Optimisation process guide

RADIO ENGINEERING SOLUTIONS

AIRCOM
International
Technology Consulting
System Optimisation
Guide
CS-SG-TPP-0042-01

CONTENTS
1.1.1 Optimisation Tips: ................................................................................................. 7
1.2 LOW SIGNAL LEVEL ....................................................................................................... 8
1.2.1 Optimisation tips: .................................................................................................. 8
1.3 ADJACENT AND CO CHANNEL INTERFERENCE ................................................................ 8
1.3.1 Optimisation tips: .................................................................................................. 8
1.4 EXTRANEOUS INTERFERENCE ......................................................................................... 9
1.5 LINK IMBALANCE ........................................................................................................... 9
1.6 HIGH HANDOVER FAILURE RATE:................................................................................. 10
1.7 HIGH NEIGHBOUR INTERFERENCE ................................................................................ 10
1.7.1 Optimisation tips.................................................................................................. 10
1.8 NO DOMINANT SERVER ................................................................................................ 10
1.9 DATABASE PARAMETERS .............................................................................................. 10
1.9.1 Optimisation Tips ................................................................................................ 11
1.10 SDCCH BLOCKING ...................................................................................................... 12
1.11 HIGH PAGING LOAD ...................................................................................................... 12
1.11.1 Optimisation Tips: ............................................................................................... 12
1.12 INCORRECT OR INAPPROPRIATE TIMER VALUES ............................................................ 12
1.12.1 Optimisation Tips: ............................................................................................... 12
1.13 TCH BLOCKING: .......................................................................................................... 13
1.14 HANDOVER MARGIN .................................................................................................... 13
1.15 CELLS TOO LARGE ........................................................................................................ 13
1.16 IMPROPER TIMER .......................................................................................................... 13
1.16.1 Optimisation Tips ................................................................................................ 14
1.17 SDCCH MEAN HOLDING TIME ...................................................................................... 14
1.18 ERLANGS, TOTAL CALLS .............................................................................................. 14
1.19 PROCESSOR LOAD ........................................................................................................ 14

CS-SG-TPP-0042-01 Page 1

1 INTRODUCTION
This document describes in the process of optimisation in GSM900/1800 mobile
telephone systems. The document covers the traditional methods of optimising a
cellular system in the basic optimisation section. It introduces the improvements
being made to Aircom’s Optimisation process with the introduction of new tools and
techniques in the advanced optimisation section. Aircom has a tool namely “ optima ”
to monitor the system performance. A series of counters are monitored on daily,
weekly and monthly basis to check the network health. The document will specify the
recommended metrices and the standard performance management procedures to
identify and rectify the problems.

Optimisation is an invaluable element of service required to maintain and improve the
quality and capacity of a network. It is essential if an operator wants to implement
changes to the network to maintain the high quality of service levels expected by
subscribers in GSM900/1800 networks. Without optimisation the network will
degrade from the commissioned state, due to the network changing radically as the
traffic on the GSM system grows, and snapshot optimisation will not keep pace with
these changes. Without optimisation the system will suffer poor call quality, many
dropped calls due to interference and inaccurate parameters resulting in poor
handover performance. These together with other problems, have the same result,
Subscriber Dissatisfaction.

Setting the parameters that control mobility has equal importance to the frequency
plan. In GSM900/1800 networks there is a series of parameters that control mobility.
Tuning these parameters for improved GSM900/1800 operations, in terms of
maximising calls carried, improved handover performance and increased call
success rate, is termed ‘Optimisation’.

The aim of optimisation is to maximise the Quality of Service (QoS) of the GSM
network. In order to do this you need to measure the QoS, compare the measured
value with the desired value, and then take steps to correct the causes of any
deviations from the desired value.


INPUTS PROCESS OUTPUTS

RF DESIGN
PROCESS

QUALITY OF INTERGRATION OF
SERVICE METRICS SYSTEM OR NEW FREQ.
PLAN

DATABASE
PARAMETERS OPTIMISATION
OPTMISATION REPORT

RF DESIGN
PARAMETERS
OMC

DRIVE TEST
ROUTES

APPROX 2 - 4 WEEK PROCESS

PERFORMANCE PERFORMANCE
ENGINEERING REPORT

SYSTEM
ACCEPTANCE
FIGURE 2.1 Basic Optimisation Process

The flowchart shown in figure 2.1 summarises the basic optimisation
process.


2 D AILY C OUNTERS :
The following metrics can be used to measure the performance of the network.
These counters should be monitored daily on a per cell basis. These counters are
defined in Optima under the heading of “ network level ”.

S.NO Daily Counters to Monitor Standard Thresholds
1 Call set up success Rate Above 90 %
2 Tch Congestion Less than 2%
3 Sdcch Congestion Less than 2 %
4 Handover Success rate Above 85 %
5 TCH Drop Rate Less than 1 %
6 TCH blocking Less than 1 %
7 SDCCH blocking Less than 1 %
8 Erlangs per hour -
9 Average no of TCH’s -
available
10 Sdcch Mean Holding Time 3 seconds
11 Total Calls -
12 Processor Load -

2.1 Call Setup Success Rate:
The call setup rate should be above 90 % for a healthy network. However a CCSR of
85%-90% is satisfactory. There could be so many reasons for a poor CSSR. Some
are described as follows.

No Access to Sdcch
CM service reject
Tch Failure assignment
Hardware Problems

2.1.1 No Access to SDCCH:

BSS detects Channel Request (in the form of RACH) from a source, requesting
resources for network transactions. After validation of the RACH, BSS will attempt to
allocate a dedicated channel (SDCCH) for the source. Once the availability of
SDCCH channel is confirmed, the BSS will send Immediate Assignment to MS
indicating the dedicated SDCCH sub-channel (via AGCH), whereby subsequent
message exchange will be performed over the dedicated SDCCH.


2.1.2 Examples of Abnormality :

2.1.2.1 Valid RACH (SDCCH Congestion)

Due to the unavailability of SDCCH, BSS will response to MS with Immediate
Assignment Reject, terminating the transactions. In which case, call setup is termed
as unsuccessful due to SDCCH congestion

Invalid RACH (Invalid Establishment Cause detected in the received RACH )

2.1.2.2 Phantom RACHs

The received RACH is in fact generated from an “unknown source”, whereby it fails
to continue the transactions after SDCCH has been allocated by the BSS. For
instances, cases of Channel Request detected by over-shooting cells, Handover
Access burst from distanced MS, hardware deficiency, UL/DL imbalance path, MS
moving out of range would carry the Phantom RACHs symptoms.

2.1.2.2.1 Optimisation tips:

Within the optima there are certain stats which can be monitored before coming to
the conclusion that there is SDCCH problem

1-) SDCCH Blocking
2-) SDCCH Congestion

If the SDCCH blocking is greater than 1 % or the SDCCH congestion is greater than
2% than that means that it is a capacity related issue and more slots should be
assigned for SDCCH.

A Tch can be allocated bypassing Sdcch. A parameter namely
Immediate_assign_mode when enabled allocates Tch bypassing SDCCH.

2.1.3 CM Service Reject:

CM Service Request [MOC] or Paging Response [MTC] to BSS/MSC. Inside the CM
Service Request message (MS initiated service request), MS informs the network the
types of services it requires (i.e. Mobile Origination Call, Emergency Call, Short
Message Transfer or Supplementary Services Activity), whereby Paging Response is
specific to MTC. Subsequently, BSS embraces the information with its own initiated
Connection Request BSSMAP Message, sends to MSC for approval. MSC will
response with either Connection Confirmed, confirming the success in link
establishment between MS-BSS-MSC, or Connection Refused, indicating the
termination of the specific network transaction.

2.1.3.1 Examples of Abnormality
Basically, the generation of excessive Connection Refused messages could relate to
MSC internal issue.


Upon successful link establishment, a number of signalling activities will proceed.
Firstly, Authentication Request will be sent to MS followed by cipher mode command.
Thirdly, MSC will send Identity Request to MS, questioning the IMSI. MS should
response with its IMSI in the Identity Response message to MSC for validation. MS-
BSS-MSC connection will be cleared down if any of the above three signalling
activities fail. In general, Authentication Request, Cipher Mode Command or Identity
Request does not always appear during call setup, the occurrences of the three
signalling activities is controlled by MSC. (e.g. Authentication Request is only
performed on every 10th call in the network)
.
Examples of Abnormality
MS related (e.g. incompatibility in ciphering algorithm, identity, etc)

2.1.4 TCH Failure Assignment:

Upon completion of MS/BSS/MSC link establishment, MSC issues Assignment
Request to BSS, requesting TCH assignment to the dedicated MS. Subsequently,
BSS will attempt to allocate free TCH for MS voice messaging. Once Assignment
Command is received by MS, stating the availability of TCH for the MS, it will move to
the dedicated TCH and responds with Assignment Complete. In turns, BSS will
submit Assignment Complete to MSC as to complete the signalling activity.

2.1.4.1 Examples of Abnormality
- TCH Congestion
BSS responses to MSC via Assignment Failure, indicating the unavailability of TCH
resources. BSS will subsequently terminate all transactions, in which case, call setup
is termed as unsuccessful due to TCH congestion.
Interference on TCH
There could be co-channel or adjacent channel interference on a particular TCH
causing TCH failure assignment.

2.1.4.1.1 Optimisation tips:

For TCH congestion certain features can be enabled like TCH queuing, Directed
Retry and Congestion Relief.In the case of TCH Queuing feature is enabled, MS will
queue in the original SDCCH, awaiting for the next available TCH. It is to be
reminded that once Queuing Timer expires, BSS will also terminates all transactions,
in which case, call setup is termed as unsuccessful due to TCH congestion. The
same situation also applies in situation where Congestion Relief feature is enabled.
In the case of Directed Retry feature is enabled, MS will perform handover to TCH of
another cell if a valid handover neighbour is detected. The best thing to do is to add
more radios in the cell to remove congestion.
Interference analysis on a particular carrier can be done through an optimisation tool
like Neptune. Once interfering frequencies are determined, the frequency plan can
be cleaned from such frequencies.


2.1.5 Hardware Problems:

Hardware failures also play the major role for Poor CSSR. Improper functionality of
any BTS hardware can affect the overall performance of site

2.1.5.1.1 Optimisation Tips:

If there are no capacity or RF issues then equipment needs to be checked. Before
starting the drive test make sure that the cell sites are free of any hardware alarms.
The important parameter to check is the Path balance. If Path balances are not fine
then start checking the power from radio to connected antennas. If we take the
example of GSM 900 scenario, the link budget defines that the radio should transmit
40 watts power and at the top of the cabinet, 20 watts are received (considering the 3
dB loss of combiner). while checking the power, if any components seems to
produce more losses than expected, change that component. Similarly check the
power at antenna feeder ports. Some time due to the water ingress, connectors get
rusty and needs to be replaced.

2.1.6 High Dropped Call Rate:

For a healthy network the drop call rate should be less than 1 %. There are again
number of reasons, which could contribute towards higher dropped call rate.

a-) Drop on Handover
b-) low signal level
c-) adjacent channel interference
d-) Co-channel interference
e-) Extraneous interference
f-) link imbalance
a-) Drop on Handover

The call may drop on handover. It’s mostly high neighbor interference on the target cell, which causes
the main problem. Sometime the mobile is on the wrong source cell (not planned for that area but
serves due to the antenna overshoot) which may result in the call drop.

2.1.7 Optimisation Tips:

Within optima, monitor the following statistics. Theses statistics are defined under
the category of BSC level statistics.

1-) Total and successful handovers on up/dl quality.
2-) Total and successful handovers on up/dl signal strength.
3-) Total and successful power budget handovers.

From the above statistics , quality or level issues can be estimated.


2.2 Low Signal Level

Signal level below –95 dbm is considered to be poor. If the mobile is unable to
handoff to a better cell on level basis, the call will possibly be dropped. Topology or
Morphology issues may also be there like if Mobile enters into a tunnel or a building,
higher RF losses will be developed.

2.2.1 Optimisation tips:

First of all Path balances should be checked. If Path balances are deviating form the
standard value than Check the BTS transmitted power with the help of wattmeter.
BTS may transmit low power because of the malfunctioning of radio or higher
combiner losses. Also check the feeder losses, antenna connectors. Enable downlink
Power Control. Power control is bi directional. The lower and upper receive level
downlink power control values should be properly defined.

1-) l_ rxlev_ dl_p

defines the lower value for receive level for the power control to be triggered .

Range = 0 to 63
Where 0 = -110 dbm
1 = -109 dbm
63 = -47 dbm

e.g if the value of 20 is set it means that the BTS will start transmitting more if it
senses that downlink receive level is below –90 dbm.

2-) u_rxlev_dl_p

defines the upper threshold value for receive level for the power control to be
triggered. (Range is same as described above)

e.g on setting the value of 50 ( equivalent to –60 dbm ) BTS will lower down the
power .

2.3 Adjacent and Co Channel Interference

Frequency planning plays a major role to combat adjacent channel and Co channel
interference. Co channel is observed mostly when mobile is elevated and receives
signals from cell far away but using the same frequencies.

2.3.1 Optimisation tips:

An optimisation tool like Neptune could be helpful in identifying the interference on a
particular area. Such frequencies can be cleaned form existing frequency plan. The


following statistics can also be monitored to confirm that there are interference issues
in the cell. These stats are defined in optima under the category of BSC stats.

1-) TCH interference at level 1

When a tch timeslot is idle it is constantly monitored for an uplink ambient noise.
During a SACCH multiframe an idle time slot is monitored 104 times. These samples
are then processed to produce a noise level average per 480 ms. An interference
band is allocated to an idle slot depending upon the interference level. The
thresholds for these levels can be set in the system parameters, interference level 1
being the least ambient and interference level 4 being the most ambient. While
planning the network care should be taken that the cells do have the proper
frequency spacing.

2.4 Extraneous Interference

Extraneous interference might be from

Other mobile networks
Military communication
Cordless telephones
Illegal radio Communication equipment

Optimisation tips:

External interference is always measured through spectrum analyser which can scan
the whole band. Some Spectrum analyser can even decode voice from AMPS
circuits or Cordless Phones.

2.5 Link Imbalance

Some time the malfunctionality of vendor hardware becomes responsible for high
CDR. One of the possible scenarios could be

Transmit and Receiving antenna facing different direction
Transmit and Receiving antennas with different tilts
Antenna feeders damage, corrosion or water ingress
Physical obstruction


2.6 High Handover Failure rate:

High Handover failures rate will probably be due to one or more of the following
reason.

a-) High Neighbour Interference
b-) No Dominant Server
c-) Database Parameters

2.7 High Neighbour Interference

While handing off to the best neighbour the interference on the target cell frequency
may result in the hand off failure.

2.7.1 Optimisation tips

When designing the cell frequencies care should be taken that there is proper
frequency spacing between the cells to avoid neighbour interference. In most of the
cases Ping pong Handover starts i.e the mobile hand off to a cell for better level and
due to interference (quality issues) hand off again to original cell. A thorough drive
test can determine the “interfering frequencies” which should be eliminated from the
frequency plan.

2.8 No Dominant Server

If cell sites are designed poorly there might be areas where neighbour being received
at the same level and some neighbours randomly look good for handoff for a certain
amount of time. Such situation is disastrous because handoff decision will be hard
and mostly it will end up in unsuccessful handovers.


Antenna tilts provide the good way to reduce the footprint of the sites. Efforts should
be made that a single dominant server should serve the specific area. Timing
advance limitation ( ms_max_range) is applied to cell areas where there is multiple
servers.

2.9 Database parameters


Received level, receive quality and Power budget algorithm are set in the system
information to define the criteria for Handover. Improper values for these criteria may
result in poor handoff.

2.9.1 Optimisation Tips

Enable the “ Per neighbour ” feature which displays the successful and
unsuccessful handovers on a per cell basis. In “optima ” monitor the following stats,
which comes under “ cell statistic category ”.

HOCNT (handover count from source cell to destination cell over reading
period)
HOSUCC (Success Count from source cell to destination cell over reading
period)
HORET (Returned Count from source cell to destination cell over reading
period)
HODROP (Dropped count from source cell to destination cell over reading
period)

All those cells can be identified which are problematic in terms of hand off so one can
focus only specific cells causing the major contribution towards poor HSSR. Ensure
that handover margins are optimised. Rule of thumb is 4 dB for adjacent frequencies
and 6 dB for cell without adjacent frequencies. The following parameters can be
played for defining the thresholds for imperative and non-imperative handovers.

1-) l_rxqual_up_h (defines the lower threshold for uplink quality handover)

Range: 0 to 1800
Step size: 0.01

e.g. a value of 500 defines the lower threshold value of 5( BER ) for a quality
handover to be triggered for uplink. The optimum value for this threshold is 500

2-) l_rxqual_dl_h (defines the lower threshold for downlink quality handover)

3-) l_rxlev_up_h (defines the lower threshold for received level uplink handover)

e.g. a value of 20 defines the threshold value of -90 dbm for a level handover to be
triggered for uplink.

Range: 0 to 63
Where 0= -110 dbm
1= -109 dbm

47= -63 dbm

The optimum value for this threshold is 15 i.e. 95 dbm. If the signal level goes below
that, a level handover is initiated.

4-) l_rxlev_dl_h (defines the lower threshold for received level downlink handover)


5-) u_rxlev_dl_ih (defines the upper threshold for downlink interference hand over)

6-) u_rxlev_dl_ih ( defines the upper threshold for uplink interference Handover )

2.10 SDCCH Blocking

SDCCH blocking is probably due to one or more reasons.

a-) No access to SDCCH
b-) Failure before assignment of TCH
c-) High Paging load
d-) Incorrect or inappropriate timer values

The first two reasons has already been discussed.

2.11 High Paging load

Irregular paging distribution in location areas results in SDCCH blocking. Higher
paging load in certain location area means higher location updates on SDCCH
resulting in SDCCH blocking.


A location area with a high paging load needs to be reduced in size to relieve
SDCCH blocking. A location area with a low paging load needs to be enlarged in size
to reduce the overall number of location areas.

2.12 Incorrect or inappropriate timer values

Timer rr_t3 111 sets the amount of time allowed to delay the deactivation of a traffic
channel (TCH) after the disconnection of the main signalling link.


The suitable value for this timer is 1200 ms (max being 1500 ms). The timer will
cause the BSS to wait before the channel in question is allocated another


connection. A lower value of timer will result in higher capacity since the channel is
held for less time before being released.

2.13 TCH Blocking:

Tch blocking may be due to the following reasons

a-) Handover and Power budget margins
c-) Cells too large
d-) Capacity Limitations ( congestion )
e-) Incorrect or appropriate timer

2.14 Handover Margin

Handover margins should be properly optimised to move the traffic to neighbouring
cell. Strict handover margins can result in lower handovers and ultimately congestion
in cell


6 db handover margin is considered to be an appropriate margin for handover. A
strict handover margin results in the strict criteria for Power budget handovers also.
Setting a lower value of handover margin will initiate ping-pong handovers, which are
not considered good for network health. ( hand over margins have already been
discussed )

2.15 Cells too large

If cells are too large meaning antenna too high or antenna too shallow, it will pull in
out of area traffic again causing congestion in the cell


Consider reducing antenna height to reduce the footprint of the site. Also increase
the antenna tilt ( the max tilt being 12 )

2.16 Improper timer

Timers are important in terms of allocating TCH resources. Timer rr_t3111 sets the
amount of time allowed to delay the deactivation of a traffic channel TCH after the
disconnection of the main signalling link. Parameter link_fail and radio_link_timeout


sets the thresholds for the number of lost SAACH messages before a loss of SAACH
is reported on abis.

2.16.1 Optimisation Tips

Setting the lower value rr_t3111 will increase capacity as the channel is held for
some time before being released. The best value for this timer is 1200 ms. The lower
values for radio link timeout and link_fail will result in early disconnection of Tch’s.
The optimised value for these timers is 3 which means that BSS will wait for 12 sacch
messages before it declares that the link has been broken.

2.17 Sdcch Mean Holding Time

Sdcch performs number of tasks like authentication, ciphering, periodic updates,
location update, IMSI attach/detach, and SMS. Holding time for different services is
different. The average holding time for call set-up should be 2.8 seconds and for
periodic location updating it is 3.6 seconds. An “ increase ” in “SDCCH mean holding
time ”can result in longer call set-ups or may end up in unsuccessful call attempts.

2.18 Erlangs, Total calls

Total calls and Erlangs stats on a per cell basis gives the insight on amount of traffic
a cell can cater. These statistics are useful in terms of future planning of sites. A high
value of Erlangs for a specific cell recommends that new load sharing sites be
designed to cater the increased demand for capacity.

2.19 Processor Load

The statistic provide the distribution of processor load on a per BSC basis. Processor
overloading can affect the overall performance of the cell sites like call set-ups,
handovers and call drops. This stat should be monitored on daily basis.


Optimisation process guide

More Related Content

What's hot (18)

Similar to Optimisation process guide (20)

Recently uploaded (20)

Optimisation process guide