Cellular Optimization Community

Uplink Interference Identification

boggie — Fri, 05 Jun 2009 02:21:00 +0000

Frequency clearance is mandatory requirement in designing, implementing and optimizing radio frequency for cellular networks. Though a good frequency plan has been defined to ensure acceptable of internal interference level, an external interference sometime disturbs the radio networks not only in the operational phase but also when the mobile operator optimizing their radio networks.

Continuous radio performance quality monitoring activities should cover major KPIs as well as interference level in the network. Identification on interference level can be performed through BSS performance statistic such as radio quality issues relate to handover activities and signaling and traffic channel drops analysis.

There are many reasons causing radio quality problem such as defective RF module on the BTS, non optimized coverage design, bad frequency planning, external interference, etc.

Unluckily, not all of root causes of the radio quality problems can be easily identified by the RF engineers, especially when the problem occurred due to external interference on the uplink direction. At the end, a field measurement test has to be conducted, e.g. using spectrum analyzer.

To ease the field measurement test in identifying interference sources in office interference level analysis shall be initialized using dedicated mode or idle mode performance data. Measurement Reports send by MS is common statistic data used by radio networks during dedicated mode, and specific vendor RF feature like Idle Interference Measurement is available as well for idle interference analysis.

Through geographical information system application available in market such as MapInfo, those RF measurement statistic data can be easily manipulated using data interpolation thus become useful information for RF engineer to conduct external interference sources identification.

Hence, corrective actions shall be taken placed in the form of frequency tuning or external frequency clearance.

Fractional Load Planning

boggie — Wed, 13 May 2009 16:29:43 +0000

Fractional Load Planning is one of specific radio design task that evaluating and analyzing available frequency spectrum and transceiver against traffic load, thus able to determine design of radio frequency assignment strategy to maintain quality service at desired level.

At early stage of activity, gathering performance measurement data and network configuration is required to allow evaluation of current relation between network load and performance.

The evaluation should lead to frequency spectrum efficiency preliminary analysis which deliver initial frequency allocation and hopping scheme. Using advanced planning tool, the initial RF strategy can be simulated and evaluated through intensive coverage and interference analysis.

The simulation would let us know whether further adjustment would be required or not, thus most suitable and benefit strategy can be determine. Once we able to find it, detail RF configuration data for new hopping system shall be defined.

On implementation of the new RF strategy derived from fractional load planning, fine tuning might be required, including consideration of activation and fine-tuning of interference controlling feature such as Power Control and DTX.

Fishbone Diagram in Network Optimization

admin — Mon, 27 Apr 2009 15:22:42 +0000

Dr. Kaoru Ishikawa, a Japanese quality control statistician, invented the fishbone diagram. Therefore, it may be referred to as the Ishikawa diagram. The fishbone diagram is an analysis tool that provides a systematic way of looking at effects and the causes that create or contribute to those effects. Because of the function of the fishbone diagram, it may be referred to as a cause-and-effect diagram. The design of the diagram looks much like the skeleton of a fish. Therefore, it is often referred to as the fishbone diagram.

Network optimization is commonly close related to root-cause analysis. Lets take for example identifying cause of dropped call that triggered by low signal strength.

Fishbone diagram is suggested to group all possible problem causes into common major categories which labeled on the ‘bone’ of the ‘fish’, e.g. 4M (methods, machines, materials, manpower); 4P (place, procedure, people, policies); 4S (surroundings, suppliers, systems, skills).

Refer to those categories, we might define all categories that more related to cellular environment to be specifically dropped call. An fishbone for dropped call analysis can be figured as it has categories of surroundings, antenna/feeder, hardware module, mobile subscriber, and software system.

An extensive analysis base on drive test and on site verification, alarm data inspection, configuration and radio parameter assessment, and network performance data analysis have to be performed to see overall causes relation and judge ‘most likely cause’.

At this point, we will possible to define root cause base on its priority on which we are going to validate through step by step proofing and finally found and decide the real root cause.

Ping Pong Handover

admin — Mon, 20 Apr 2009 17:46:39 +0000

As the cellular network growing, it requires addition of new sites to expand capacity and or coverage services. New mobile operators were expecting to grow their network in exponential order to gain market share against mature competitors. Without any intensive network performance maintenance, common problem such ping pong handover would be degrade user experience like bad voice quality ( SQI speech quality index, MOS mean opinion score ) or even dropped call.

Ping Pong handover is shown from the successful handover back to old cell within pre-defined time of total handover, e.g. less than 10 seconds. Since not all BSS vendors provide such performance counters it might be identified by simple metric that shall be expressed as total successful handover over number of call or connection, e.g. more than 200% indicates ping pong handover.

The actual ping pong handover is easily identified through drive test with sense of optimization engineer without require any calculation. Coverage holes might lead to Ping-Pong handover especially for slow moving mobiles. It may be cause by shadowing by high building. Several possible equal signals from two or more base stations might lead to Ping-Pong handover as well which is so called not optimized cell dominance.

The following action items shall be used for optimization solutions:

Check the parameter setting such as hysteresis, offset, priority layer in dual-band case, etc. If poor parameters setting found, then correct the related parameters. Compare to default parameters design.
Check the output power BSPWR and BSPWRT which normally put on maximum value or compare to design.
Check path balance on transceiver link, e.g. loss in uplink and downlink chain. Defective RF modules might be the reason of imbalanced link.
Check the site location such as possible coverage holes or no dominant cells.
Perform drive test on that particular area to check the signal strength, if the average of signal strength of the neighbors are the same, no dominant cell is found.
For Long term action if no dominant cell shall be adding new site, while short term action shall be uptilt the most dominant cell as the serving cell for that area and reduce coverage for others.

Maximizing Radio Resource Utilization

admin — Mon, 20 Apr 2009 17:35:21 +0000

Have you ever found many of cells with very low traffic utilization when assessing system load? What about amount of cells with very high traffic utilization? It is often occurred both on new mobile operator or the mature one, whilst network utilization already in line with the design target.

Lack of accuracy of the design inputs or assumptions (i.e. traffic model) in network design phase for new network deployment, fast increment or degradation amount of subscribers on particular provider due to end of promotion or other reason causing churn, global economic changing, and soon, it can be many reasons trigger non-optimized resource utilization. These all mean missing direction in addressing cost of investment for infrastructure (e.g. CAPEX), in contrary mobile operator able to allocate them for subscriber acquisition cost which is remain high for 2G network.

As the network owner or managed service provider, you can start to intensively evaluate whether resource utilization of your network have been optimized or not yet by simply go to cell level load assessment rather than rely on network level utilization. The following techniques are recommended to maximizing radio resource utilization:

Market Research

Through market research able to provide update data of potential subscriber demand which increase accurateness of traffic model or other inputs to size the network capacity as required. You can’t just relay on the ongoing rollout process base on non up to date market data. Do your market research in more detail with help of specialized third parties in market survey.

Radio Resource Management Features

How efficient RRM features within the base station sub-system have been implemented? You check whether very useful radio resource management features have been implemented or not. Consider to activate network feature such queuing, preemption, directed retry, traffic handover, PS and CS dynamic resource allocation, etc. Define its efficiency metric for each function to keep them active or need to be stopped.

Capacity Enhancement

Quick capacity enhancement through new channel coding such half rate or dual rate can be a tradeoff between quality and capacity. However this way shall be used as temporary solution. Considering to deploying AMR will able to compensate loss of quality while providing higher capacity when it is combined with half rate coding. Reconfiguration of traffic channel or signaling channel can also be performed with less capacity gain. In medium to long term period, utilization issue might be solved through introduction of new cell or site, or simply TRX expansion.

Load Balancing

Shifting traffic from one cell to other cell known as load balancing can be done through many ways such: idle and dedicated radio parameters tuning in single or dual band network, coverage tuning with antenna setting adjustment, traffic sharing through handover due to traffic, etc. These are major activities during tuning activity to relief capacity issues such high blocking, very low or high utilization.

Capacity Balancing

Identification of cells that having condition of very low and high traffic utilization is direct action to resolve non-optimized radio resource utilization cases. Term of donor and receptor are applicable for those identified cells. The identification has to consider long term traffic trend and increment factor of traffic during normal session (average daily) and peak session (Lebaran, Christmas, New Year, etc.). TRX downgrade and upgrade on the particular cells is main activity of capacity balancing. Channel type reconfiguration can also be done with less capacity gain.

Handover Failure due to same BCCH-BSIC

admin — Sun, 19 Apr 2009 06:21:31 +0000

Limited available radio frequency ARFCN owned by mobile operators would often increase challenge of RF planning and optimization. Beside difficulty when assigning frequency, one of impact of this is high handover failure due to configuring same BCCH and BSIC on the neighbor.

We can simply identify such impact by make a special query that meet such condition:

“High handover failure in one direction AND 1^st tier + 2^nd tier neighbor have same BCCH-BSIC”

On this situation, mobile station would be “fooled” during composing measurement report to observed cell and handover shall be wrong be triggered and possibly increases handover failure.

Wireless network optimization can be started with network configuration audit which cover inspection of high handover failure rate in one direction with above mentioned situation is met.

Dropped Call due to Sudden Drop

admin — Thu, 16 Apr 2009 18:06:54 +0000

On circuit switch service, when a call is abnormally disconnected, a Clear Message with cause code Call Control be treated as normal Disconnection is sent to the MSC – named Clear Request Message.

Refer to Ericsson system, the following Urgency condition is checked at that time and the relevant counter is incremented as a consequence:

Excessive TA
Low Signal Strength
Bad Quality
Sudden Drop

As named, stright forward meaning for the dropped call is described for the first three items. However, Sudden Drop is quite not easy to understand.

Sudden Loss are drops that have not been registered as bad quality, signal strength, timing advance. The term Sudden Loss is used because if the network cannot establish a connection with the lost MS after a pre-defined period, the sudden loss counter is incremented if the last reported measurement from the MS does not fulfill any of the reasons mentioned.

A connection is marked as Sudden loss if none of the three types of urgency states (that is excessive TA, low signal strength or bad quality) are indicated and the locating procedure indicates missing measurement results from the MS.

Drops due to ‘Other’ reasons are generally associated with hardware problems and disturbances, number of drops due to ‘Other’ reasons is obtained by subtracting the drops with known reasons from the total number of drops.

Main contributors in sudden and other TCH drop:

Very sudden and severe drop in signal strength, such as when subscribers enter into buildings, elevators, parking garages, etc.
Very sudden and severe occurrence of interference or bad quality.
MS running out of battery during conversation.
Handover Lost.
BTS HW faults.
Synchronization or Abis link fault (transmission faults).
MS Faults.

Pilot Polution Identification

admin — Thu, 16 Apr 2009 16:40:02 +0000

In idle or cell_FACH mode, phenomenon of the pilot pollution is that a UE cannot firmly camp on a cell at one location because of receiving many pilot channels with similar quality (or signal strength), i.e., Ec/No (or RSCP).

The condition to determine the area has pilot pollution in idle or cell_FACH mode is that third pilot appears in the cell re-selection region.

In cell_DCH mode, phenomenon of the pilot pollution is that a UE at one location frequently changes its active set cells (active set update rate is very high) because of receiving many pilot channels with similar quality (or signal strength), i.e., Ec/No (or RSCP). It causes high signaling load in RRC and Iub interfaces and the capacity of RNC is consequently reduced.

The condition to determine the area has pilot pollution in cell_DCH mode is the number of pilots within the Reporting Range 1b: threshold for drop window range is larger than Max Active Set.

Possible solutions to overcome pilot pollution shall be:

The direct solution is to remove the cells overlapping by changing the antenna configurations or reducing pilot powers of the unwanted cells.
The most effective solution is to increase the pilot channel power Primary CPICH power of the desired cell.
For the pilot pollution in idle or cell_FACH mode, Q Offset 2sn (CPIC RSCP) parameter of the desired cell can be increased to create a dominated cell.

Imbalance Traffic Load

admin — Tue, 14 Apr 2009 20:51:06 +0000

One of DCS1800 site implementation objective is to relief TCH Blocking Rate on existing GSM900 sites by using its achieved capacity gain. However on the initial capacity design such as defining TRX configuration for the new DCS, the radio planners sometime missed to carefully take into account important factors contributing traffic absorption on the DCS cells, e.g. geographical subscriber distribution, coverage estimation, etc.

Facts that currently faced by dual-band mobile operators might be imbalanced traffic load on co-located dual-band cells. Generic technical tips to identify Imbalanced Traffic Load at co-located dual-band cells (e.g. GSM and DCS cell on the same sector) shall be suggested into two kinds of situations on which they are triggered:

Due to TCH Utilization: a particular co-located & “co-direction” dual-band pairing cells having 200% difference of utilization.
Due to TCH Blocking Rate: a particular co-located & “co-direction” dual-band cell with TCH Blocking Rate less than 0.005% (e.g. ~ 0%) while the other cell with more than 2% of traffic channel blocking rate.

Coverage tuning, TRX re-balancing, and radio parameter tuning are widely being used to optimize traffic sharing between those co-located dual-band cells.

Does your radio network require frequency re-plan now?

admin — Tue, 14 Apr 2009 17:04:19 +0000

A very tight competition in mobile telecommunication industry nowdays is shown in which new players enter the market almost every year and a tariff war is imminent, quality of service and efficient, lean operating structure have become the antes of the game.

Telco operators are forced to maximize the use of their existing network resources while maintaining good grade of service for subscribers in order not to loose them to competitors. In this kind of business climate, bad quality could mean disaster.

Radio interface capacity and quality become one of major issue to accommodate fast subscribers growth while maintaining superior quality of experience. Assignment of radio frequency strategy has to be able to carry out the needs of this situation including when it has to be redesigned. The following considerations shall be taken to evaluate whether your RF network requires such re-plan or not:

Spectral Efficiency

Spectral efficiency quantifies the amount of traffic a network can carry for given spectrum. Even though the most traditional way to express it has been Erlangs per megahertz per square kilometer (Erl/MHz/km2), there have been many other ways to quantify the same concept in the wireless industry, and one of most easy can be used to direct measure of how loaded the spectrum is Effective Frequency Load (EFL).

EFL quantifies how loaded each frequency available in the system is, and can be calculated as:

EFL (%) = (Erl_BH/Tot#freq) * [1/Ave# (TSL/TRX)]

The higher EFL value on the network, the more probability of system performance and service performance will be degraded since it directly indicates collision probability on the air interface. It is easy to make cross relation between this and mentioned indices such Dropped Call Rate or FER based on actual performance statistic.

When this metric considered on the higher level which impacts to performance degradation, a new frequency allocation and hopping scheme strategy should be choosen on the radio engineering level togather with BCCH allocation startegy.

System Performance

Most installed radio sub-system from some vendors has a capability to provide actual radio measurement report through their BSS performance management. One of most useful radio quality metric is RXQUAL sample on UL and DL of air interface.

Grade of service quality being used by most of telecommunication industries in the world is to deliver highest probability value on which their system able to serve subscribers within guaranteed best quality.

Good RXQUAL Sample = RXQUAL (0 to 5) / RXQUAL (0 to 7) > 98%

A generic value applicable for this is up to 98% probability of accumulative radio quality successfully being performed within RXQUAL 0 to 5.

Distribution of handover attempt due to DL and UL quality over the whole handover type can also be used as indicator of air interface quality. Generic value of this metrix might be set around 20% to 25% to indicate bad radio interface quality.

An external factor of interference sources which affect on a large of geographical area might also require modification of allocated radio frequency.

End -user Service Performance

Performance metrics that shall be used under this consideration are easily derived from end to end test throughout drive test such as speech quality index (SQI) or mean opinion score (MOS), setup failure and dropped connection rate, etc. Customers complain collected from the customer services line will be helpful to identify lack of service performance.

Finding on service performance degradation indicates need of radio optimization which is actualized through radio frequency redesign. It should be combined to current system performance to justify the radio reenginering activity.

Capacity Growth Demand

A fast susbcriber growth requires more capacity through network expansion or site capacity upgrade to maximize the grade of service accessibility. The requirement on radio capacity expansion often meets a bottleneck in the situation of current RF scheme on the network limitting addition of TRXs and triger changing of existing RF design.

Redesigning frequency hopping system into higher efficient scheme such SFH 1×1 sometime can be choosen to accommodate more TRXs can be installed into the current system.

Network Rollout Challenges

Nowdays, complexity of network infrastructure become local regulator concern that oftenly leads to failure in executing site acquisition of network rollout. When such cases are found many and halt a need of new site for capacity expansion, it may trigger temporary solution to upgrade radio capacity on existing cells.

More over a fast rollout program sometime meets dfficulties to put new sites on service due to limittation of frequency allocation while new capacity is urgently required.

These challanges require solution on designing new RF strategy or just simply local frequency tuning. Sometime above mentioned factors mixup on the same period that forces mobile operators to perform global frequency reengineering.