The minimum throughput reduction factor is computed using the input data: minimum required throughput per user defined in the service properties, the average throughput per timeslot deduced from the throughput curves stored in the GPRS/
EDGE configuration properties for each coding scheme, the total number of downlink timeslots defined in the properties of the mobile terminal (See defintion above) and the required availability defined in the service properties.
It is at the stage of calculating the average timeslot capacity per transmitter that Atoll studies each covered pixel for carrier power or carrier-to-interference ratio. According to the measured carrier power or carrier-to-interference ratio, Atoll deduces the maximum throughput available on that pixel through the throughput vs. C or throughput vs. C/I curves of the GPRS/EDGE configuration.
The throughput per timeslot per pixel TPTS, Pixel can be either a function of carrier power C, or carrier power C and the carrier-to-interference ratio C/I, depending on the user-defined traffic analysis RF conditions criteria. Therefore,
TDP
Timeslots
Traffic demand per pixel (kbps) Throughput per pixel (kbps)
TSMax TerminalType = TSDL TerminalType CarriersDL TerminalType
TSMax TerminalType
Or
and
The required availability parameter defines the percentage of pixels within the coverage area of the transmitter that must satisfy the minimum throughput condition. This parameter renders user-manageable flexibility to the throughput requirement constraint.
To calculate the minimum throughput reduction factor for the transmitter, Atoll computes the minimum throughput reduction factor for each pixel using the formula:
Once the minimum reduction factor for each pixel is known, Atoll calculates the global minimum reduction factor that is satisfied by the percentage of covered pixels defined in the required availability. The following example may help in understanding the concept and calculation method.
Example: Let the total number of pixels, covered by a subcell S, be 1050. The reliability level set to 90%. This implies that the required minimum throughput for the given service will be available at 90% of the pixels covered. This, in turn, implies that there will be a certain limit on the reduction factor, i.e. if the actual reduction factor in that subcell becomes less than a minimum required, the service will not be satisfactory.
Atoll computes the minimum reduction factor at each pixel using the formula mentioned above, and outputs the following results:
So for a reliability level of 90%, the corresponding RFmin will be the one provided at least 90% of the pixels covered, i.e.
945 pixels. The corresponding value of the resulting RFmin in this example hence turns out to be 0.9, since this value covers 962 pixels in total. Only 87 of the covered pixels imply an RFmin of 0.98. These will be the pixels that do not provide satisfactory service.
This calculation is performed for each service type available in the subcell coverage area. The final minimum throughput reduction factor is the highest one amongst all calculated for each service separately.
The minimum throughput reduction factor RFmin value is a minimum requirement that must be fulfilled by the network dimensioning process when the Reduction Factor KPI is selected in the dimensioning model.
5.6.2.2.5 Step 5: Served PS Traffic
Atoll calculates the served packet switched traffic using the number of timeslots available to carry the packet switched traffic demand. As the result of the above iterative step, Atoll always finds the best possible answer in terms of number of timeslots required to carry the packet switched traffic demand unless the requirement exceeds the maximum limit on the
RF
minNumber of pixels
0.3 189
0.36 57
0.5 20
0.6 200
0.72 473
0.9 23
0.98 87
Figure 5.12Minimum Throughput Reduction Factor TPTS Pixel =f C
TPTS Pixel =f C TPTS Pixel f C ----i
=
RFmin Pixel TPuser min TPTS Pixel TSTerminal
---=
number of the packet switched traffic timeslots defined in the dimensioning model properties. Hence, there is no packet traffic overflow unless the packet switched traffic demand requires more TRXs than the maximum allowed
5.6.2.2.6 Step 6: Total Traffic Load
This step calculates the final result of the dimensioning process, i.e. the total traffic load. The total traffic load L is calculated as:
Where,
• STC is the served circuit switched traffic
• STP is the served packet switched traffic
• TSC, dedicated is the number of dedicated circuit switched timeslots
• TSP, dedicated is the number of dedicated packet switched timeslots
• TSS is the number of shared timeslots
5.7 Key Performance Indicators Calculation
This feature calculates the current values for all circuit switched and packet switched Key Performance Indicators as a measure of the current performance of the network. It can be used to evaluate an already dimensioned network in which recent traffic changes have been made in limited regions to infer the possible problematic areas and then to improve the network dimensioning with respect to these changes.
The concept of this computation is the inverse of that of the dimensioning process. In this case, Atoll has the results of the dimensioning process already committed and known. Atoll then computes the current values for all the KPIs knowing the number of required TRXs, the respective numbers of shared and dedicated timeslots and the circuit switched and packet switched traffic demands.
The computation algorithm utilizes the parameters set in the dimensioning model properties and the quality curves for the throughput reduction factor, delay and the blocking probability.
The following conventional relations apply:
If,
• TSC, dedicated is the number of timeslots dedicated to the circuit switched traffic,
• TSP, dedicated is the number of timeslots dedicated to the packet switched traffic,
• TSS is the number of shared timeslots for a transmitter,
Then, the number of timeslots available for the circuit switched traffic, TSC, is defined as:
And the number of timeslots available for the packet switched traffic, TSP, is given by:
5.7.1 Circuit Switched Traffic
For each subcell, Atoll has already calculated the effective traffic overflow rate and the blocking rate during the dimensioning process. Also knowing the circuit switched traffic demand, TDC, and the number of timeslots available for circuit switched traffic, TSC, the blocking probability can be easily computed using the Erlang formulas or tables.
5.7.1.1 Erlang B
Under the current conditions of circuit switched traffic demand, TDC, and the number of timeslots available for the circuit switched traffic, TSC, the percentage of blocked circuit switched traffic can be computed through:
In a network dimensioning based on Erlang B model, the circuit switched traffic overflow rate, OC, is the same as the percentage of traffic blocked by the subcell calculated above.
5.7.1.2 Erlang C
Similarly, under the current conditions of circuit switched traffic demand, TDC, and the number of timeslots available for the circuit switched traffic, TSC, the percentage of delayed circuit switched traffic can be computed through:
L STC+STP
TSC dedicated +TSP dedicated +TSS
---=
TSC = TSS+TSC dedicated
TSP= TSS+TSP dedicated
% of blocked traffic
TDC
TSC TSC
!
---TDC
k ---k!
k=0 TSC
---=
If the circuit switched traffic demand, TDC, is higher than the number of timeslots available to accommodate circuit switched traffic, the column for this result will be empty signifying that there is a percentage of circuit switched traffic actually being rejected rather than just being delayed under the principle of Erlang C model.
The circuit switched traffic overflow rate, OC, will be 0 if the circuit switched traffic demand, TDC, is less than the number of timeslots available for the circuit switched traffic, TSC.
If, on the other hand, the circuit switched traffic demand, TDC, is higher than the number of timeslots available to carry the circuit switched traffic, TSC, then there will be a certain percentage of circuit switched traffic that will overflow from the subcell. This circuit switched traffic overflow rate, OC, is calculated as:
5.7.1.3 Served Circuit Switched Traffic
The result of the above two processes will be a traffic overflow rate for the circuit switched traffic for each subcell, OC. The served circuit switched traffic, STC, is calculated as:
5.7.2 Packet Switched Traffic
Identifying the total traffic demand, TDT, (circuit switched traffic demand + packet switched traffic demand) as:
The following two cases can be considered.
5.7.2.1 Case 1: Total Traffic Demand > Dedicated + Shared Timeslots
In the case where the total number of timeslots available is less than the total traffic demand, there will be packet switched data traffic that will be rejected by the subcell as it will not be able to accommodate it. The following results are expected in this case:
5.7.2.1.1 Traffic Load
The traffic load will be 100%, as the subcell will have more traffic to carry than it can. This implies that the system will be loaded to the maximum and even saturated. Hence the user level quality of service is bound to be very unsatisfactory.
5.7.2.1.2 Packet Switched Traffic Overflow
In a 100% loaded, or even saturated subcell, the packet switched data calls will start being rejected because of shortage of available resources. Hence there will be a perceptible packet switched traffic overflow in this subcell, OP. This overflow rate is calculated as show below:
5.7.2.1.3 Throughput Reduction Factor
The resulting throughput reduction factor for a 100% loaded or saturated subcell will be 0. Hence, the throughput perceived by the packet switched service user will be 0, implying a very bad quality of service.
5.7.2.1.4 Delay
Again for a 100% loaded or saturated subcell, the delay at the packet switched service user end will be infinite as there is no data transfer (throughput = 0).
5.7.2.1.5 Blocking Probability
All the data packets will be rejected by the system since it is saturated and has no free resources to allocate to incoming data packets. Hence, the blocking probability will be 100%.
5.7.2.1.6 Served Packet Switched Traffic
With the packet switched data traffic overflowing from the subcell, there will be a part of that traffic that is not served. The served packet switched data traffic, STP, is calculated on the same principle as the served circuit switched traffic:
% of traffic delayed TDCTSC TDC
TSC TSC! 1 TDC TSC
--- –
TDCk
---k!
k=0 TSC–1
+
---=
OC TDC–TSC TDC
---=
STC = TDC1–OC
TDT = TDC+TDP
OP 1 TSC dedicated +TSP dedicated +TSS ST– C TDP
---– 100
=
STP = TDP1–OP
5.7.2.2 Case 2: Total Traffic Demand < Dedicated + Shared Timeslots
In the case where the total traffic demand is less than the number of timeslots available to carry the traffic, the subcell will not be saturated and there will be some deducible values for all the data KPIs. In a normally loaded subcell, the packet switched data traffic will have no overflow percentage. This is due to the fact that the packet switched data traffic is rather placed in a waiting queue than be rejected.
Therefore, there will be a within limits packet switched traffic load, LP, calculated as under:
The second parameter for computing the KPIs from the quality curves of the dimensioning model is the number of equivalent timeslots available for the packet switched data traffic, NP, which is calculated in the same manner as in the dimensioning process as well:
These parameters calculated, now Atoll can compute the required KPIs through their respective quality curves.
5.7.2.2.1 Traffic Load
The traffic load is computed knowing the total traffic demand and the total number of timeslots available to carry the entire traffic demand:
5.7.2.2.2 Packet Switched Traffic Overflow
In a normally loaded subcell, no packet switched data calls will be rejected. The packet switched traffic overflow will, therefore, be 0.
5.7.2.2.3 Throughput Reduction Factor
The resulting throughput reduction factor for a normally loaded subcell is calculated through the throughput reduction factor quality curve for given packet switched traffic load, LP, and number of equivalent timeslots, NP.
5.7.2.2.4 Delay
The resulting delay the subcell is calculated through the delay quality curve for given packet switched traffic load, LP, and number of equivalent timeslots, NP.
5.7.2.2.5 Blocking Probability
The resulting blocking probability for a normally loaded subcell is calculated through the blocking probability quality curve for given packet switched traffic load, LP, and number of equivalent timeslots, NP.
5.7.2.2.6 Served Packet Switched Traffic
As there is no overflow of the packet switched traffic demand from the subcell under consideration, the served packet switched traffic will be the same as the packet switched traffic demand:
5.8 Neighbour Allocation
The intra-technology neighbour allocation algorithm takes into account all the TBC transmitters. It means that all the TBC transmitters of the .atl document are potential neighbours.
The transmitters to be allocated will be called TBA transmitters. They must fulfil the following conditions:
• They are active,
• They satisfy the filter criteria applied to the Transmitters folder,
• They are located inside the focus zone,
• They belong to the folder on which allocation has been executed. This folder can be either the Transmitters folder or a group of transmitters or a single transmitter.
Only TBA transmitters may be assigned neighbours.
LP
STC–TSC dedicated TDP
Timeslots
+
TSP
---=
NP TSP TSTerminal
---=
Traffic Load TDT
TSC dedicated +TSP dedicated +TSS
---=
STP= TDP
Note:
• If no focus zone exists in the .atl document, Atoll takes into account the computation zone.
5.8.1 Global Allocation for All Transmitters
We assume a reference transmitter A and a candidate neighbour, transmitter B.
When automatic allocation starts, Atoll checks following conditions:
6. The distance between both transmitters must be less than the user-definable maximum inter-site distance. If the distance between the reference transmitter and the candidate neighbour is greater than this value, then the candidate neighbour is discarded.
7. The calculation options,
Force co-site transmitters as neighbours: This option enables you to force transmitters located on the reference transmitter site in the candidate neighbour list. This constraints can be weighted among the others and ranks the neighbours through the importance field (see after).
Force adjacent transmitters as neighbours: This option enables you to force transmitters geographically adjacent to the reference transmitter in the candidate neighbour list. This constraints can be weighted among the others and ranks the neighbours through the importance field (see after).
Force neighbour symmetry: This option enables user to force the reciprocity of a neighbourhood link. Therefore, if the reference transmitter is a candidate neighbour of another transmitter, the later will be considered as candidate neighbour of the reference transmitter.
Force exceptional pairs: This option enables you to force/forbid some neighbourhood relationships. Therefore, you may force/forbid a transmitter to be candidate neighbour of the reference transmitter.
Delete existing neighbours: When selecting the Delete existing neighbours option, Atoll deletes all the current neighbours and carries out a new neighbour allocation. If not selected, the existing neighbours are kept.
8. There must be an overlapping zone ( ) with a given cell edge coverage probability where:
• SA is the area where the received signal level from the transmitter A is greater than a minimum signal level. SA is the coverage area of reference transmitter A restricted between two boundaries; the first boundary represents the start of the handover area (best server area of A plus the handover margin named “handover start”) and the second boundary shows the end of the handover area (best server area of A plus the margin called “handover end”)
• SB is the coverage area where the candidate transmitter B is the best server.
Notes:
• Adjacence criterion: Geographically adjacent transmitters are determined on the basis of their Best Server coverages in 2G (GSM GPRS EDGE) projects. More precisely, a transmitter TXi is considered adjacent to another transmitter TXj if there exists at least one pixel of TXi Best Server coverage area where TXj is the 2nd Best Server. The ranking of the adjacent neighbour transmitter increases with the number of these pixels. The figure below shows the above concept.
• When this option is checked, adjacent cells are sorted and listed from the most adjacent to the least, depending on the above criterion. Adjacence is relative to the number of pixels satisfying the criterion.
• This criteria is only applicable to transmitters belonging to the same HCS layer. The geographic adjacency criteria is not the same in 3G (UMTS HSPA, CDMA2000) projects.
SASB
Atoll calculates either the percentage of covered area ( ) if the option “Take into account Covered Area” is
selected, or the percentage of traffic covered on the overlapping area for the option “Take into account Covered Traffic”. Then, it compares this value to the % minimum covered area (minimum percentage of covered area for the option
“Take into account Covered Area” or minimum percentage of covered traffic for the option “Take into account Covered Traffic”). If this percentage is not exceeded, the candidate neighbour B is discarded.
The coverage condition can be weighted among the others and ranks the neighbours through the importance field (see number 4 below).
9. The importance values are used by the allocation algorithm to rank the neighbours according to the allocation reason, and to quantify the neighbour importance.
Atoll lists all neighbours and sorts them by importance value so as to eliminate some of them from the neighbour list if the maximum number of neighbours to be allocated to each transmitter is exceeded. If we consider the case for which there are 15 candidate neighbours and the maximum number of neighbours to be allocated to the reference transmitter is 8.
Among these 15 candidate neighbours, only 8 (having the highest importances) will be allocated to the reference transmitter.
As indicated in the table below, the neighbour importance depends on the neighbourhood cause; this value goes from 0 to 100%.
Except forced neighbour case (importance = 100%), priority assigned to each neighbourhood cause is now linked to the (IF) Importance Function evaluation. The importance is evaluated through a function (IF), taking into account the following 3 factors:
• Co-site factor (C) which is a Boolean factor,
• Adjacency factor (A) which deals with the percentage of adjacency,
• Overlapping factor (O) meaning the percentage of overlapping Figure 5.13Overlapping Zones