Part3_Advance Network Optimization_Solution Finding

Advance Network Optimization Solution Findings NPO Refresher Course July, 1st to 3rd 2010 Voda Vodafon fone e MS – RoB RoB

Jignesh Parmar [email protected] Nokia Siemens Networks National NPO, Ahmedabad, India NSN Internal Document 1 © No N okia Siem ens Networks

Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

Performance Optimization - Introduction Typical performance problems which have to be improved Indoor coverage problems

CS / PS Capacity problems Downlink coverage problems Downlink interference problems Uplink Signaling (access) problems

NSN Internal Document 2 © Nokia Siemens Networks


Performance optimization - basic CS/PS flows Paging Optimization

SDCCH Optimization

TCH Optimization

• Accessibility

• Accessibility • Signaling

• Retainability • Quality • Traffic / TSLs allocation • Data rates

CS – Basic Call Flow PCH AGCH RACH

Get SDCCH

Establish SDCCH connection

Get TCH

Establish TCH connection

Call phase

Release phase

TBF Session

Release phase

PS – MO Uplink TBF Ready State

Send RACH


Establish immediate assignment

Get PDCH

Requested TSLs to Allocate


Performance optimization Performance optimization will include here:

CS

PS

• Accessibility

• MM and SM Signaling

• Signaling problems

•GPRS Attach •PDP Context activatios

• Retainability • TCH Dropped calls

• RLC/MAC TSL data rate •TBF Failure

• Quality

• E2E Data Rate

• interference

•Throughput

• Traffic • Traffic handling

• Multislot usage •Territory downgrades/upgrades •PS Blocking

• Mobility NSN Internal Document 4 © Nokia Siemens Networks

•Cell reselection Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

Performance optimization – Accessibility Signaling

CS

If problems with RACH, AGCH, PCH capacity • Check if combined signaling is existing. If it is existing, it should be removed Here is combined sdcch in TSL0 Should be TSL0 = BCCH TSL1= SDCCH

• Check if TRXSIG is big enough, default value for TRXSIG is 32k.If lots of SMS traffic + HR is used 64k should be used.

• Check if there are any bad interference problems – Bad interference problem can cause lot of repetitions => more blocking • Check CCCH related parameters

• Check if location areas are optimized – Sizes – Borders NSN Internal Document 5 © Nokia Siemens Networks


Performance optimization – Retainability TCH drop

CS

Most Common TCH drop reasons can be seen here Drop Reason TCH radio fail call TCH radio fail old TCH tc fail call TCH tc fail old Lapd fail BTS fail user act BCSU reset cnfg act Act fail call Abis fail call Abis fail old A if fail call A if fail old


Counter

Description

c1013

Transactions ended due to radio failure.

c1014 c1029 c1030 c1046 c1047 c1048 c1049 c1050 c1081 c1084 c1085 c1087 c1088

Transactions ended due to old channel failure in HO. Transaction failures due to transcoder failure. Transaction failures due to transcoder failure on old channel during HO. Transaction failures due to Lapd problems. Transaction failures due to BTS problems. Transaction failures due to user actions. Transaction failures due to BCSU reset. Transaction failures due to radio network Channel activation failures during call. Abis failures during call. Abis failures on old channel during TCH HO. A if failures during call. A if failures on old channel during TCH HO.


Performance optimization – Retainability TCH drop

CS

TCH DROP Distribution

Excersise What can be typical reason for TCH radio drops? NSN Internal Document 7 © Nokia Siemens Networks


Performance optimization – Quality RX Quality – Level Distribution Rx Quality x Rx Level

CS

NOTE! Same level – quality distribution for both UL and DL HW Problem: Bad Quality

Good Quality

for all Rx Levels Coverage Problem: Bad quality and Low Rx Level

HW Problem All samples below 100dBm CL10  <-100dBm

High Rv Level Interference Problem: Bad quality and High Rx Level NWD report 204 model



Performance optimization – RX Quality - Level CS Here are some RX quality – level distribution examples Like Normal distribution CL10 CL15 CL20 CL30 CL40

10645 47043 56204 200863 12785

CL10 CL15 CL20 CL30 CL40

Q0 5323 23522 28102 67544 6393


Q0 200863 16543 2654 4543 5114

8516 37634 44963 160690 10228

6813 30108 35971 128552 8182

5450 24086 28776 102842 6546

HW Problem, TRX/combiner etc is broken 4360 9865 6574 17654 456

3488 7543 10324 9876 112

2791 5643 345 145 24

2232 2345 65 28 23


3

5323 11761 14051 33772 3196

4

Q0 5323 23522 28102 67544 6393

HW problem, Q4 amount of samples is strange

1

Q1 4258 18817 22482 54035 5114

Q2 3406 15054 17985 43228 4091

Q3 2725 12043 14388 34582 3273

Q4 2180 18765 36542 150073 9876

Q5 1744 3772 5162 4938 56

Q6 1395 2822 173 73 12

4258 9409 11241 27017 2557

3406 7527 8993 21614 2046

2725 6022 7194 17291 1636

2180 9383 18271 75037 4938

1744 1886 2581 2469 28

4563 65432 65438 18765 3659

9765 7675 63562 14523 2648

Q4 2180 13234 2123 3634 4091

Q5 1744 10588 1699 2908 3273

Q6 1395 8470 1359 2326 2618

Q7 1116 6776 1087 1861 4532

Interference problem Q7 1116 1173 33 14 12


Q1 4258 13234 2123 3634 4091

Q2 3406 10588 1699 2908 3273

Q3 2725 8470 1359 2326 2618

HW Problem, TRX/combiner etc is broken

2

Q1 160690 13234 2123 3634 4091

Q2 128552 10588 1699 2908 3273

Q3 102842 8470 1359 2326 2618

Q4 17654 6776 1087 1861 2095

Q5 9876 2822 173 73 12

Q6 7865 1173 33 14 12

Q7 6543 234 65 28 23

1.

Lots of Q4 samples. Distribution is not OK  HW problem. Site reset will help

2.

Lots of bad signal level samples. TRX/Combiner is broken. If no HW problem  Site is totally in wrong place, site is like transferring traffic to another cell ( cause level HO). Typically a HW problem

3.

Lots of Q6 and Q7 samples. Distribution is not OK, no Q5 samples  HW problems. Site reset or broken TRX

4.

Typical normal interference problem., lots of q4…q7 samples in good signal level



Performance optimization – Interference, UL CS q0 -100dBm 10645 -95dBm 47043 -90dBm 56204 -80dBm 200863 -70dBm 1234 -47dBm 24

q1 8516 37634 44963 160690 987 19

q2 6813 30108 35971 128552 790 15

q3 5450 24086 28776 102842 632 12

q4 4360 9865 16574 17654 505 10

q5 3488 7543 5676 3653 404 8

q6 2791 5643 845 145 323 6

q7 2232 2345 65 28 259 5

q0 -100dBm 10645 -95dBm 47043 -90dBm 56204 -80dBm 200863 -70dBm 1234 -47dBm 24

q1 8516 37634 44963 160690 987 19

q2 6813 30108 35971 128552 790 15

q3 5450 24086 28776 102842 632 12

q4 4360 9865 16574 17654 505 10

q5 3488 7543 5676 3653 404 8

q6 2791 5643 845 145 323 6

q7 2232 2345 65 28 259 5

Bad interference problems. By POC parameter interference can be decreased, how power is adjusted etc. Also optimum MS power feature improves UL interference, no full power is sent after HO.


q1 8516 37634 44963 160690 987 19

q2 6813 30108 35971 128552 790 15

q3 5450 24086 28776 102842 632 12

q4 4360 9865 16574 17654 505 10

q5 3488 7543 5676 3653 404 8

q6 2791 5643 845 145 323 6

q7 2232 2345 65 28 259 5

Bad quality sample due to signal level problems. Diversity should be checked, also possibilities to use LNA to improve UL signal level. Antenna place should be also checked if there are some obstacles near the antenna.



MS power control can be seen here. If Power is reduced → no bad UL problems. BSC border is increasing good level samples

Performance optimization – Interference

CS

UL examples -100dBm -95dBm -90dBm -80dBm -70dBm -47dBm

-100dBm -95dBm -90dBm -80dBm -70dBm -47dBm


UL_q0 490 8410 38512 219137 509591 244302

UL_q0 70004 80339 106883 246892 254875 23288

UL_q0 251992 123653 62938 27005 2831 751

UL_q1 137 3768 5249 4600 2504 582

UL_q1 5020 392 233 227 155 21

UL_q1 30668 1151 247 51 3 0

UL_q2 368 3225 3066 2453 1812 711

UL_q2 3500 286 318 314 298 27

UL_q2 20552 403 144 65 9 4

UL_q3 605 3023 2323 2311 3050 1363

UL_q3 2564 461 602 860 935 103

UL_q3 18417 692 288 149 28 10

UL_q4 1014 2767 1692 1470 1271 1713

UL_q4 2232 328 199 298 203 109

UL_q4 18156 519 353 177 87 25

UL_q5 1378 2304 1570 1724 1465 873

UL_q5 1946 250 146 338 135 11

UL_q5 18286 351 129 82 10 1

UL_q6 16983 221 62 40 6 0

UL_q6 1830 1561 1191 1864 1649 671

UL_q6 2190 275 234 395 166 41

UL_q7 12478 72 29 16 0 0

UL_q7 2586 1859 1268 1221 337 87

UL_q7 2579 92 64 60 15 0

DL_q0 27862 89094 142228 222462 44504 5994

Some UL interference in good signal level. UL power control is not working properly. Ul power control is good indicator, if power is adjusted, there are no big problems in UL direction

There is no UL interference, or just a little. MS is adjusting power properly, there are only little samples in good UL signal level.

DL_q1 5537 4236 2732 1523 91 49

DL_q2 5529 3853 2550 1355 123 17

DL_q3 6069 3576 2356 1552 151 17

DL_q4 6109 2893 1510 716 81 12

DL_q5 6107 2264 1114 812 101 29

DL_q6 5389 1372 774 1081 191 65

Huge amount of UL bad samples. Cells is not working properly, it is like transferring traffic, UL quality/ UL level HO’s are triggering immediately. These kind of cells must be investigated.



DL_q7 4119 584 404 343 88 10

Performance optimization – Interference, DL

CS


q1 8516 37634 44963 160690 10228 1123

q2 6813 30108 35971 128552 8182 583

q3 5450 24086 28776 102842 6546 452

q4 4360 9865 16574 17654 456 261

q5 3488 7543 5676 3653 112 76

q6 2791 5643 845 145 24 26

q7 2232 2345 65 28 23 2

Really bad interference problem → signal level is really good (<-70dBm) and usually no better cell available → no HO → samples can be seen in the table. Interference is causing drops.


q1 8516 37634 44963 160690 10228 1123

q2 6813 30108 35971 128552 8182 583

q3 5450 24086 28776 102842 6546 452

q4 4360 9865 16574 17654 456 261

q5 3488 7543 5676 3653 112 76

q6 2791 5643 845 145 24 26

q7 2232 2345 65 28 23 2

Bad interference problem → signal level good (<-80dBm) and sometimes no better cell available. If better cell available and quality samples are 4 or worse → HO (reason quality or interference, depends on the parameter) Interference is causing drops.

q0 10645 47043 56204 200863 12785 4583

q1 8516 37634 44963 160690 10228 1123

q2 6813 30108 35971 128552 8182 583

q3 5450 24086 28776 102842 6546 452

q4 4360 9865 16574 17654 456 261

q5 3488 7543 5676 3653 112 76

q6 2791 5643 845 145 24 26

q7 2232 2345 65 28 23 2


q1 8516 37634 44963 160690 10228 1123

q2 6813 30108 35971 128552 8182 583

q3 5450 24086 28776 102842 6546 452

q4 4360 9865 16574 17654 456 261

q5 3488 7543 5676 3653 112 76

q6 2791 5643 845 145 24 26

q7 2232 2345 65 28 23 2



Situation is “network is working properly” If there are quality 4 or worse samples → quality HO. Most of the samples are q4 samples. If lots of q5..q7 samples → interference problem and interference must be analyzed / removed. If quality HOs but no q5..q7 samples → better cell is available → no interference problems. In these signal levels overlapping exists and if handover reason is no PBGT, it will be quality HO. By parameter amount of quality HOs can be adjusted

Bad quality samples due to signal level problems. If PBGT overlapping is not existing → lots of quality HOs + level HOs (margin are lower than in PBGT). Not interference problem, more coverage problem.

Check how much samples vs. HOs better cells available or not. Advance Network Optimization/JP/NNPO/ 1 to 3 July 2010 @ VF RoB st

rd

→

are

Performance optimization – Interference

CS

DL, examples -100dBm -95dBm -90dBm -80dBm -70dBm -47dBm

DL_q0 12057 44818 98587 225919 88708 15881

DL_q1 2827 4811 7107 7450 1014 84

DL_q2 3108 5041 7400 7731 971 109

DL_q3 3952 5866 8334 8445 998 122

DL_q4 4783 6587 8470 7726 751 104

DL_q5 6200 7223 8781 7441 688 167

DL_q6 7013 7259 7825 5695 689 199

DL_q7 8156 6781 6162 3369 367 184


DL_q0 8006 24951 57559 200602 304206 108047

DL_q1 1636 1636 2171 5771 5464 2134

DL_q2 1681 1627 1884 4686 4310 1908

DL_q3 2197 1767 1651 4130 3796 1623

DL_q4 2379 1037 781 1736 1315 689

DL_q5 2510 431 330 566 350 230

DL_q6 2025 175 161 254 153 129

DL_q7 1290 53 47 97 105 64


DL_q0 7055 20109 34531 107539 177614 58718

DL_q1 1398 1307 1053 875 283 78

DL_q2 1374 1274 745 518 316 91

DL_q3 1906 1161 587 630 663 198

DL_q4 2163 694 273 161 61 54

DL_q5 2003 332 131 98 32 40

DL_q6 1468 211 94 113 29 54

DL_q7 832 84 41 47 9 32


There are almost as much samples Q5 and Q7 samples as Q 4 samples → even interference is really bad or there is no better cell available ( no ho’s after bad quality samples). These kind of interference cells should be optimized, otherwise there are lots of drops etc

There are no as much Q5…Q7 samples as Q4 samples → after interference samples Quality HO is done or the interference situation is not so bad, for example sampling is Q0,Q2,Q4,Q2,Q5,Q0,Q2,Q3,Q4,Q2 → quality HO is not triggering

There are bad quality samples only due to signal level problems.


Performance optimization, Interference

CS

(internal / external)

•Internal Interference – Interference can be seen from stats or can be measured by scanner. – Neighbor cells (DL) or mobiles (UL) are causing interference. – By frequency / network planning interference can be decreased.

•External Interference – Interference can be seen from stats or can be measured by scanner. – External radio frequencies are causing interference  Military use  In the border area, interference is coming from other country.  Some external wireless system (for example some wireless industry system)

is causing interference  Increased I level can be also due to external interference



Performance optimization – Traffic Traffic handling How to balance traffic between other cells in same site? •HO parameters

Blocking in cellA cellA

– PBGT – Quality /interference margins – Level margins

•HO Features

Traffic will be pushed to neighbor cells

cellB

– Umbrella – Traffic reason HO – DR

•Traffic layers – 1800 (dual band) – CBCCH NSN Internal Document 15 © Nokia Siemens Networks

CS


cellC

Performance optimization – Traffic Traffic handling How to balance traffic between other sites? • HO parameters

CS

cellA Some blocking in cellA

– PBGT – Quality /interference margins – Level margins

Traffic will be pushed To neighbor cells

• HO Features – Umbrella – Traffic reason HO – DR cellB

cellC



Performance optimization – PS Before PS Optimization basic GSM optimization should be done

1. As high signal level as possible • It means that even the indoor signal level should be high enough to have MCS9 for getting the highest data rate on RLC/MAC layer.

2. As low interference as possible • The aim of having high C/I is to avoid throughput reduction based on interference.

3. Enough capacity • Enough BSS hardware capacity (interface and connectivity) is needed to provide the required capacity for PSW services in time. Both CSW and PSW traffic management should be harmonized with the layer structure and long term plans. NSN Internal Document 17 © Nokia Siemens Networks


Performance optimization – GPRS Attach

PS

Typical failures outside BSS • User subscription – non-GPRS users and users from other operators with no roaming agreement

• Protocol error – Mainly caused by MS not responding during attach, some possible reasons:

 Radio contact lost  MS out of coverage

– Collision of signaling message – MS sends Attach Request or Detach Request when the previous Attach Request is still in process

Typical recommendations: • PS Core settings • Further RF optimization • GPRS detach feature



Performance optimization – PDP Context Activation

PS

Typical failures: • PS Core • Subscribers use incorrect APN during PDP Context Activation Request – Subscribers/MS keep reattempt with the incorrect APN can badly affect overall PDP Activation Success Rate – APN of other operators often seen, could because of subscribers using the same MS with different SIM



Performance optimization – TBF Failure

PS

Typical reasons for TBF failures • Radio reasons – Bad coverage – Interference problems – Blocking

TBF Failures

• Parameter discrepancies – MCA parameters – POC – GPRS_POC  IFP, TFP

Establishment

No

Failures

Response

TBF Attempts

• SGSN problems – Gb, Gn traces will be needed



Flush

CSW Traffic

Suspend Normally Ended TBFs

Performance optimization – Throughput

PS

• Radio reasons, throughput / timeslot is bad – Bad coverage – Interference problems ⇒

Decreased coding scheme => less throughput even many timeslots used

• TSL allocation problems, total throughput is bad – Check territory usage • Check downgrades due to CSW traffic and PSW rejects • Add more dedicated timeslots for data if possible

– Check TBF sharing • If lots of TBF/ TSL => check that CDED parameter <>0 • Too much sharing => territory not big enough. If possible, add more dedicated timeslots

• Transmission problems – EDAP congestion – Gg congestion



Performance optimization – Multislot Usage

PS

Territory downgrade and upgrade • Territory downgrade due to CSW traffic rise (c1179) – CS traffic handling • Territory downgrade due to less PSW traffic(c1181) – Territory optimization, territory is perhaps not big enough – CS Traffic handling • Territory upgrade request rejection beyond default territory

– Upgrade request beyond Default territory for additional resources (c1174), which can be rejected because of: 1. No (E)GPRS capable resource left (No (E)GPRS enable TRX or the maximum (E)GPRS capacity reached) 2. PCU and EDAP capacity limitation (256 Abis TSL per PCU) 3. High CSW load 0

1

Default territory

2 1174

3

4

5 1179

1181 1180



6

7

Performance optimization – PS Blocking

PS

• Hard blocking – no resources available – More TSLs for data must be available  Check unavailability in the cell  More dedicated timeslots for data  Amount of CS traffic must be reduced (traffic handling with features)

– RF optimization => better throughput => less blocking

• Sof blocking – TSL allocation problems – Check parameters  Territory parameters, CDED, CDEF  Check also also (E)GPRS parameters  Check also timers

• Too much delays => more blocking



Performance optimization – Cell Reselection Typical main causes: • Bad overshooting – Dominance areas are not clear – Too much cell reselections

• Link balance problems – UL / DL

• Cell reselection parameters are not properly planned – Check C1,C2, NCCR parameters – Check IFP, TFP parameters

• BSC / PCU area optimization is not properly done – Delays will be increased if areas are not properly planned NSN Internal Document 24 © Nokia Siemens Networks


PS

Traffic and Capacity Optimization



Traffic and Capacity Optimization GSM Architecture Traffic and Capacity Optimization levels • Ater, Transcoder (TC), A-interface • BSC level • Site level • TRX level • Timeslot level • Signaling

MSC

TC

A Ater Air interface

BTS 2G

BSC

Abis


BCSU


Gb

SGSN

Traffic and Capacity Optimization Introduction • Ater, TC, A – Capacity and pools

• BSC – BSC Type, BSC load, Gb links – BCSU, capacity – EDAP/PCU Type, Load

• Site – Antennas, BTS HW/SW BTS configuration, Dual band, CBCCH, SDCCH traffic etc

• TRX – Where capacity will be needed • Timeslot – FR/HR, AMR FR/HR, territory, CMAX, SDCCH traffic • Signaling – CCCH, TRXSIG, SS7 links NSN Internal Document 27 © Nokia Siemens Networks


Traffic and Capacity Optimization Ater, TC, A Abis interfaces

Ater Interface • All traffic from BSC must be transferred to transcoders. Number of E1’s in Ater / Asub interface is depending on: • TCH Traffic from all abis IF’s • CCS7 traffics • O&M links • Utilization rate

• 1 PCM timeslot in Ater/Asub interface: • 64Kbit/s • 1 PCM timeslot in Abis interface : • FR= 16 kbit/s • HR = 8 kbit/s => 4 or 8 PCM timeslots in abis = 1 PCM timeslot in Ater NSN Internal Document 28 © Nokia Siemens Networks


Traffic and Capacity Optimization Ater, TC, A Required transcoding (TC) capacity is derived from • Total number of required Ater Channels. • Number of Ater channels is based on total Erlangs in BSC and blocking in A-interface

• Configuration options for TCSM3i for standalone installations can be seen below TCSM3i units

(capacity step)

1 2 3 4 5 6 7 8 9 10 11 NSN Internal Document 12 29

© Nokia Siemens Networks

ETSI channels ANSI channels

960 1920 2880 3840 4800 5760 6720 7680 8640 9600 10560 11520

760 1520 2280 3040 3800 4560 5320 6080 6840 7600 8360 9120

ET 16

ET 16 TR3E

(16 E1 PCMs)-1 BSC

(16 E1 PCMs) several BSCs

3 5 8 10 13 15 18 20 23 25 28 30

3 6 9 12 15 18 21 24 27 30 33 36

8 16 24 32 40 48 56 64 72 80 88 96


Note. In some solution Transcoder can be implemented to Media Gateway MGW

Traffic and Capacity Optimization Ater, TC, A A Interface Number of required A- Interface PCMs is derived from total number of required Ater Channels. 4:1 multiplexing 1 : 120

For Example • Total Erlangs in BSC = 3920 Erl • Blocking in A-interface =0.1% →

→ →

ErlangB table(3920,0.1QoS ) → 4042 Ater Channels

Ater Interface PCMs A interface PCMs(1:4)


ETSI Channels 960 1920 2880 3840 4800 5760 6720 7680 8640 9600 10560 11520

4:1

Ater interface PCMs 8 16 24 32 40 48 56 64 72 80 88 96

A interface PCMs (4:1)

= 4042(ETSI) / 120 = 34 = 34*4 = 136


32 64 96 128 160 192 224 256 288 320 352 384

Traffic and Capacity Optimization BSC • Example of BSC capacity evolution Static limits # TRX #BTS #BCF #BSCU #SS7L #LAPD # (logical) PCU # abis 16k chann #E1/T1 lines Erlangs

BSC2i

BSC3i 660

512 512 248 6 16 992 16 4096 144 3040

660 660 504 6 16 1236 24 6144 256 3920

BSC3i 1000/2000 1 cab 2 cab 1000 2000 1000 2000 1000 2000 5 10 16 16 2240 4480 50 100 12800 25600 384 800 5940 11880

BSC3i 3000 3000 3000 3000 6 16 5760 30 30720 800 17820

How Erlangs are calculated? • This depends on used traffic profile • Different traffic profile => different results Example, How to calculate BH call attempts = (Total_Erl x 3600s) / mean_holding_time(s) NSN Internal Document 31 © Nokia Siemens Networks

Traffic Profile example Mean holding time Mobile originated calls (MO) Mobile terminated calls (MT) Handovers per call Location updated per call IMSI detach per call Paging response SMS (req(subs/1 hour) Terminated SMS 80% QoS


120s 70% 30% 1.5 2 0.1 63% 1 80% 2%

Traffic and Capacity Optimization BSC - Example of Utilization BSC Utilization • TRX Usage • TCH Usage

Actual TRX utilization Actual TCH Utilization

If utilization values are too high => BCS Capacity optimization will be needed

Utilization analysis can be used for long term BSC optimization Capacity management vs. configuration optimization



Traffic and Capacity Optimization BSC - EDAP / PCU / Gb

• EDAP capacity can be optimized based on KPI values – DL / UL MCS selection limited by EDAP (dap_7a, dap_8c) – DL / UL MCS selection limited by PCU (dap_9, dap_10) – Peak DL EDAP usage (c76004) – Peak UL EDAP usage (c76005) – Territory upgrade rejection due to lack of PCU capacity (blck_32) – Not too much capacity nor too less capacity

• EDAP / PCU performance can be decreased due to Gb link size – UL/ DL Gb load: frl_7a/ frl_8a – This should be checked always with EDAP / PCU optimization



Traffic and Capacity Optimization BSC - EDAP / PCU

• EDAP blocking – If EDAP utilization is below 100 => increasing EDAP does not provide any help => PCU optimization needed

– As a rule of thumb the following estimation may be used to calculate number of E1’s. Nbr of TRX

Amount of E1s

12 TRX GSM / GPRS

1 E1

9 TRX GSM / EGPRS 6 TRX GSM / Heavy EGPRS

Can be used also T1s

18 TRX GSM / GPRS

1.5 E1


24 TRX GSM / GPRS

2 E1



Note!


E1:32 channel T1:24 channel

Traffic and Capacity Optimization BSC - Gb

• Gb utilization – If values near thresholds, it is recommended to increase Gb link size OR Implement Gb over IP Maximum received load % ( DL )


Gb bandwidth [kbps]

Restricting Gb link utilisation [%]

128

25

256

61

384

68

512

68

640

70

768

75

896

85

1024

90


Traffic and Capacity Optimization Site - Indoor solution Indoor solutions Before

4 TRX, 20Erl

After

2 TRX, 5Erl

Indoor Site 4TRX, 15Erl

Note! Indoor solution is collecting traffic inside (heavy traffic) office => Less frequencies is needed outside the building



Traffic and Capacity Optimization Site - Dual Band Network Before

GSM 900 layer (4TRX:24Erl (Bad blocking), GSM 900 layer (2TRX:2Erl)

Actions:

After GSM 900 layer (2TRX:7Erl (no blocking) GSM 1800 layer (4TRX:21Erl)

• • • •

1800 layer was added From 900 layer traffic to 1800 layer 3 TRX was removed from 900 layers Total traffic was increased

GSM 900 layer (1TRX:1,5Erl

Note! Dual band layers can be also GSM 900 and WCDMA



Traffic and Capacity Optimization Site – Traffic layers Traffic layers

3G Layers • 2100 Macro • 2100 Micro • 2100 Indoor

2G Layers • • • •

Note! Traffic between layers can be handled by parameters NSN Internal Document 38 © Nokia Siemens Networks


900 Macro 1800 Macro 900/1800 Micro 900/1800 Indoor

Traffic and Capacity Optimization TRX – capacity where it is needed Site configuration 2+2+2 in all sites Bad Blocking area 8Erl 5Erl

3Erl

• More capacity is needed

9Erl 4Erl

2Erl

1Erl

5Erl 5Erl

6Erl 4Erl

1Erl

1Erl

Extra capacity 4Erl 3Erl NSN Internal Document 39 © Nokia Siemens Networks


• 2 trx / cell is too much •=> TRX can be transferred to the cells where these are more needed

Traffic and Capacity Optimization Timeslot – AMR FR/HR Half rate (HR) is recommended to use when: • More capacity will be needed but additional TRXs can not be added – Interference problems, Frequency reuse – HW limitations

• Temporary needs for additional capacity – Special events – Due to daily traffic profile, additional capacity will be needed just for short time • Big cruise is passing by, tens of calls during short time



Traffic and Capacity Optimization Timeslot – AMR FR/HR • Channel Mode Adaptation is an HO algorithm that aims at select the correct channel rate (FR or HR). • The selection of the channel rate depends on 2 main factors: load and quality Codec Good Quality

load

AMR FR

AMR FR

packing unpacking

AMR HR

AMR HR

Bad Quality NSN Internal Document 41 © Nokia Siemens Networks


Traffic and Capacity Optimization Timeslot – FR/HR & PS territory

Strategy – which timeslots for data and which for speech?

• BFG –parameter • Whether the BCCH TRX or other TRXs are preferred in GPRS channel allocation .

• FRL –parameter • With this parameter the percentage of full rate TCH resources that must be available for traffic channel allocation is defined.

• FRU –parameter • With this parameter the percentage of full rate TCH resources that must be available for traffic channel allocation is defined.

• Note! Territory and dual rate in same TRX(HR just after territory) => not good, should be changed TRX1

tsl0

tsl1

bcch

sdcch

tsl0

tsl1

NSN InternalTRX2 Document 42 © Nokia Siemens Networks

tsl2

tsl2

tsl3

tsl3

tsl4

tsl4

tsl5

tsl5

tsl6

tsl6


tsl7

tsl7

How to allocate CS and PS channels?

Traffic and Capacity Optimization – AMR and PS interworking, CS TCH Allocation Calculation CS TCH allocation calculation (CTC) •

Defines how the GPRS territory is seen when calculating FR resources.

EGPRS used/default

MBCCHC

x

x

x

x

x

TCHD

TCHD

TCHD

TCHD

TCHD

TCHD

5 CS calls CTC 0 Free FR TCH resources

CTC2

1

1

2

Working FR TCH resources 6

7

7

% of free FR resources

16.7%

14.3%

28.6%

HR Preferred?

N

Y

N

x EGPRS downgraded

CTC1

MBCCHC

TCHD

x

x

TCHD

TCHD

x TCHD

x

CTC Value 0 = Only CSW used RTSLs are used to calculate CDEF resources 1 = CSW and PSW used RTSLs are used. PSW used RTSLs are seen as occupied resource when calculating Territory downgrade due to PSW RTSLs CS traffic 2 = CSW used and PSW used RTSLs are used. PSW used RTSLs are seen as idle resource when calculating resources

x

TCHD

TCHD

downgrad ed

6 CS calls

FRL = relative amounts of free FR TCH resources in proportion to working FR TCH resources FRL = 15% in this example

Free FR TCH resources

1

1

1

Working FR TCH resources 7

7

7

% of free FR resources

14.3%

14.3%

14.3%

HR Preferred?

Y

Y

Y



Traffic and Capacity Optimization Timeslot – AMR FR/HR HR will be allocated when there are few free FR TCHs available. • The purpose is to avoid congestion / blocking • HR => FR when there are again enough free FR TCHs • Thresholds are set by parameters (FRL, FRU) Free FR TCHs Upper limit for free FR TCHs

Lower limit for free FR TCHs

Time No packing of FR calls


Packing of FR calls


No packing of FR calls

Traffic and Capacity Optimization signaling -paging Paging TRXSIG can be dimensioning based on following table: • After implementation paging capacity KPIs and occurrence of LAPD overload must be checked TRXSIG

Comments

16k

Can be used if 32k is not possible to use

32k

Default value. Must be used if HR is used or high

64 k

Must be used if HR is used and high SMS traffic

Note! If the cell is part of the large location area, high paging load is expected. Thus small signaling links (16kbps) shall be avoided. NSN Internal Document 45 © Nokia Siemens Networks


Traffic and Capacity Optimization signaling -sdcch • Capacity of SDCCH in the cell is depending on the amount of TRX’s in the cell • Capacity of SDCCH in the cell is also depending heavily on Traffic profile, for example amount of SMS’s Number of TRX in the cell 1 2 3 4 5

SDCCH Traffic (Erl) Number of SDCCH in the cell Erl mErl/subs subs 1 sms/subs/hour 5 sms/subs/hour 10 sms/subs/hour 1 sms/subs/hour 5 sms/subs/hour 10 sms/subs/hour 2.2 25 88 0.4 0.8 1.3 1 1 1 8.2 25 328 1.6 3.1 4.9 1 2 2 14.9 25 596 3.0 5.6 8.9 1 2 3 21 25 840 4.2 7.9 12.6 2 2 3 27 25 1080 5.4 10.2 16.2 2 3 4

• SDCCH blocking KPIs should be monitored & avoided – And reason for blocking, SMS, LU, paging etc should be known.

• Note! Exceptions are cells that cover ports of entry, such as airports, where a large number of subscriber location updates are expected NSN Internal Document 46 © Nokia Siemens Networks


Traffic and Capacity Optimization signaling - sdcch Increased Dynamic SDCCH • SDCCH resources are used for call establishment, location updates and short messages (SMS) • Dynamic SDCCH allocation feature configures additional SDCCH resources according to the traffic situation in a cell

• When a BTS needs temporarily larger SDCCH capacity, then idle TCH resources are configured for the SDCCH use • When the congestion situation is over extra SDCCH resources are configured immediately back to TCH resources • By using Increased Dynamic SDCCH Capacity, you can use up to 32 SDCCHs per non-BCCH TRX and 24 per BCCH TRX NSN Internal Document 47 © Nokia Siemens Networks

BCCH SDCCH/8 TCH

TCH

TCH

TCH

TCH

TCH

BCCH SDCCH/8SDCCH/8

TCH

TCH

TCH

TCH

TCH

• In case of S DC C H c ongestion, one free traffic channel can be changed dynamically to S DC C H/8 • When S DC C H/8 is no longer needed it is changed dynamically back to TC H


Coverage Optimization



Solution finding – coverage optimization

20 km away from site • Signal level -95dBm? • Are both cases critical ones? • Are both cases coverage problem

20km

cases? Bad coverage – what does it mean? 20km



SiteA

Solution finding – coverage optimization Signal level – quality distribution 12478 samples

251992 samples • signal level < -100dBm • Quality 0

• signal level < -100dBm • Quality 7 => UL coverage problem?

=> UL coverage problem?

Exercise How we should here optimize coverage? • Any problems? QUALITY L E V E L L A N G I S

UL DL



UL_q0 251992 123653 62938 27005 2831 751

UL_q1 30668 1151 247 51 3 0

UL_q2 20552 403 144 65 9 4

UL_q3 18417 692 288 149 28 10

UL_q4 18156 519 353 177 87 25

UL_q5 18286 351 129 82 10 1

UL_q6 16983 221 62 40 6 0

UL_q7 12478 72 29 16 0 0


DL_q0 27862 89094 142228 222462 44504 5994

DL_q1 5537 4236 2732 1523 91 49

DL_q2 5529 3853 2550 1355 123 17

DL_q3 6069 3576 2356 1552 151 17

DL_q4 6109 2893 1510 716 81 12

DL_q5 6107 2264 1114 812 101 29

DL_q6 5389 1372 774 1081 191 65

DL_q7 4119 584 404 343 88 10


Solution finding – coverage optimization Signal level <-90…-95dBm is many times level threshold • Typical values for ul/dl level HO (UL, -95dBm, DL, -90dBm) • UL / DL coverage should be always analyzed separately • Here, lots of bad UL signal level samples => UL coverage should be improved QUALITY L E V E L L A N G I S

UL DL


UL_q0 251992 123653 62938 27005 2831 751

UL_q1 30668 1151 247 51 3 0

UL_q2 20552 403 144 65 9 4

UL_q3 18417 692 288 149 28 10

UL_q4 18156 519 353 177 87 25

UL_q5 18286 351 129 82 10 1

UL_q6 16983 221 62 40 6 0

UL_q7 12478 72 29 16 0 0


DL_q0 27862 89094 142228 222462 44504 5994

DL_q1 5537 4236 2732 1523 91 49

DL_q2 5529 3853 2550 1355 123 17

DL_q3 6069 3576 2356 1552 151 17

DL_q4 6109 2893 1510 716 81 12

DL_q5 6107 2264 1114 812 101 29

DL_q6 5389 1372 774 1081 191 65

DL_q7 4119 584 404 343 88 10



Solution finding – coverage optimization Solutions to improve coverage • Antenna changes  Tilting  Adding antenna with bigger gain

• New cell / site  Costly

• DL /UL amplifiers  DL, boosters  UL, Mast head amplifier (MHA)

• Diversity  UL

• Reducing losses  Combiner types  Feeder types

• Better site / antenna place NSN Internal Document 52 © Nokia Siemens Networks


Solution finding – coverage optimization indoor coverage •Many networks have insufficient indoor coverage due to high building penetration loss. This includes subways, under ground facilities and tunnels. •Outdoor coverage is normally ok in urban and city area. Interference is more an issue for outdoor users.

•Micro cell and Indoor cell are good solution to improve indoor coverage and network quality. – Antenna location is main factor for successful indoor coverage planning

Indoor BTS Indoor Panel Antenna

Optical Fiber

Remote Unit NSN Internal Document 53 © Nokia Siemens Networks


Master Unit

RF Cable

Solution finding – coverage optimization indoor coverage Other features can be also used for coverage optimization • Extended cell • Smart radio concept • Antenna hopping



Examples of wrong parameter set impact on network operation/performance – call setup, qual, bands

1. No calls happening in a cell • Cell Barred • Non existent (LAC, Cell ID) in MSC • DMAX = 0

2. Very few calls happening in a cell • RxLevAccesMin • Wrong MNC, MCC, LAC declaration

3. Very low traffic in a cell • msTxPwrMax = 0, bsTxPwrMax = 30

4. Bad quality in UL after rehoming 5. Few traffic in 1800 layer of a dual band 900/1800 network



Example 1a No calls happening in a cell: • The cell has been barred Handover

 C all S etup S Y S ( C I N e l l F O B a 3 r ( y e r A c c B C C s ) e s s H = )





Example 1b No calls happening in a cell: • CI different between MSC and BSC or non existent (LAC, Cell ID) in MSC


MSC does not find (LAC, CI) in its database Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

Example 1c No calls happening in a cell: • MsMaxDistanceInCallSetup (DMAX) = 0 Despite the coverage of the cell, no calls will be established!

DMAX = 0


Call Setup




RxLevel = -70 dBm

Example 2a Very few calls happening in a cell: • RxLevAccesMin Mapping

OSS database

BSS MML

Lowest range

0

-110 dBm

1

-109 dBm

...

...

62

-48 dBm

63

-47 dBm

Highest range

Be careful when setting parameter through xml or dat file! OSS database unit should be used to specify parameter value!

Cell coverage

RxLevAccessMin = -47 dBm



Example 2b Very few calls happening in a cell: • Wrong MNC, MCC, LAC declaration in network - MNC: 01 ≠ 1 !!! (In OSS correct value)

Note! Be careful when setting parameter through xml or dat file! OSS database unit should be used to specify parameter value!



Example 3 Very low traffic in a cell: •msTxPwrMax = 0, bsTxPwrMax = 30

l t a e ∆D

Parameter

bsTxPwrMax bsTxPwrMin minMsTxPower msTxPwrMax

Value_______

0 … 30 dB 0 … 30 dB 0 … 36 dBm 0 … 36 dBm

Attenuation Values (dB) BTS Max Power = BTS Power – (bsTxPwrMax = 0 dB) BTS Min Power = BTS Power – (bsTxPwrMin = 30 dB)

t e u l s o b A

Power Values (dBm) MS Max Power = (msTxPwrMax = 33 dB) MS Min Power = (minMsTxPower = 13 dB)



Default value

0 dB 30 dB 0 dB

Example 4 Bad quality in UL after rehoming: • DiversityUsed parameter not set to yes anymore

After rehoming RDIV parameter was set to default value (No) and UL quality was affected.

RDIV = Y

Uplink Diversity improves quality of signal received.



Example 5 Few traffic in 1800 layer of a dual band 900/1800 network: • Idle Mode: C2 parameters not set properly (temporaryOffset, penaltyTime) • Idle / dedicated mode parameters should be according to strategy

BCCH BCCH

slow moving mobile 1800: micro-Layer NSN Internal Document 63 © Nokia Siemens Networks


fast moving mobile 900: macro-layer

Frequency Optimization



Solution finding – frequency optimization GSM network is based on good frequency planning Bad frequency planning is causing high interference levels • Lots of drop calls • CS/PS quality is bad • Features are not working properly • Maybe new sites are built due to bad interference problems • Lots of customer complaints • Finally less customers => Frequency plan should be properly done



Solution finding – frequency optimization Basic Theory How to analyze my frequency plan • Frequency reuse distance • Frequency allocation reuse • Frequency load • Effective frequency load



Solution finding – frequency optimization Theory, Effective reuse • Since the frequency band is always limited, the frequencies have to be reused in the network.

• As the reuse distance becomes smaller, there are more frequencies available for each cell, so more capacity can be provided. • The effective reuse is essentially the same as the conventional frequency reuse distance.

Reff

N freqsTOT =

N TRXave

where: Reff = effective reuse NfreqsTOT = total number of used frequencies NTRXave = average number of TRXs in a cell

Note! The smaller the effective reuse, the higher the capacity in terms of the number of TCHs provided by one frequency in the network. NSN Internal Document 67 © Nokia Siemens Networks


Solution finding – frequency optimization Theory, Frequency Allocation Reuse RF Hopping only •

Frequency allocation reuse indicates how closely the frequencies are actually reused in a network.

FAR

N freqsTOT =

N freqs / MA

where: FAR = frequency allocation reuse NfreqsTOT = total number of used frequencies Nfreqs/MA = average number of frequencies in MA-lists

• •

It indicates the severity of a worst case C/I in the cell border. If the network doesn’t utilize fractional loading, the frequency allocation reuse is the same as the effective reuse.



Solution finding – frequency optimization Theory, Frequency load • The C/I is low when frequency collisions occur. – In order to guarantee an adequate quality, the collision probability has to be made low. – The collision probability depends on the load of the hopping frequencies called a frequency load.

• The frequency load describes the probability that a frequency channel is used for transmission at one cell at one time.

• The frequency load is a product of two other loads: – hard blocking load (the average busy hour TCH occupancy in most of the cases) – fractional load

L freq

=

LHW L frac ⋅

where: Lfreq = frequency load LHW = the busy hour average hard blocking load (see next slides) Lfrac = fractional load (see next slides) NSN Internal Document 69 © Nokia Siemens Networks


Solution finding – frequency optimization Theory, Hard blocking load • The hard blocking load is calculated as

L HW

T hopTCH =

N hopTCH

where: LHW = hard blocking load T hopTCH = average number of used TCHs in the busy hour N hopTCH = total number of TCHs in the hopping TRXs



Solution finding – frequency optimization Theory, Fractional load • Fractional load means that the cell has been allocated more frequencies than TRXs. This is only possible for RF hopping TRXs.

• The fractional load is very useful when the number of TRXs is low. By utilizing fractional load, it is possible to provide enough frequencies to hop over (to get FH gain) to even a cell with just one hopping TRX.

L frac

=

N TRX N freqs / cell

where: Lfrac = fractional load N TRX = number of TRXs in a cell N freqs/cell = number of frequencies allocated to a cell (MA-list length) NSN Internal Document 71 © Nokia Siemens Networks


Solution finding – frequency optimization Theory, Frequency Load,HW and fractional load 75 %

•“HW load”isis75% 75% •“HWload” •Fractional loadFL FLisis •Fractionalload 33TRX / 5 F = 0.6 = TRX / 5 F = 0.6 =60% 60% •“Frequency load”isis •“Frequencyload” HWL HWL**FL FL==45% 45%

25 %

TRX-1

BCCH

1

2

3

4

5

6

7

f1

TRX-2

0

1

2

3

4

5

6

7

f2, f3, f4, f5, f6

TRX-3

0

1

2

3

4

5

6

7

f3, f4, f5, f6, f2

TRX-4

0

1

2

3

4

5

6

7

f4, f5, f6, f2, f3

Active Activeslots slots


Empty Emptyslots slots


(E)GPRS is on the BCCH layer in this case

Solution finding – frequency optimization Theory, Effective Frequency Load • Coverage limited network has low EFL • Interference limited network has high EFL • It is calculated by the equation mentioned below:



Solution finding – frequency optimization Visual BCCH Plan Inspection

584 re-used close together

584 re-used close together

Heavy re-use of 582



Solution finding – frequency optimization BCCH Carrier Utilisation

60

Channel distribution

50 e 40 c n e r 30 r u c c 20 O

C arriers also used in MA List

10 0

3 4 5 6 7 8 9 0 1 2 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 1 1 1 1 1 1 1 2 2 2 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5



Solution finding – frequency optimization BSIC Utilization BSIC allocation distribution • As can be seen from the figure, amount of BSIC combinations are not spread smoothly • BSIC planning is not properly done => risks to double BCCH+BSIC combinations is increased

Note! Good radio network is based on properly done BSIC + BCCH Planning NSN Internal Document 76 © Nokia Siemens Networks


Solution finding – frequency optimization Same color = same frequency

Signal level is good, but quality is bad => Frequency plan is not properly done



Solution finding – frequency optimization Different frequency planning methods

Exercise How these planning methods differs from each other? • Based on Prediction tools – Planning Tools with propagation models

• Based on Interference matrix – Optimizer

• Mapinfo – Visualization

Any other methods? NSN Internal Document 78 © Nokia Siemens Networks


Solution finding – frequency optimization Frequency Hopping Network capacity and performance is typically limited by co-channel interference, multipath fading, delay spread and noise

Interference

Interference Diversity

F1

F3

F1

F1

Frequency Hopping benefits are based on: Interference Diversity • where interference is averaged over multiple frequencies Frequency Diversity • which reduces the needed fading margins

average

F2 F2 F3

F2 F 3

MS_1

Signal Level

MS_2

MS_3

Frequency Diversity F1 F2

Interference averaging and reduced immunity to signal fading gives the possibility to reduce the C/I margins and tighten the frequency reuse schemes


F3

MS Location

Distance

Bursts sent on frequency F2 may be degraded or lost, but the initial signal is still be reconstructed from the bursts on frequencies F1 and F3.


Solution finding – frequency optimization Frequency Hopping Baseband Hopping 0 TRX 1

1 2

7 Timeslot f1

BC CH

TRX 2

f2

TRX 3

f3

TRX 4

f4

HSN1 Timeslot 0 hops over TRXs 2-4 only BCCH does not hop HSN2 Timeslots 1-7 hop over all TRXs

RF Hopping 0 TRX 1

1

2

BC CH

7

Timeslot

f1 – no hopping

TRX 2 TRX 3 TRX 4 NSN Internal Document 80 © Nokia Siemens Networks

f2,f3..fn – hopping according mobile allocation list One hopping sequence number only Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

Solution finding – frequency optimization DFCA – (Dynamic frequency & channel allocation) Random FH TRX 1

BCCH

TRX 2

Random FH over a fixed frequency list

TRX 3 TRX 4

DFCA TRX 1 TRX 2 TRX 3 TRX 4


•• Loose Looseinterference interferencecontrol control •• Relies Relieson onrandom randomspreading spreadingofofthe the interference interference

BCCH Cyclic FH over individually selected frequency lists and MAIOs for each connection

C/I > C/I target

•• Accurate Accurateinterference interferencecontrol control(C/I (C/Iestimations) estimations) •• Each mostsuitable suitable Eachconnection connectionisisassigned assignedwith withthe themost radio channel (MA list, MAIO, TSL) radio channel (MA list, MAIO, TSL) Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

Frequency optimization– Consistency check Cells with co-channel and adj-channel frequencies

A

Cell A

-6dB -9dB

Cell B B

F R E Q=10

F R E Q=9,10,11

• Co-channel frequencies: HO is not possible in case of co-BCCH • Adj-channel frequencies: Ho is possible but might fail due to interference. • If Co / Adj-channel frequencies are existing, these must be removed



Frequency optimization– Consistency check Check BCCH-BSIC and BCCH reuse distance Small distances may be dangerous!

FRE Q = 234 BS IC = 42

FRE Q = 234

FR E Q = 234 BS IC = 42

R euse 4/12 NSN Internal Document 83 © Nokia Siemens Networks


Frequency optimization – Consistency check Other Checks • • • • •

Neighbors with co-channel frequencies Neighbors with adj-channel frequencies Neighbors of a same cell with co-BSIC, co-BCCH Neighbors from other Vendor Frequencies / co-BSICs near the country border ( if available)



Examples of wrong parameter set impact on network operation/performance – frequencies

1. Drop call rate increase after new frequency plan implementation •

Double BA List activated

2. Impossibility to unlock some BTS after a RF-frequency hopping implementation 3. Impossibility to unlock some BTS after a frequency retune • •

NON-EDGE TRX, with GTRX = Y TRX, with GTRX = Y, not attached to any EDAP pool



Example 1 Drop call rate increase after new frequency plan implementation: • In the meantime, measurementBcchAllocation has been changed to idle and MA list defined with old BCCH frequency band OLD MA List BCCH

TCH

TCH

BCCH

TCH

OLD FREQUENCY PLAN

NEW FREQUENCY PLAN

These BCCH frequencies will not be measured by old MA List These TCH frequencies will be wrongly measured



Example 2 Not possible to unlock some BTS after a RF-frequency hopping implementation:

• After frequency hopping retune, some BCCH frequency equal to TRX frequency of non-BCCH TRXs • After changing TRX frequency, verify BSIC and TSC (Training Sequence Code) is changed accordingly.



Example 3a Impossibility to unlock some BTS after a frequency retune: • NON-EDGE TRX, with GTRX = Y, in a cell with EGENA set to yes SEG EGENA = Y

B

BTS

(E)GPRS territory GTRX = Y GTRX = Y


SEG EDGE TRX(s) NONEDGE TRX(s)

EGENA = Y

B

(E)GPRS territory GTRX = Y GTRX = N


BTS EDGE TRX(s) NONEDGE TRX(s)

Example 3b Impossibility to unlock some BTS after a frequency retune: • TRX, with GTRX = Y, not attached to any EDAP pool in a cell with EGENA set to yes SEG EGENA = Y 1

B

2

(E)GPRS territory GTRX = Y GTRX = N

BTS EDGE TRX(s) NONEDGE TRX(s)

Add TRX-1 to EDAP pool



EDAP pool

Neighbor Optimization



Solution finding – neighbor optimization • Keep the neighbor list in order. – Note! If lots of overshooting => dominance areas are not clear => neighbor list are getting bigger. – Bad overshooting should be avoided – Missing neighbors – the worst situation

• Use the Handover Adjacency Statistics to identify and remove neighbors

– Bad overshooting cells can cause problems

• ISHO optimization Avoid: • neighbor definitions that are co-BCCH and co-BSIC or Adj-BCCH and co-BSIC that can lead to wrong neighbor reporting • One-way neighbors



Solution finding – neighbor optimization

GREEN = source RED BLUE

=neighbor cell = no neighbor

As can be seen, neighbors are not fully optimized. There are missing neighbors and on the other hand, some far away cells are neighbors NSN Internal Document 92 © Nokia Siemens Networks


Solution finding – neighbor optimization Overshooting cells => added to neighbor list. Neighbor planning is more difficult to do, if lots of overshooting cells

GREEN = source RED BLUE

=neighbor cell = no neighbor


In this area there might be big problems due to unoptimized neighbors Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

Solution finding – neighbor optimization Challenges • Sites by the sea • Dominance areas are reaching far away => neighbor cells can be far away SEA

• Indoor solutions • Signal levels from outdoor cells near windows can be high

• High buildings • Lots of cells can be heart near the window

• 3G cells • Limited amount of neighbors NSN Internal Document 94 © Nokia Siemens Networks


Only few neighbor cells => in the sea there will be problems

Solution finding – neighbor optimization Neighbor optimization based on statistics • HO attempts, blocking and fails can be analyzed as source or target cells • Neighbor removing can be done based on stats

In case Number of HO Attempts is very low to a certain cell, consider removing this cell from Adjacency list.



Neighbor Planning – Consistency check Non symmetrical adjacencies Find all non symmetrical adjacencies : • cell B is neighbour of Cell A • cell A is not neighbour of Cell B

Cell A

Cell B

• If there are non symmetrical adjacencies, these must be changed to symmetrical adjacencies NSN Internal Document 96 © Nokia Siemens Networks


Neighbor Planning– Consistency check Neighbours with co-channel and adj-channel frequencies or co-BSIC, co-BCCH Cell B f=10

f=10, BSIC=2,2

Cell A f=24

Cell C

f=9,10,11

f=10, BSIC=2,2

Neighbours with co-BSIC, co-BCCH: • Might generate ghost-access and HO drop calls. • High sdcch abis failure and high handover failure could be indicators for this case. NSN Internal Document 97 © Nokia Siemens Networks

f=24

Neighbours with co-channel or adjacent channel frequencies: •Will cause interference between neighbour cells


Solution finding – neighbor optimization Different methods Exercise How these planning methods differs from each other? 1.Add maximum nbr of neighbors and after 2 weeks remove extra neighbors based on HO statistic 2.First only nearest (dominance area) and if problems, add more neighbors

In neighbor optimization following selection methods can be used • Based on Optimizer • Based on Map • Based on OSS Statistics



Neighbor planning – Consistency check Other Checks • • • • •

Check cells in MSC vs cells in BSC Check external adjacencies in MSC Non symmetrical adjacencies Neighbors of a same cell with co-BSIC, co-BCCH 3G neighbors



Handover Optimization



Solution finding – HO optimization HO Strategies – Traffic handling (900-1800 layers)

PBGT, quality, level

1800 layer

Umbrella Traffic Reason IUO PBGT, quality, level

900 layer



Solution finding – HO optimization Optimizing HOs, Reduce unnecessary HOs old

• DL Level HO typically (-94dB) • Quality HO 3dB • HO Level margin 2dB => if Cell1 -98dBm => Cell2 >-96dBm => HO => might be HO fail if for example some imbalance

new ⇒ HO level margin => 24dBm (=disabled in this case) Quality HO 3dB (HO is done only if quality problems) ⇒ No HO, AMR is used => no drops and quality is OK ⇒

⇒

PBGT is still working. Value can be decreased if problems)

Unnecessary HO is prevented in this area

-100dBm Cell1

Cell2

route -100dBm



Solution finding – HO optimization Optimizing HOs, HO parameters • Umbrella ( AUCL parameter) • AMR FR/HR – Packing, unpacking parameters • Interference / quality HO – If interference level -75dB => interference problems in good signal level – Rest are quality HOs => bad quality can be also due decreased signal level. – This difference is good to know

• Level / PBGT – If margin is same as PBGT, the only difference is level, See example in previous slide where Level HO disabled. – It is useful to know if overlapping in good signal level and bad signal level  Can be noticed if overshooting  If no Level HOs => real coverage problems without overlapping NSN Internal Document 103 © Nokia Siemens Networks


Solution finding – HO optimization AMR + Power Control & HO control co-ordination in UL Quality (BER) (PxNx:6/16)

UURF

0

Packing of the call

Power down quality (PxNx:4/6)

IHRF

1

No action (PxNx: 2/3)

LURF

3

Power up quality IHRH

4

(PxNx: 4/6)

Unpacking the call QURF

) 1 / 1 : x N x P (

5

Handover trigger

5 9 r ) e C v o r O d e H ( n g g R a i r U H t L

-110dBm (AMR) Qual HO HYS -6dB NSN Internal Document 104 © Nokia Siemens Networks

) 1 / 1 : x N x P (

5 9 ) p C u O r P e l ( w e R o v e U P l L@ VF RoB Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010

) 1 / 1 : x N x P (

0 7 ) C O P ( R U U

RxLevel n w o d r e l w e o v e P l

-47dBm

HO Optimization – Consistency check Synchronized handover Non-synchronized Handover

– MS sends access bursts (HO_ACCESS) (with varying TA) until it receives PHYSICAL_INFO Synchronized Handover

– MS sends a few access bursts (HO_ACCESS) and then starts transmission with previous TA Non-synchronised handover leads to a longer communication interruption than synchronised handover (200ms vs. 100ms) Synchronized HO should be activated between sectors of the same site. If activated on inter-site adjacencies the handover can fail Synchronized HO

Non- Synchronized HO

Site A NSN Internal Document 105 © Nokia Siemens Networks

Site B


Solution finding – HO optimization Different methods Exercise 1 In which case the network is working better?

(theoretical)

• All handovers are quality handovers • All handovers are level handovers Exercise 2 How these differs from each other? • Quality HO vs. Interference HO • PBGT vs. Level HO • Traffic handling : decreased power 2 dB vs. decreased HO margin 2dB



HO planning – Consistency check Other Checks • • • • • •

Check cells in MSC vs cells in BSC Check external adjacencies in MSC Neighbors with co-channel frequencies Neighbors with adj-channel frequencies Neighbors of a same cell with co-BSIC, co-BCCH Check site with same MA List in different sectors and different HSN • Check site with same MA List in different sectors and MAIO collision



Examples of wrong parameter set impact on network operation/performance – Handover

1. No handover from a cell towards all its neighbours •

PLMN permitted = No

2. High Handover failures after implementation of new adjacency plan •

SYNC = YES

3. No handover happening from an interfered cell •

hoMarginQual set to 0

4. 100% of handover failures in an adjacency relation •

Co-BSIC co-BCCH declarati

5. High number of handovers •

hoThresholdsLevUL = hoThresholdsLevDL

6. Handover not happening when DL signal level of neighbour much greater than serving cell •

POC DL activated



Example 1 No handover from a cell towards all its neighbours: • PLMN permitted not set properly. MML Default value is the NCC of the BTS. So HOs will not happen towards neighbours with different NCC. No Measurement reports of Cell B are sent to BSC. So no HOs occur from Cell A to Cell B

Plmn permitted 0 = No …. Plmn permitted 3 = Yes Plmn permitted 4 = No …

Cell A NCC 3

Cell B NCC 4

Set Plmn permitted 4 = Yes

Plmn permitted 7 = No

PLMN permitted parameter consists actually of 8 parameters (0 …7) related to the NCC part of the BSICs of neighbour cells. MS only reports measurements of cells with NCC permitted (plmn permitted = YES).



Example 2 High Handover failures after implementation of new adjacency plan • All adjacencies have been implemented with synchronised parameter set to yes. HOs between cells of different sites will probably fail. Synchronized Handover MS

NETWORK ACTIVE CALL HANDO CMD HANDO ACC

x

PHYS INFO

Old Channel, Old Cell

Set SYNC = YES only between sectors of the same site. SYNC = NO

New Channel, New Cell

HANDO COM ACTIVE CALL


Recommendation is to have Sync HO within BTSs in the same BCF and Non Sync HO between BTSs in different BCFs. Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

SYNC = YE

Example 3 No handover happening from an interfered cell • hoMarginQual set to 0 A

Ho Margin Qual = -4

PBGT margin = 6dB

B Ho Margin Qual = 0 (MML Default)

In an interfered cell, despite high signal strength, quality is not good. So HO Margin Qual should permit HO to a cell that despite it’s lower signal strength, may have a better quality. NSN Internal Document 111 © Nokia Siemens Networks


Example 4 100% of handover failures in an adjacency relation • Co-BCCH declaration (Co-BSIC)

Cell A


 Cell B

FREQ=10

FREQ=10

BS IC = 33

BS IC = 33


Example 5 High number of handovers. hoThresholdsLevUL = hoThresholdsLevDL Quality

0

1

hoThresholdsLevUL , high thresholds will anticipate HO s: -95 -> -100

2

Level Ho

No Action Needed

3

4

5

hoThresholdsLevDL : -95 -> -90

Quality Ho

Interference Ho

6

7 0 8 6 4 1 0 0 0 1 1 1 1 -

2 0 8 6 4 2 0 0 9 9 9 9 1 1 -

0 8 6 4 9 8 8 8 -

2 0 8 8 8 7 -

6 4 7 7 -

2 0 8 6 4 2 0 8 7 7 6 6 6 6 6 5 -

6 4 2 0 5 5 5 5 -

HoThresholdInterferenceUL/DL HoThresholdLevUL


HoThresholdLevDL

HoThresholdLevUL/ DL Advance Network Optimization/JP/NNPO/ 1st to 3rd July 2010 @ VF RoB

dBm

Example 6 Handover not happening when DL signal level of neighbour much greater than serving cell • Power control DL activated

A Power Control -6dB B -3dB



-9dB

(E)GPRS Optimization



Objectives After PS optimization module learning • You know the impact of GSM performance on (E)GPRS performance • You know the main assessment activities • You know how the signaling, throughput and mobility can be optimized



GSM Network as the physical layer of (E)GPRS The optimal GSM network from PSW services point of view has: 1. As high signal level as possible 

It means that even the indoor signal level should be high enough to have MCS9 for getting the highest data rate on RLC/MAC layer.

2. As low interference as possible 

The aim of having high C/I is to avoid throughput reduction based on interference.

3. Enough capacity 

Enough BSS hardware capacity (interface and connectivity) is needed to provide the required capacity for PSW services in time. Both CSW and PSW traffic management should be harmonized with the layer structure and long term plans.

4. As few cell re-selection as possible  

The dominant cell coverage is important to avoid unnecessary cell-reselections in mobility. The prudent PCU allocation can help to reduce the inter PCU cell reselections. Dominant cell structure can help to maximize the signal level and reduce the interference, too.

5. Features 

All the features should be used which can improve the PSW service coverage, capacity and quality in general.

6. The GSM network is the physical layer of (E)GPRS, so the optimization of GSM network can improve the performance of (E)GPRS, too.



(E)GPRS Optimization– TSL data rate optimization Timeslot optimization is based on basic optimization – Interference optimization – Coverage optimization RLC/MAC Data Rate (FTP Download on 2 TSLs)

Data rate is heavily depending on network quality • Better quality Better throughput ⇒ Less blocking ⇒

120 100 80 s p b k

No Interference C/I 25 dB

60

C/I 20 dB C/I 15 dB

40 20 0 -65

-70

-75

-80

-85

-90

Signal level (dBm) NSN Internal Document 118 © Nokia Siemens Networks


-95

-100

-105

(E)GPRS Optimization – Network Element and Configuration Assessment HLR

BSS

BTS • GPRS territory

• PCU variant & dimensioning

• BTS HW considerations (TRX & BB-card)

• PCU strategy in mixed configuration

• BTS SW (EPCR)

• QoS profile TC

TCSM

• BSS SW and features

HLR/ AC/ EIR MSC/VLR

• GPRS settings

SGSN • Unit capacity (PAPU etc.) • BSS Gb Flow control

MS/Client parameters Gs

• GPRS/EDGE capability and release

IP/MPLS/IPoATM backbone GGSN

•Multislot support RF

BTS

BSC Abis

Gb

2G SGSN

Gn Gi Application Applicatio Servers n Servers

RF interface • Coverage Gb interface

•C/I • Capacity • Traffic volume • Mobility

Abis interface • EDAP size / dimensioning

• Bearer size • IP v.s. FR • Dimensioning

• # of E1/T1s • GPRS/EDGE traffic NSN Internal Document 119 © Nokia Siemens Networks

(colocated


Server • load • settings (Linux/Win)

(E)GPRS Optimization - Introduction (E)GPRS Network Optimization

• Signaling capacity & resource allocation improvement • Data Rate – Connectivity Capacity (MS-SGSN) – Multiplexing and multislot usage maximization

• Mobility improvement



(E)GPRS Optimization - Signaling Capacity & Resource Allocation improvement • Signaling – PCH, AGCH, RACH and SDCCH (NMOII)

• Resource Allocation – Cell (re)-selection – BTS selection – Scheduling



(E)GPRS Optimization - RF Signaling Paging Measurements (NMO I) Traffic Volume • •

cs_paging_msg_sent (c3058) (CS pagings from Gb) ps_paging_msg_sent (c3057) (PS pagings from Gb)

Congestion (CS + PS) •

max_paging_gb_buf (003050)

Paging success ratio on CS •

PAC_PAG_REQ_FOR_CS_PAG (c72083) / (cs_paging_msg_sent) (c3000)

Success ratio of Paging on Gs interface (2G SGSN) sum(DL_MESSAGES_DISCARDED_IN_GS(11000)) Sgsn_961a = ----------------------------------------------------------------------- * 100 sum(CS_PAGING_MSGS + DL_TOM_MSGS)

Solution for reducing PCH rejection and load • • •

Usage of combined structure, modifying MFR and PER parameters LA/RA re-planning Cell splitting



(E)GPRS Optimization - RF Signaling AGCH Measurements Immediate Assignment Traffic Volume (with Rejection) • imm_assgn_sent (c3001) - Imm Assign) • imm_assgn_rej (c3002) - Imm Assign Rejected

• packet_immed_ass_msg (c72084) - P-Imm Assign • packet_immed_ass_rej_msg (c72087) - P-Imm Assign Congestion blck_21b =

packet_immed_ass_nack_msg ------------------------------------------------------------------------packet_immed_ass_nack_msg + packet_immed_ass_ack_msg

Solution for reducing AGCH rejection and load • Usage of combined structure, modifying AG and CALC parameters • Immediate Assignment messages are shared between PCH and AGCH • PBCCH implementation (in case of high (E)GPRS signaling traffic)



(E)GPRS Optimization - RF Signaling RACH Measurements Traffic Volume PACKET_CH_REQ (c072082) (PSW) CH_REQ_MSG_REC (c003004) (CSW)

Load RACH_4 = 100 * avg(ave_rach_busy(C3014)/res_acc_denom3(c3015)) avg(ave_rach_slot(c3006)/res_acc_denom1(c3007))

Repetitions of PS channel requests (load and quality) RACH_9 = UL_TBF_WITH_RETRY_BIT_SET (c072020) / PACKET_CH_REQ (c072082 )

Solution for reducing RACH rejection and load • Usage of non combined structure, modifying RET parameter • PBCCH implementation (in case of high (E)GPRS signaling traffic) • Reducing high UL interference (and DL interference if the repetition is too high) NSN Internal Document 124 © Nokia Siemens Networks


(E)GPRS Optimization - RF Signaling SDCCH Measurements (NMO II with LA Update) Traffic Volume – SDCCH seizure attempts (c1000) – Average available SDCCH (ava_45a)

Congestion – Blocking on SDCCH, before FCS (blck_5) – Time congestion on SDCCH (cngt_2)

Solution for reducing SDCCH load – Increase of Periodic RA update timer (PRAU) / MS Reachable timer (MSRT) 

– – – –

Drawback: more PAPU capacity is needed and more paging will be generated, if the MS is out of service

More SDCCH TSL allocation and/or Dynamic SDCCH feature usage Combined RAU (NMO-I with Gs for (E)GPRS) (Resume feature decreases the amount of RAUs) LA and border re-planning



(E)GPRS Optimization - Resource Allocation Introduction Basic Allocation related Topics • PSW Activation • Territory settings • Channel pref. Cell-(re)selection • C1, C2 • C31/C32 • NCCR

RLC/MAC

Provide enough capacity to PSW traffic in general (find balance between CSW and PSW)

Allocate the traffic to the most appropriate resource

BTS Selection • MultiBCF and CBCCH • PCU algorithm

Separate GPRS and EGPRS and share the resources

User Channel Scheduling • Priority based QoS


Service and user prioritization


(E)GPRS Optimization– GPRS Territory • Circuit Switched traffic has priority outside dedicated territory • GPRS dedicated time slots (% of total cell capacity) can be defined. Only (E)GPRS can use, no CSW

• Dedicated TSL is subset of Default TSL • Territories consists of consecutive timeslots • GPRS can be set to favour the BCCH Transceiver -> minimum interference Default GPRS capacity threshold

Extra GPRS capacity TRX 1

BCCH

TS

TS

TS

TS

TS

TS

TS

TRX 2

TS

TS

TS

TS

TS

TS

TS

TS

Free time slots in Circuit Switched territory Territory upgrade in interval of Territory Upgrade Guard Time


Default GPRS Capacity (%) Territory downgrade forced by the Circuit Switched traffic


Dedicated GPRS Capacity (%)

Circuit Switched Territory Circuit / Packet Switched Territory

(E)GPRS – Cell (re) -Optimization selection Cell (re) - Selection



(E)GPRS Optimization – BTS Selection and TSL Allocation BTS selection and TSL allocation is Segment Initial BTS Selection Reallocation of TBFs among the BTS 1. BTS Load reallocation 2. Uplink Rx level reallocation

3. Downlink Rx level reallocation 4. Downlink RX level received first time reallocation



(E)GPRS Optimization – Resource Allocation - Priority based QoS TBF1 with SSS=6 TBF2 with SSS=1

1

2 127 12 7

8

Latest Service Time = Current Time + Scheduling Step Size

3

4

5

6

12

12

12

12

11

12

9

10

7

8

13 12

18 13

10 11 (time) 12

9 18 14

18 15

18 18 16

17

Latest service time

t

i 52 TDMA frames = 240 ms= 12 blocks



t

The scheduling is done based on latest service time, one TBF at a time is served by the RTSL

i

(E)GPRS - Throughput optimization • Connectivity Capacity (MS-SGSN) • TSL data rate improvement and multislot usage maximization (BSS) • E2E data rate (applications)



(E)GPRS Optimization Connectivity Planning – Maximized Capacity • The connectivity optimization for maximum capacity is based on the proper set of CDEF and DAP size. • To provide enough capacity for territory upgrade the 75 % utilization in the connectivity limits is recommended by NSN • PCU Connectivity capacity limits can be seen below Outputs Abis channles (radio TSLs) EDAP pools BTS (cell, segment) TRXs

Max limit* 256 16 64 128

Utilization 75% 75% 75% 75%

*PCU & PCU-S handle 128 radio TSLs only with S11.5 *PBCCH is not implemented



Limit 192 12 48 96

unit TSLs pcs pcs pcs

(E)GPRS Optimization Connectivity in different PCUs PCU variant PCU

PCU-S

PCU-T

PCU2-U

PCU-B

PCU2-D


BSS11

BSS11.5 ownwards

BTS

64

64

TRX Radio TSLs Abis 16 kbps channels

128 256 256

128 128 256

Gb 64 kbps channels BTS

31 64

31 64

TRX Radio TSLs Abis 16 kbps channels Gb 64 kbps channels

128 256 256 31

128 128 256 31

BTS TRX Radio TSLs

64 128 256

64 128 256

Abis 16 kbps channels Gb 64 kbps channels

256 31

256 31

BTS TRX Radio TSLs

N/A N/A N/A

128 256 256

Abis 16 kbps channels Gb 64 kbps channels

N/A N/A

256 31

BTS TRX

2 x 64 2 x 128

2 x 64 2 x 128

Radio TSLs Abis 16 kbps channels Gb 64 kbps channels BTS TRX Radio TSLs Abis 16 kbps channels Gb 64 kbps channels

2 x 256 2 x 256 2 x 31 N/A N/A N/A N/A N/A

2 x 256 2 x 256 2 x 31 2 x 128 2 x 256 2 x 256 2 x 256 2 x 31

BSC Type BSCE, BSC2, BSCi, BSC2i

BSCE, BSC2, BSCi, BSC2i



BSC3i

BSC3i


(E)GPRS Optimization Connectivity Planning – Cells / PCU The following configuration types were defined:

The following table shows the number of EGPRS cell / PCU calculations for the different configuration types. EGPRS BSS Configuration Portfolio (ETSI)

extra small (XS)

Extra small Small Medium

(CDEF (CDEF (CDEF

Large Extra large Data monster

(CDEF (CDEF (CDEF

small (S)

lowest cost per EGPRS cell: maximize low cost: high EGPRS cells EGPRS cells per per PCU PCU ratio

Use case

Parameters CDEF DAP

medium (M)

1 RTSL, DAP 2 RTSL, DAP 4 RTSL, DAP

→

→

3 TSL) 4 TSL) 6 TSL)

→

→

→

→

→ → →

large (L)

High data volume Basic EGPRS site EGPRS site

4 RTSL, DAP 4 RTSL, DAP 4 RTSL, DAP extra large (XL)

Extra high data volume site

→ → →

8 TSL) 11 TSL) 16 TSL) data monster monster Data (DM) (DM)

Data hot spot site)

Data hot spot site

1 3

2 4

4 6

4 8

4 11

4 16

# of cells / PCU with utilization cells per DAP #DAPs per PCU #EGPRS cells per PCU PCU Abis ch utilization average cells per DAP #DAPs per PCU #EGPRS cells per PCU PCU Abis ch utilization cells per DAP #DAPs per PCU #EGPRS cells per PCU PCU Abis ch utilization

2 16 32 88% 2.5 15 37 85% 3 15 45 88%

2 11 22 86% 2.5 10 25 82% 3 10 30 86%

2 7 14 88% 2.5 6 15 80% 3 6 18 84%

2 5 10 78% 2.5 5 12 81% 3 5 15 86%

2 4 8 81% 2.5 4 10 84% 3 4 12 88%

2 3 6 84% 2.5 3 7 86% 3 3 9 89%

Performance figures (kbps) (n+n+n) single user peak RLC/MAC (#RTSL in DL) cell peak RLC/MAC (gross)

186 354

237 325

237 597

237 748

237 829

NSN Internal Document 134 © Nokia Siemens Networks notes fo applying

237 373 areas for medium area for low areas for low+ EGPRS traffic, EGPRS traffic, EGPRS traffic, supports 4 & 5 no support 4 RTSL RTSL MS st to 3rd July 2010 @ VF RoB Advance Network Optimization/JP/NNPO/ 1 performance MS FTP max FTP/HTTP max commitments throughput throughput

areas for high EGPRS traffic

Hot spot site

areas for extra hot spot site for For EGPRS high EGPRS traffic EGPRS

(E)GPRS Optimization - KPIs • DL MCS selection limited by EDAP (dap_7a) – Similar formula for DL (dap_8c)

Target values: Good: < 75 min/GByte Bad: > 150 min/GByte

– DL_TBFS_WITH_INADEQ_EDAP_RES (c076008)  c76008 includes lack of PCU resources (c76020

DL_MCS_LIMITED_BY_PCU ) as one reason.  If c76008 is updated but peak EDAP usage is less than 100%, reason is that c76008 has been updated because of lacking PCU resources,



(E)GPRS Optimization - KPIs • DL MCS selection limited by PCU (dap_9) – Similar formula for UL ( dap_10)

Target values: Good: < 15 min/GByte Bad: > 30 min/GByte



(E)GPRS Optimization - Multiplexing Channel Allocation Algorithm tends to separate EDGE TBFs and GPRS TBFs on different RTSL to avoid multiplexing, if only one PS Territory exists in the cell or there is high load.

• UL GPRS USF on DL EGPRS TBF • TSL sharing - GPRS/EGPRS TBFs’ multiplexing on a TSL The algorithm checks the need for re-allocation every TBF_LOAD_GUARD_THRSHLD, in order to separate sessions. The max amount of TBFs per TSL can be limited:

Parameter Name

Abbreviation

Description

Maximum Number of DL TBF

MNDL

maximum number of TBFs that a radio time slot can have in average, in a GPRS territory, in the downlink direction.

1..9, default:9

Maximum Number of UL TBF

MNUL

maximum number of TBFs that a radio time slot can have in average, in a GPRS territory, in the uplink direction.

1..7, default:7



Range and Default

(E)GPRS Optimization – Multiplexing – Measurements (KPIs) Amount of TBFs / TSL Uplink TBFs pr timeslot tbf_37d Downlink TBFs pr timeslot tbf_38d GPRS TBF multiplexed with EGPRS TBF 8PSK coding scheme downgrade due to GPRS multiplexing rlc_61 DL EDGE TBFs in GPRS territory tbf_60 UL EDGE TBFs in GPRS territory tbf_59 Ratio of DL GPRS TBFs in EDGE territory tbf_58a Ratio of UL GPRS TBFs in EDGE territory tbf_57a



(E)GPRS Optimization - Multislot usage Territory Downgrade and Upgrade The territory downgrade heavily depends on the size of dedicated and default territory.

•

Territory downgrade due to CSW traffic rise •

•

Territory downgrade due to less PSW traffic •

•

Downgrade request below Default territory because of rising CSW (c1179) Downgrade request back to the Default territory when there is no need for additional channels anymore (c1181)

Territory upgrade request rejection beyond default territory •

Upgrade request beyond Default territory for additional resources (c1174), which can be rejected because of: 1. 2. 3.

No (E)GPRS capable resource left (No (E)GPRS enable TRX or the maximum (E)GPRS capacity reached) PCU and EDAP capacity limitation (256 Abis TSL per PCU) High CSW load Default territory

0

1

2

3

1174

4

5 1179

1181 1180



6

7

(E)GPRS Optimization Free TSL Size (after CS Upgrade and Downgrade) When a downgrade or upgrade procedure is requested following parameters can reduce or increase the border between CSW and PSW territories:

TSL number after CS downgrade TRX number

free TSL for CS downgrade (%) (CSD)

70 95 99

1

2

3

4

5

0

0

0

1

1

1

1

1

2

2

1

1

2

2

3

1

2

3

4

5

0

1

1

1

2

1

2

2

3

4

1

2

3

4

5

2

3

4

5

6

TSL number after CS upgrade TRX number

free TSL for CS upgrade (sec) (CSU)


1 4 7 10


(E)GPRS Optimization Multislot Usage – Measurements (KPIs) Actual Territory • ava_44 • Peak PS territory (c2063) Recommendation: Ava_44 and c2063 can be compared with the CDEF settings. If too big difference, then CDEF should perhaps be changed, or more capacity should be added to the cell.

Multislot Blocking • UL / DL multislot allocation blocking – hard (tbf_15a, tbf_16a) • DL multislot blocking – soft (blck_33a) Recommendation: Too much multislot blocking shows that the territory is perhaps not enough. And it is also interesting to see how many timeslots which are requested (helpful in determining the size of the default territories) • Requested timeslots for GPRS TBFs c72039-c72149, c72040-c72150, c72041c72151, c72042-c72152 • Requested timeslots for EDGE TBFs c72149, c72150, c72151, c72152 NSN Internal Document 141 © Nokia Siemens Networks


(E)GPRS Optimization - Mobility Optimization The aim of mobility optimization is to reduce the cell outage time during cell re-selection. Cell outage can be reduced by

• Providing enough signaling capacity for cell re-selection (the RACH, PCH, AGCH and SDCCH channel are not limiting the signaling flow)

• Rebalancing BCFs among PCUs properly (the important neighbors are allocated to the same PCU)

• Reallocating LA/RA borders properly • Enabling Network Assisted Cell Change (NACC) feature



(E)GPRS Optimization – Outage Definition Used in Measurements Three delays can be calculated from logs: GERAN

MS

Cell outage: • In one-phase access: the time

between the last EGPRS Packet Downlink Ack/Nack message and the first Packet Uplink Ack/Nack. • In two-phase access: the time between the last EGPRS Packet Downlink Ack/Nack message and the first Packet Uplink Assignment.

Data outage:

Las t IP packets E G PR S Packet Downlink Ack/Nack R outing Area Update R equest

Data Outage

Packet Uplink Assignment R outing Area Update A ccept R outing Area U pdate C omplete

the time between the last and the first EGPRS Packet Downlink Ack/Nack message.

E G PR S Pac ket Downlink Ack/Nack

Application outage: the time between the last and the first successfully received FTPpacket. NSN Internal Document 143 © Nokia Siemens Networks

Application Outage


F irst IP packets

Cell Outage

(E)GPRS Optimization Network Assisted Cell Change (NACC) NACC is in Rel 4 of 3GPP GERAN , mandatory for R4 mobiles. Nokia’s S11.5 implementation is based on Rel4. Both, autonomous and network controlled cell reselections are supported. Support is for intra-BSC cell changes (Support of inter-BSC NACC specification is available in Rel 5) NACC support is for MSs in RR Packet Transfer Mode only

NACC

shortens the cell reselection in two ways:

•

Sending neighbour cell system information on PACCH to MS in packet transfer mode while it is camped on the serving cell

•

By supporting PACKET SI STATUS procedure in a target cell



(E)GPRS Optimization– CS Traffic vs. PS Traffic

CS and PS peak values at the same time (BSC level data) Bad for PS timeslot allocation ⇒ Lots of downgrading ⇒

Exercise How to optimize PS performance on area level?



Alarms



Alarms - Introduction Alarm analysis is the 1st thing to be done in worst cell troubleshooting! No use optimizing parameters if you have a HW problem! Bad performance of a cell may be caused by faulty equipment. • Check that there are no performance-affecting alarms in the cell monitored and

also in neighbour cells! • Alarms are usually transferred to the NetAct database. • Monitoring can be done through MML Commands, Nokia NetAct Top Level User Interface

Note! Constant monitoring is needed in order to avoid critical alarms in any network element.



Alarms – Groups

Alarm number in:

Notices (NOTICE)

Disturbance Failure printouts printouts (DISTUR) (ALARMS)

switching equipment 0–799

1000–1799

O&M equipment

800 - 899

1800–1899

900 - 999

1900–1999

transmission equipment diagnosis report number base station/ transmission equipment alarms

Diagnosis reports (DIAGN)

Base station alarms

2000–2799 3000–3799 2800–2899 3800–3899 2900–2999 3900–3999

Numbers Transmission reserved for equipment possible alarms external alarms

4000–4799

4800–4899

4900–4999

3700–3999 7000–7999

8000–8999

power equipment

5000–5499

external equipment

5500–5999



Alarms – Printout Fields • Alarm Number • BCF number, BTS number, TRX number, Alarm Object, Unit, Date, Time, Alarm Number • Urgency level

Note:

* Low Priority ** Med Priority

• Printout type

*** High Priority

ALARM fault situation CANCEL fault terminated DISTUR disturbance NOTICE notice

The urgency level is output in all alarm printouts except notices (NOTICE). The urgency levels of terminated alarms are indicated by dots (.) instead of asterisks (*).

• Event type COMM communication failure QUAL quality of service PROCES processing failure EQUIPM equipment failure ENVIR environmental failure NSN Internal Document 149 © Nokia Siemens Networks


Alarms - Alarms in MML Alarms in BSC Level: – ZAHO: PRINT ALARMS CURRENTLY ON – ZAHP: PRINT ALARM HISTORY

Alarms in BTS Level: – ZEOL: LIST ALARMS CURRENTLY ON – ZEOH: LIST ALARM HISTORY

Verify if BCF, SEG, BTS, TRX and RTSL are LOCKED or UNLOCKED; WO (working), BL- USR (blocked by user) or Restarting.

BCF, SEG or BTS configuration and status – ZEEI: OUTPUT RADIO NETWORK CONFIGURATION

TRX and RTSL configuration and status – ZERO: OUTPUT TRANSCEIVER PARAMETERS



Alarms - Alarms in MML HW Tests: • ZUBK: HANDLE ABIS LOOP TEST (Parameters: BTS, TRX, RTSL, Fixed or Dynamic Abis connection and Abis TSL and Sub-TSL, looping time)

• ZUBS: START TRANSCEIVER TEST (Parameters: BTS, TRX, RTSL, test mode, RTSL,test selection, diversity path selection,test connection, RF test signal attenuation,BTS RX level,STM antenna attenuation,BS TX power attenuation, loop duration)



Alarms - Examples • 2993: BTS AND TC UNSYNCHRONIZATION CLEAR CALLS ON ABIS INTERFACE – Transcoder and transmission alarm. Abis test is needed.

• 7045: TRX/FU DATA TRANSFER ERROR BETWEEN FU AND CU16 – Base Station alarm. Observed when TCH failure rate is very high. A possible solution can be to change the TRX.

• Example of a BTS alarm printout:



Alarms - Top Level User interface • The Top-level User Interface contains graphical views of the network, in which network elements are represented hierarchically with symbols

• One of the main functions of the Top-level User Interface in network monitoring is to show the alarm situation in all managed objects



Implementation / Documentation



Implementation / Documentation • Optimization process is an iterative process, so there is no need to create troublesome final report after every circle. Main items must be still written down after every circle.

• All the activities/ findings should be listed – What should be done and when – “recommendation for the changes to be done" and "working orders".

• All improvements should be shown – Implementation time should be shown with improvements


Network Report Tasks And Results


Solution Verification



Solution Verification • The purpose is to check was solution succeed – Is the trend positive or negative after aimplementations – Check that collected data is valid EGPRS RLC Throughput kbit/s/TSL BSC KRASNODAR 60

50

• Parameter changes were done 20.11

40

s / t i b k

• Throughput after changes was improved

30

20

DL 10

UL

0 .

.

.

.

.

.

.

.

0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 1 NSN Internal Document 9 . 2 1 . 2 3 . 2 5 . 2 7 . 2 9 . 3 1 . 2 . 1 157 © Nokia Siemens Networks

1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 1 . 1 2 . 1 2 . 1 4 . 6 . 8 . 1 0 . 1 2 . 1 4 . 1 6 . 1 8 . 2 0 . 2 2 . 2 4 . 2 6 . 2 8 . 3 0 . 2 . 4 .


Monitoring



Monitoring – Introduction The basic idea to monitor network is to see how the optimization project is ongoing and how the network is performing • KPI monitoring • Testing & post processing • Alarm monitoring • Schedule monitoring • Investments checking • Resource monitoring



Monitoring – Reporting Suite

• Before Reporting Suite can be started, connection to NetAct must be done • Reporting Suite can be used for KPI monitoring



Monitoring – Reporting Suite

• Different KPI group can be seen here • Example of final KPI report



Tools to be used



Tools to be used

Example of tools to be used in Solution finding Statistics (Reporting Suite, Metrica etc.)

Performance optimization Traffic and Capacity optimization Coverage Optimization Frequency Optimization Neighbor Optimization HO Optimization (E)GPRS Optimization


x x x x x x x

Planning tools (NetAct Planner, Asset etc)

Configuration tools (Plan Editor, RAC tools, etc)

x x x x x


Optimizer

Drive test tools (Nemo Outdoor. TEMS, Actix etc)

x

x

x x x x x

x x x x x

x x x x x

Part3_Advance Network Optimization_Solution Finding

Recommend Documents