ASSIGNMENT # 02

SUBJECT: Digital
Switching

TOPIC:
Blocking Probability

SUBMITTED TO: Sir.

Saad Mustafa

SUBMITED BY: Naveed

Akbar Mughal

REG #:
SP08-BTN-065
DATE OF SUBMISION: 2nd MAY
2011

Comsats Institute of Information

Technology Abbottabad
Question 1):
What does blocking mean in networks? What is the importance of blocking in
interconnecting devices?

Answer:
Blocking means in network to block certain users to access the resources of network while too
many users are accessing the network simultaneously. A network that has
fewer transmission paths than would be required if all users were to communicate
simultaneously. Blocking networks are used because not all users require service simultaneously.
Certain statistical distributions apply to the patterns of user demand.

Importance of Blocking in Interconnecting Devices:

Blocking is very important in interconnecting devices such as switches, routers, hubs etc.
Blocking is used so that the network utilization can be efficient and effective. Especially it is
used to avoid collision in devices like switches, hubs etc. When the amount of dat that a devices
can handle exceed then there can be loss of data, collision, jitter, delay and other network issues
can be expected which degrade the performance of network. Blocking is implemented to avoid
all of above problems and also the available bandwidth can be efficiently utilize.

Multistage interconnection networks or “link systems” or “Switching Fabrics” are required when
the capacity to be switched in a given node exceeds the capabilities of a single
switching element. When small PABXs or routers may be built from a single switching
processor, larger nodes have to be organized with multiple stages of switching elements,
linked together. Such link systems present the inconvenience of internal blocking, which
means that there is a probability that no paths may be found available at a given moment
between a given inlet and a given outlet. The art of switch design consist in finding link and
switching elements configurations that minimize or even cancel this Internal Blocking
probability. The calculation of internal blocking probabilities in multistage connecting
networks is therefore a fundamental problem in the switching industry for switching fabric
design, when non blocking conditions are not feasible or would be too costly. This used to be
the case with analogue switches; it is presently the case with very large asynchronous
switches like ATM switches or IP routers. In 1950, C. Jacobaeus of LM Ericsson [JAC]
established a method for computing internal blocking probabilities in a point-to-point
selection within a three-stage interconnection network. In 1955, C. Y. Lee of Bell
Laboratories [LEE] proposed a much simpler, although more approximated method, to
compute internal blocking probabilities. The Lee approximation had the advantage of being
easily extended to any number of stages and of giving a worst-case assessment since it usually
overestimates the internal blocking probability. Consequently the Lee method has been very
extensively used in the industry for switching fabric design.

Question 2):
What does blocking probability means? Discuss Lee graphs and Jacobaeus blocking
probabilities?

Answer:
All the switching systems are designed to provide a certain maximum probability of blocking for
the busiest hour of the day. It is one of the aspects of the grade of service of the telephone
company. There are variety of techniques to evaluate the blocking probability of a switching
matrix. Every network has a maximum probability of blocking. Depends on the accuracy,
required availability, geographical area, priority, complexity and applicability of different
network structures, the techniques are varying. Here, two techniques are described.

 Lee graphs. It was proposed by C.Y. Lee. It is a most versatile and straight forward
approaches of calculating probabilities with the use of probability graphs.

 Jacobaeus method. It was presented in 1950 by C. Jacobaeus. It is more accurate than

Lee graph method.

Both these techniques are approximate techniques and provide reasonably accurate result.

Lee graphics:
C.Y. Lee’s approach of determine the blocking probabilities of various switching system is
based on the use of utilization percentage or loadings of individual links.

LEE’s Assumptions:
 1) Traffic Distribution is uniform, constant or Poisson traffic i.e. inter-arrival rate follows a
pattern.

2) The network is considered to be a multiple-server queuing system.

3) In steady state, call arrival rate is equal to departure rate.

4) During the network operation, a certain WORKLOAD is maintained which is defined by

Network Utilization.

Network Utilization is the ratio of “Total no of busy links to no of input ports”.

Let p be the probability that a link is busy. The probability that a link is idle is denoted by

q=1–p
When any one of n parallel links can be used to complete a connection, the blocking probability
B is the probability that all links are busy is given by

n
B = p .....................................(1)
n = no of parallel links used to complete a connection.

when a series of n links are all needed to complete a connection, then

n
B = 1 – q ................................(2)

3 stage n/w:
For a probability graph of three stage network, shown in Fig. below, the probability of blocking
is given by

k
B = (1 – q2) ...........................(3)
where q′ = probability that an interstage link is idle and

q′ = 1 – p′ ...............................(4)
p′ = probability that any particular intersatge link is busy

k = number of centre stage arrays.

If p is known, the probability that an interstage link is busy is given by

p′ = p/β ..................................(5)
where,

β = k/n .....................................(6)
β is the factor by which the percentage of interstage links that are busy is reduced.

Fig. Probability graph of three stage network.

Substituting in (5) in (4) , we set

q′ = 1 –p/β .............................(7)
Substituting (7) in (3) we get complete expression for the blocking probability of a three stage
switch interms p as

B= [1 - [1- p/β]2]k.........(8)
with inlets of 10% busy, the switch size of N with n = 8, h = 5, β = 0.625 requires 2560
crosspoints.
MERITS:
The merits of this method are

(i) It provide accurate results

(ii) Its formulas are directly relate to the network structures

(iii) It provides insight of the network and thus provides ideas to change the structure

for high performance.

Jacobaeus:
The Lee’s graph approach is not much accurate. Because the probability graphs entail several
simplifying assumptions. The important one which gives erroneous values of blocking is the
assumption that the individual probabilities are independent. In fact the probabilities not
independent and highly dependent when significant amounts of expansion are not present.
According to C. Jacobaeus the blocking probability of a three stage switch is

B =[( n!)2/ k! ( 2n- k) !] pk(2-p) .......(9)

where n = number of inlets (outlets) per first (third) stage array

k = number of second stage array

p = inlet utilization.

More accurate techniques can be used for systems with high concentrations and high blocking.
As the high blocking probabilities not having much practical value, those techniques are not
considered here.
References:

 R Syski, Introduction to Congestion Theory.

 S. S. Katz, Improved Networks and End-to-End Probability Theory

 John C. Bellamy, Digital Telephony

 Claude Rigault , CLOS NETWORKS