Acta Electrotechnica et Informatica Vol. 8, No.

3, 2008, 5356 53


ISSN 1335-8243 2008 FEI TUKE
EVALUATION OF REAL CALL SET UP SUCCESS RATE IN GSM


Martin KOLLR
TMM RA 5 MDC, Siemens Program and System Engineering, Ltd. ,
Lomen 1, 041 00 Koice, tel. +421 55 7248 243, E-mail: [email protected]


ABSTRACT
The Call Set up Success Rate is one of the most important Key performance Indicators (KPIs) used by all mobile operators.
However there is no standard measurement possible for this parameter. Therefore the different operators can measure it differently.
In this paper, a definition of a real Call Set up Success Rate, possibility of its implementation using the current technologies in
GSM and difference between the real and implemented Call Set up Success rate is provided. The real in this case means that Call
Set up Success Rate is calculated as ratio of the assigned TCHs to the channel requests. The error in measurement of real Call Set up
Success Rate with activated Direct TCH Assignment feature is evaluated as well. The proposed error modeling can be used in order
to guarantee a quality in measurement of Call Set up Success Rate.

Keywords: GSM, Call Set up Success Rate, Error, TCH Assignment, Direct TCH Assignment


1. BASIC STRUCTURE OF GSM NETWORK

A part of GSM network is shown in Fig. 1. The Base
Transceiver Station, or BTS, contains the equipment for
transmitting and receiving of radio signals (transceivers),
antennas, and equipment for encrypting and decrypting
communications with the Base Station Controller (BSC).
Typically a BTS will have several transceivers (TRXs)
which allow it to serve several different frequencies and
different sectors of the cell [1, 4].



Fig. 1 Structure of GSM network

The Base Station Controller (BSC) provides,
classically, the intelligence behind the BTSs. The BSC
handles allocation of radio channels, receives
measurements from the mobile phones, controls
handovers from BTS to BTS (except in the case of an
inter-BSC handover in which case control is in part the
responsibility of the Anchor MSC). A key function of the
BSC is to act as a concentrator where many different low
capacity connections to BTSs (with relatively low
utilisation) become reduced to a smaller number of
connections towards the Mobile Switching Center (MSC)
via A interface [1 - 6].
2. CALL SET UP IN GSM

The successful call set up consists of two procedures.
The simplified description of these procedures is provided
in the next text in such a way that the focus is only on the
parts necessary to understand the philosophy of Call Set
up Success Rate calculation correctly.
First procedure is Immediate Assignment procedure
which is used to create a signaling connection between the
Mobile station (MS) and the network. It can be initiated
only by the MS sending a CHANNEL REQUEST
message on the Random Access channel (RACH) to the
BTS that it requires a signaling channel (SDCCH). This
message contains the information field establishment
cause and random reference. The establishment cause
gives the reason why the MS is requesting a SDCCH [2].
Possible reasons are:
- emergency call
- call re-establishment
- answer to paging
- originating speech call
- originating data call
- location updating
- other procedures, which can be completed with
an SDCCH.

Then it comes next signalization between the MS and
network in order to activate the signaling channel,
recognize the service being requested by the MS, etc. The
successful seizure of SDCCH is acknowledged by sending
the Establish Indication message from MS to BTS and
then to BSC. Further coordination procedure
(authentication, ciphering etc.) are now performed on the
SDCCH [2, 5, 6].
Second procedure is Assignment procedure which is
used to occupy a radio resource (speech channel). The
MSC is initiator of this procedure. The MSC sends an
ASSIGNMENT REQUEST message to the BSC
requesting the assignment of a radio resource (RR). Then
it comes next signalization between BTS and BSC in
order to allocate and activate a suitable RR (Traffic
channel - TCH). If the TCH is successfully seizured by
54 Evaluation of Real Call Set up Success Rate in GSM

ISSN 1335-8243 2008 FEI TUKE
MS, the BSC sends the ASSIGNMENT COMPLETE
message [2, 5, 6].
Fig. 2 summarizes the above described procedures.

MS BTS BSC MSC
CHANNEL REQUEST
Further signalization in order to activate the signaling channel, recognize the
service being requested by the MS
ESTABLISH INDICATION
Further coordination procedure (authentication, ciphering etc.) are now performed
on the SDCCH.
ASSIGNMENT REQUEST
Further signalization between BTS and BSC in order to allocate and activate a
suitable RR (Traffic channel - TCH).
ASSIGNMENT COMPLETE
ASSIGNMENT COMPLETE


Fig. 2 Message flows for the call set up in GSM.

Note:
Some of the vendors (for example Siemens) provide
the possibility to assign TCH directly. In this case the
result of Immediate Assignment procedure is occupation
of Fast Associated Control channel (FACCH)1 instead of
SDCCH. The Assignment procedure for the Direct TCH
assignment stays unchanged. On the Siemens side the
Direct TCH assignment is used for emergency calls when
usually authentication and ciphering are excluded and the
call set up is provided up to 1 second instead of standard
time frame from 3 to 4 seconds.

3. DEFINITION OF REAL CALL SET UP
SUCCESS RATE AND COMPARISON WITH
THE EXISTING ONE

As mentioned in the previous paragraph the successful
call set up consists of the following parts:
- Successful Immediate assignment procedure (the
result is occupation of SDCCH or FACCH in
case of Direct TCH assignment)
- Successful authentication and ciphering on
SDCCH or FACCH (these procedures can be
excluded in case of Direct TCH assignment)
- Successful TCH assignment

Therefore the Real Call Set up Success Rate (CSSR) is
given by the formula:

1
It is called as Direct TCH Assignment despite the result of
Immediate Assignment procedure is occupation of FACCH
because to transmit some information on FACCH means to
transmit signaling blocks on TCH. During the conversation
phase (TCH is occupied) the user data blocks (speech) are
transmitted.
qSpeech Re NumCH
g NumTCHAssi
CSSR =
, (1)

where NumTCHAssig represents the number of
successfully assigned TCH (number of ASSIGNMENT
COMPLETE messages in the Fig. 2) and
NumCHReaqSpeech represents the number of CHANNEL
REQUEST messages (see Fig. 2) but related only to
request for a mobile originated (MO) or mobile terminated
(MT) call. The other procedures, which can be completed
with an SDCCH, like SMS MT, SMS MO, location
updating etc. are not counted because they do not
represent the request for the speech call. In other words,
the Call Set up Success Rate represents for how big part
from the total call attempts the call set up was successful.
Practical implementation of the Eq. 1 means problem
because up to now it is not possible to distinguish between
the requests for the speech call and other ones. In fact, for
example the establishment cause equal to sequence
111xxxxx1, where x is 1 or 0, can represent originating
call or procedures that can be completed with a SDCCH
according to [3].
One of the possibilities how to solve this problem
could be using of simplified formula:

*
NumTCHAssig
CSSR
NumCH Re q
=

(2)

where NumCHReq represents total number of CHANNEL
REQUEST messages and NumTCHAssig represents the
number of TCH assignments (number of ASSIGNMENT
Complete messages in the Fig. 2). Let us consider that
NumCHReq = NumCHReqSpeech +
NumCHReqNonSpeech, where
NumCHReqNonSpeech is the number of CHANNEL
REQUEST messages not used for MT or MO speech call.
Then the Eq. 2 can be modified to the next form:

qNonSpeech Re NumCH qSpeech Re NumCH
g NumTCHAssi
CSSR
*
+
=
.
(3)
Under condition that
NumCHReqNonSpeech/NumCHReqSpeech 20 % the
formula can be modified using binomial series as follows:

qSpeech Re NumCH
qNonSpeech Re NumCH
qSpeech Re NumCH
g NumTCHAssi
CSSR
*
1
. (4)

Afterwards the absolute error in measurement of Call Set
up Success Rate using Eq. 2 will be

qSpeech Re NumCH
qNonSpeech Re NumCH
* CSSR
q Re NumCH
g NumTCHAssi
qSpeech Re NumCH
qNonSpeech Re NumCH
qSpeech Re NumCH
g NumTCHAssi
CSSR CSSR
*
==

= = 1
(5)
In case that NumCHReqNonSpeech is equal to zero the
Eq. 2 provides exactly the Call Set up Success Rate but
this is not real case because in practice the ratio:
NumCHReqNonSpeech/NumCHReqSpeech is on the
range of tenths of percent which can lead to the big
systematic error. Therefore the mobile operators break
Acta Electrotechnica et Informatica Vol. 8, No. 3, 2008 55

ISSN 1335-8243 2008 FEI TUKE
away from using of the Eq. 2. In principle the Eq. 2 can be
used for calculation of Call Set up Success Rate only in
regions with the
NumCHReqNonSpeech/NumCHReqSpeech 1 %.
Second possibility could be to use in the denominator
of the Eq. 1 the number of TCH attempts
(NumTCHAttempt) instead of NumCHReqSpeech.
NumTCHAttempt represents the number of ASSIGNMENT
REQUEST messages (see Fig. 2). But in this case the
result of the calculation will be TCH Assignment Success
rate which is something different than Call Set up Success
Rate. Even some of the operators have separate KPIs for
Call Set up Success Rate and TCH Assignment Success
Rate.
It seems that the best approach promises indirect
calculation of NumCHReqSpeech according to model in
Fig. 3.

ImmAssSuccRate SDCCHSuccRate NumCHReqSpeech
NumTCHAttempt
Immediate Assignment
Procedure
Authentication, ciphering etc.
performed onthe SDCCH

Fig. 3 Model for calculation of NumCHReqSpeech

In Fig. 3, ImmAssSuccRate represents Immediate
Assignment Success Rate given by formula:

Im
NumEstInd
mAssSuccRate
NumCH Req
=

, (6)

NumEstInd represents the number of ESTABLISH
INDICATION messages (see Fig. 2). In other words, the
Immediate Assignment Success Rate represents how big
part from the total number of requests for channel was
successful during the immediate assignment procedure.
SDCCHSuccRate represents SDCCH Success Rate given
by formula:

ate SDCCHDropR ate SDCCHSuccR =1 , (7)

where SDCCHDropRate is SDCCH Drop rate and
provides how big part from the total number of SDCCH
dropped during the procedures (authentication, ciphering
etc.) performed on SDCCH.
From the model in Fig. 3 then we have

qSpeech Re NumCH *
* ate SDCCHSuccR * te mAssSuccRa Im mpt NumTCHAtte =
(8)

From Eq. 8 then for NumCHReqSpeech we can write that

ate SDCCHSuccR * te mAssSuccRa Im
mpt NumTCHAtte
qSpeech Re NumCH =
. (9)

Including the Eq. 9 into the Eq. 1 it follows that
ate SDCCHSuccR * te mAssSuccRa Im *
mpt NumTCHAtte
g NumTCHAssi
CSSR =

(10)
which can be also expressed in the form
ate SDCCHSuccR * te mAssSuccRa Im * Rate TCHAssSucc CSSR =
(11)
because
mpt NumTCHAtte
g NumTCHAssi represents TCHAssSuccRate as it
was early mentioned. Currently the Eq. 11 is the best
approach and is provided for example by Siemens. A
disadvantage can be higher effort on BSC or equipment
(for example an offline tools) where the Call set up
Success rate is to be calculated because three KPIs (or six
partial measurements) enter the Eq. 11. The Eq. 11
provides exactly the Call Set up Success Rate in case the
Direct TCH Assignment feature is disabled. Possible
distortion when this feature is enabled is described in the
next paragraph.

4. CALL SET UP SUCCESS RATE WHEN DIRECT
TCH ASSIGNMENT IS ENABLED

As mentioned in the paragraph Call Set up in GSM
in case when Direct TCH Assignment feature is enabled
the result of Immediate Assignment procedure for some of
channel requests can be occupation of FACCH instead of
SDDCH. The Direct TCH Assignment is used in order to
decrease time needed for call set up in such a way that
authentication and ciphering are excluded. It means that
FACCH is occupied only very short time in comparison to
SDCCH occupation which leads to FACCH Success Rate
close to 100 %. The model for calculation of
NumCHReqSpeech is for this case presented in Fig. 4
where
NumCHReqSpeechSDCCH and
NumCHReqSpeechFACCH represent those Channel
requests that result in occupation of SDCCH and FACCH,
respectively when Immediate Assignment procedure has
been finished.

ImmAssSuccRate SDCCHSuccRate
NumCHReqSpeechSDCCH
+
NumCHReqSpeechFACCH
NumTCHAttempt
Immediate Assignment
Procedure
Authentication, ciphering etc.
performed on the SDCCH
+
FACCH occupied
SDCCH occupied

Fig. 4 Model for calculation of NumCHReqSpeech when Direct
TCH assignment is enabled

From the model in Fig. 4 it follows

te mAssSuccRa Im * CH qSpeechFAC Re NumCH ate SDCCHSuccR *
* te mAssSuccRa Im * CH qSpeechSDC Re NumCH mpt NumTCHAtte
+ +
=

(12)
which can be also expressed in the form
56 Evaluation of Real Call Set up Success Rate in GSM

ISSN 1335-8243 2008 FEI TUKE

+
=
ate SDCCHSuccR
CH qSpeechFAC Re NumCH
CH qSpeechSDC Re NumCH
* ate SDCCHSuccR * te mAssSuccRa Im mpt NumTCHAtte
(13)

Using the Eq. 7 we have

+
=
ate SDCCHDropR
CH qSpeechFAC Re NumCH
CH qSpeechSDC Re NumCH
* ate SDCCHSuccR * te mAssSuccRa Im mpt NumTCHAtte
1
(14)
Since SDCCHDropRate does not exceed 0.5 % in practice
we can modify it using binomial series as follows:

ate SDCCHDropR * CH qSpeechFAC Re NumCH *
* ate SDCCHSuccR * te mAssSuccRa Im qSpeech Re NumCH *
* ate SDCCHSuccR * te mAssSuccRa Im mpt NumTCHAtte
+

(15)

and



(16)


Including the Eq. 16 into the Eq. 1 we have

+
=
ate SDCCHDropR *
qSpeech Re NumCH
CH qSpeechFAC Re NumCH
*
* ate SDCCHSuccR * te mAssSuccRa Im * Rate TCHAssSucc CSSR
* *
1
(17)

The Eq. 17 represents exact formula that should be
used for calculation of Call Set up Success rate when
Direct TCH Assignment feature is enabled. However the
ratio
qSpeech Re NumCH
CH qSpeechFAC Re NumCH currently cannot be
measured (see the paragraph Definition of Real Call Set
up Success Rate and Comparison with the Existing one)
therefore the Call Set up Success rate must be calculated
using the Eq. 11 with the absolute error

ate SDCCHDropR *
*
qSpeech Re NumCH
CH qSpeechFAC Re NumCH
* * ate SDCCHSuccR *
* te mAssSuccRa Im * Rate TCHAssSucc
CSSR CSSR
* *
=
= =
(18)

and relative error

ate SDCCHDropR *
qSpeech Re NumCH
CH qSpeechFAC Re NumCH
CSSR
=

=
(19)

From the Eq. 19 it follows that the relative error in the
worst case will be equal to SDCCHDropRate which is
maximum 0.5 %.

5. CONCLUSION

The number of channel requests for speech calls
cannot be measured therefore there is no simple formula
for calculation of Call Set up Success Rate. Hence more
complex formula into which enter Immediate Assignment
Success rate, TCH Assignment Success Rate and SDCCH
Success rate must be used. Currently it is the best
approach despite higher effort on the processor part in
equipment where the Call Set up Success Rate is to be
calculated is expected. Additionally it does not cover the
case when Direct TCH Assignment feature is enabled.
However as it has been shown in this paper the relative
error in the worst case will not exceed 0.5 %. The
proposed formula has been implemented in the base
stations provided by Siemens.

REFERENCES

[1] Halonen T., Romero J., Melero J.: GSM, GPRS and
EDGE Performance. John Wiley & Sons, Ltd, 2003.
[2] 3GPP specification: 23.108
[3] 3GPP specification: 04.18
[4] Dobo, ., Dha, J., Marchevsk, S., Wieser, V.:
Mobiln rdiov siete. EDIS vydavatestvo U,
2002.
[5] Kyriazakos, S., Papaoulakis, N., Nikitopoulos, D.,
Gkroustiotis, E.: Performance Evaluation of
Operational GSM and GPRS Systems under Varying
Traffic Conditions. IST Mobile and Wirless
telecomunications Summit, 2002, Thessaloniki
Greece.
[6] Kyriazakos, S., Papaoulakis, N., Nikitopoulos, D.,
Gkroustiotis, E.: A Comprehensive Study on
Performance Evaluation of Operational GSM and
GPRS Systems under Varying Traffic Conditions.
IST Mobile and Wirless telecomunications Summit,
2002, Thessaloniki Greece.

Received July 19, 2007, accepted February 12, 2008


BIOGRAPHY

Martin Kollr was born in Spisk Nov Ves, Slovakia,
on 7 December 1974. He gained an Ing. (M.Sc.) degree in
electronics and multimedia telecommunications from the
Faculty of Electrical Engineering and Informatics (FEI),
Technical University (TU) Koice and a Ph.D. degree in
measuring techniques from the FEI TU Koice, in 2000
and 2003, respectively. From 2003 till 2006, he was an
assistant professor at the Department of Theory of
Electrical Engineering and Measurement, FEI, TU Koice.
Since 2006 he has been an analyst -architect at Siemens
Program and System Engineering. His research interests
include performance measurement techniques for GSM
and GERAN, nonlinear circuit theory, smart sensors based
on a flip-flop circuit and testing the analog-to-digital
converters.
* * 1 *
NumTCHAttempt
NumCH Re qSpeech
NumCH Re qSpeechFACCH
ImmAssSuccRate SDCCHSuccRate SDCCHDropRate
NumCH Re qSpeech
=

+