lost due to congestion.
 Congestion
• If traffic Erlangs is offered to a group of trunks having
grade of service is B the traffic lost is AB and traffic
carried is A (1-B ) Erlangs .
• Larger is the grade of service , the worse is the service
given , the worse service is normally specified for traffic at
busy hour, for other time it is much better.
• If B is too large , user make many unsuccessful calls . If it
is too small unnecessary expenditure is incurred on
equipment which is rarely used.
 Congestion
Example
During the busy hour , 1200 calls were offered to a
group of trunks and 6 calls were lost . The average call
duration is 3 minutes. Find :
1- The traffic offered , 2- The traffic carried
3- The traffic lost , 4- The grade of service
5- The total duration of period of congestion.
Solution
1- A = n * h/ T = 1200 * 3/60 = 60 E
2- 1194 * 3/60 = 59.7 E
3- 6 * 3/60 = 0.3 E
4- B = 6/1200 = 0.005
5- 0.005 * 3600 = 18 Seconds
 Lost Call System
• Erlang determine the grade of service ( i.e. loss
probability) of a lost call system having N trunks when
offered traffic A are shown in fig below. Solution depend
on following assumptions.

Traffic N
offered A Outgoing
Erlang trunks

Fig 4.5 lost call system

- Pure chance traffic

- Statistical equilibrium
- Full availability
- Calls which encounter congestion are lost.
 Lost Call System
• The assumption of pure chance traffic implies that call
arrival and termination are independent random event. If
call arrivals are independent random events, their
occurrence is not affected by previous calls. This traffic is
therefore sometimes called memory less traffic .
• Statistical equilibrium implies that the probabilities do not
change.
• Full availability means that every call that arrives can be
connected to any outgoing trunk which is free.
• Each switch have sufficient outlets to provide access to
every outgoing trunk.
• The lost call assumption implies that any attempted call
which encounter congestion is immediately cleared from
system. When this happen user is likely to make another
attempt shortly afterwards.
• The traffic offered during busy hours is slightly greater
than that if it would have been there were no congestion.
 Lost Call System
• If there are x calls in progress , then

P(x) = A x P (0) (1)

x!

There will not be a negative number of calls and there can not
be more than N calls so 0 ≤ x ≤ N so
N A
x

N ∑ x! P (0)
∑ P (x) = 1 = X = 0
X=0 1

N Ax (2)
Hence P (0) =
∑ x!
X=0
 Lost Call System
• The probability of lost call which is grade of service B .
This is given the symbol E 1.N (A) which denote the loss
probability for full availability group of N trunks offered A
erlangs.
(3)
B = A /N!
N
N
∑ AK/K!
K=0

The grade of service of loss system with N full availability

trunks
, offered A erlang of traffic is given by E 1.N (A) . From equation
3 E 1.N -1 N-1
= A∑N-1 A /K!
K
/( N-1 )!
K=0
 Lost Call System
N
A / K!
K A N-1 / (N-1)! + AN / N!
∑ =
K=0 E 1.N-1 (A)

Substituting in equation (3 )

E 1.N (A) = A E 1.N-1 (A)

N + AE1.N-1 (A)
 Lost Call System
Example
A group of 5 trunks is offered 2 erlang of traffic , find.
1- The grade of service
2- The probability that only one trunk busy
3- The probability that only one trunk free
4- The probability that at least one trunk free
Solution
1- From equation (3) B = E 1.N (A) = 25 / 5!
1 + 2/1 + 4/2 + 8/6+ 16/24 + 32/120

= 0.2667/ 7.2667 = 0.037

2- From equation (2)
P (x) = Ax/ x!
N
∑ AK/ K!
 Lost Call System
P (1) = 21/1!
= 2 / 7.2667 = 0.275
7.2667
3- p (4) = 24/4! = 16/24
7.2667 = 0.0917
7.2667
4- P (x < 5) = 1 – P(5) = 1-B
= 1 – 0.037 = 0.963
 Queuing Systems
• Erlang determined the probability of encountering delay when
traffic A is offered to a queuing system with N trunks as shown
in fig below.
• Erlang’s solution depend upon following assumptions:
1. Pure chance traffic
2. Statistical equilibrium
3. Full availability
4. Calls which encounters congestion enter a queue and are stored
there till a trunk become free.

Queue
N
Traffic
Trunks
offered A
Erlangs
Fig . Queuing System
 Queuing Systems
• The assumption of pure chance traffic implies that call arrival
and termination are independent random event. If call arrivals
are independent random events, their occurrence is not affected
by previous calls. This traffic is therefore sometimes called
memory less traffic .
• Statistical equilibrium implies that the probabilities do not
change.
• Full availability means that every call that arrives can be
connected to any outgoing trunk which is free.
• If A ≥ N calls are entering the system at a greater rate than they
leave. As a result ,the length of the queue must continually
increase toward infinity.
 Queuing Systems
• Let x be total number of calls entering the system. When
x < N , then x calls are being served and there is no delay .
When x > N , then all the trunks are busy and incoming
calls encounter delay , there are N calls being served and x-
N calls in queue.
• If x ≤ N :
There is no queue and there is behavior of system is same
as that of a lost call system in the absence of congestion.
P (x) = Ax
P ( 0 ) For 0 < x < N (4.9)
x!
• If x ≥ N :
The probability of call arrival in a very short period
of time , δt , from equation on next slide :
 Queuing Systems
P(a) = A δt /h
where h is mean service time
• The probability of transition from x-1 to x calls in the system during δt is
given by :
P (x-1 → x ) = P (x-1) A δt /h
Since all trunks are busy , only N calls are being served can terminate
P (e) = N δt /h
and the probability of a transition from x to x-1 is given by
P ( x → x-1 ) = P(x) P(e) = P (x) δt /h
• For statistical equilibrium , P (x-1 → x) = P (x→ x-1).
Therefore P (x) N δt /h = P (x) A δt /h
and

P(x) = A
P (x-1)
N (4.10)
 Queuing Systems
P (N) = AN P (0) From equation 4.7
N!
P (N+1) = A A N+1
P (N) = P (0)
N N. N!
A A N+2
P (N+2) = P (N+1) =
N P (0)
N2. N!
x
For x ≥ N : A
x NN P(0) (4.11)
P (x) = A P (0) = N
N!
Nx-N . N!
If there is no limit to the possible length to the queue , then x
can have any value between zero and infinity.
∞
∑ P (x) = 1
X=0
 Queuing Systems
• Thus from equation (4.7 ) and (4.11) :
1 N -1
= NN A N ∞ k
P (0) ∑ A x
(4.12)
X=0 xỊ + NỊ N ∑
K=0
Where k = x – N , since A/N ≤ 1 then :

∞ A k A -1
∑ = 1 -
N
K=0 N

1 N -1
= AN A -1
P (0) ∑ Ax + 1
NỊ N
X=0 xỊ
P(0) = NAN + N -1 X -1
NỊ (N –A ) A (4.13)
∑
X = 0 XỊ