Fine-tuning of OCB exchanges
It was observed that overall CCR (Call Completion Ratio) of OCB Exchanges (Local,
Level-1 & Level-2 TAXs) could be improved by 5 to 10% by ‘fine-tuning’ alone without any
increase in circuits/equipments. Traffic details and other system performance indicators of
one Level-1 TAX observed for two sample days, one before fine-tuning and another after
fine-tuning, are tabulated below. It can be seen that effective transit traffic (CT) had increased
whereas BHCA had reduced due to reduction in infractuous failed calls.
Sl.No. Traffic/ Performance indicator Before After
1 Transit traffic, CT 11,421 E 12,778 E
2 BHCA (Busy Hour Call Attempts) 5,81,133 5,46,640
3 CCR (Call Completion Ratio) 32.36 38.38
4 Processing Efficiency: T5t/(T5a+T5t) *100 84.64 96.78
5 Circuit Seizure Efficiency: T6t/T4t *100 87.59 93.99
6 Transit Traffic per 1000 BHCA 19.65 23.28
7 Call failure due to Exch. Control regulation 35,914 104
T37a+T37t (See Note)
8 Call failed due to circuit congestion T42t 42,553 15,735
9 No. of inter-TAX routes with call rejections 10 NIL
(FRD) even though traffic was very less
10 No. of total STP messages lost due to threshold 428 NIL
11 No. of PUPEs going to frequent ‘load 4 1 (See Note)
modification’ (in SRFIL)
Note: These can be completely avoided by further fine-tuning after taking the observations
once again during the hour when ‘load modifications’ are observed, since busy-hour for Cell
routes are different from busy-hour for Landline routes.
Some of the problems and ‘fine-tuning’ details are as follows:
Sl.No. Problems/Observations Solutions/Suggestions
1 Some PUPEs were going to ‘load modification’ Large size NFSCs were split
frequently: into different NFSCs with not
a) Some NFSCs were large size with more than 1500 more than 500 CICs in each
CICs. When the traffic in that NFSC was high with NFSC. Some sublevels in that
call rejections, PUPE handling that NFSC (in direction were given direct
FLUIL) was overloaded. routing to each NFSC with
b) One large size NFSC was split into small groups but overflow routing (ACD) to
traffic was not distributed. NFSC1 was first choice last NFSC. Direct traffic to
for all levels in that route. Overflow of NFSC1 was last NFSC was limited to
routed to NFSC2, overflow of NFSC2 was routed to about 50% of CICs during
NFSC3, etc. By this arrangement, number of non-peak hours. There was no
seizures of PUPEs for selecting a free CIC had ACD for the last NFSC. Thus
increased more than double the actual call attempts. 50% of CICs in the last NFSC
One large size NFSC was split into small groups and could cater for the overflow
c) Traffic was distributed between different NFSCs of traffic of other NFSCs.
same route. But, number of CICs in each NFSC was Traffic and number of CICs in
not matched with the actual traffic thro’ that NFSC. all other NFSCs were
Page 1 of 3
When there were free CICs in some NFSCs, calls readjusted such that there was
were rejected in some other NFSC for want of free no overflow during normal
CICs. traffic periods.
d) SCSLA: In many NFSMs with 2 COCs and Law=2, For NFSC with Law 3, with
both the COCs were not sharing the load equally. NFSM law as 3, both the
One COC was handling about 80% traffic whereas COCs were sharing the traffic
second COC was handling only 20% of traffic. better.
e) FSMIN: In some NFSMs, both COCs were created One COC was suppressed and
in same TSV. recreated in different TSV
corresponding to the PUPE
f) COCs were not created in the last PUPE. which was having minimum
number of COCs.
g) OCIN: CTO=TTRA-j+TREF-j: Readings were COCs with STP messages
taken for all PUPEs on 2 consecutive days. (e.g. IN, Cell, etc) were
Difference between the 2 readings of TTRA-j gave distributed in all PUPEs so
the number of STP messages handled by PUPE that difference of TREF-j was
corresponding to TSM-j. Difference between the 2 0 for all PUPEs. (Loss of STP
readings of TREF-j gave number of STP messages messages may result in failure
lost by PUPE TSM-j, due to ‘STP threshold’ (total of some IN calls as well as
threshold for the exchange is 1000). Four PUPEs not-charging of some ACC &
were found handling heavy STP messages whereas VCC calls.)
two PUPEs were not handling any STP message.
Some STP messages were lost in two PUPEs.
h) For all I/C calls from one Cell operator, ‘Confusion’ DFA4 was deleted from BW
C7 message (Header 2F) was received as response to NFSCs of private operators
‘Charge’ message sent from TAX. This had since they didn’t need
increased number of C7 messages. ‘charge’ information from
j) Number of SAM messages were more for all O/G In the ACHs of all C7 routes
calls to dependant C7 exchanges and for Cell calls. which did not have further
RCR in the ACHs were kept uniformly 8 for both transit on R2, RCR were
R2 and C7 routes. made equal to RDC i.e. 11 for
01 to 09 codes, 10 for 92/93/
94/98, 12 for 95 codes, etc.
k) OCAGE during busy hour for all SMAs with After fine-tuning as explained
PUPEs showed processor load (ML=PUPE, in paras 1 a) to j), OCAGE
TYCO=N, SFCA=PUP, TOML= %) was more than was taken once again. Some
95% for some PUPEs whereas it was less than 70% NFSCs with more call
for some other PUPEs i.e. load on PUPEs were not attempts (FPD+FRD) were
balanced. PUPEs with higher loads were going to shifted from heavily loaded
‘load modification’ frequently. PUPEs to less loaded PUPEs
2 FRD even with free CICs: In some inter-TAX With DIS=FITX in all inter-
routes, call rejections FRD were observed even TAX ACHs, ‘Charge’
though traffic were less and free CICs were message received from distant
available. In SARLA observation it was found that exchange (due to using wrong
‘Charge’ message was received for some calls, from ‘charge rate numbers’ in
distant end. For calls from dependant stations to CDOT) were ignored and
such routes, call was routed thro’ alternate route if calls were successful.
ACD was available. If ACD was not available, call CTA=PX was given in all
Page 2 of 3
was failing. NFSCs where transmission of
‘Charge’ message was not
3 PCD parameter was used in some BW NFSCs When PCD was deleted, FRD
assuming that PCD would reserve some percentage in such routes were not
of CICs exclusively for O/G purpose. (But, PCD is observed. PCD is not
to be used only when certain % of CICs are to be recommended in normal
reserved for I/C calls. These CICs will not be cases.
available for O/G calls, even if free.)
4 ASR of I/C R2 NFSC from CDOT Group Dialing Reducing the ‘Digit Seize
Centre was less. OFCLA showed some partial Forward’ to 6 for R2MF in
dialing calls with only 4 or 5 digits even though CDOT/E10B was improving
‘Digit Seize Forward’ was kept as 10 at CDOT. To ASR. Whereas, increasing
receive 12 digits of called number and 10 digits of ‘Digit Seize Forward’ to 9/10
CLI on R2MF signaling, it took about 7 seconds. for C7 Tk.Gp. was increasing
CDOT RAX parented to MBM for transit working ASR.
was timing-out and released the call while digits
were being received by TAX. Also, some
subscribers were disconnecting the call prematurely
when there was ‘no-tone after dialing’.
5 CICs in DFAU : Some CICs to Hutch (AXE It was informed that such
Switch) were permanently in DFAU. Circuit testing CICs were remaining in ‘seal’
by CIRDEM was showing ‘Busy’. CICs were not status at AXE end. When all
coming to normal status by blocking & unblocking. COCs in the route were
By withdrawing and adding to another NFSC, or by blocked simultaneously and
disconnecting the PCM and then restoring, circuits then unblocked at AXE end,
were coming to normal status. But, when first O/G all CICs had come to normal
call seized the circuit, no response was received status.
from distant end and the CIC was going to DFAU
once again. Unblocking messages were sent
periodically by system every one minute.
6 Spare codes were received from some Level-2 Deeper analysis upto working
TAXs due to analyzing only upto 2/3 digits for inter- SDCA codes was done at L-2
circle STD codes of above 500 KMs. TAXs.
7 OFAML messages were output frequently from By putting the PUPE in INDL
some PUPEs to shift to reserve PUPE,
8 To reduce load on different equipments Increased RCA in preanalysis.
Page 3 of 3