less samples cause a low 3g to 2g ho success rate

Upload: degdag-abdulrahman

Post on 14-Apr-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    1/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 1

    Less Samples Cause a Low 3G to 2G HO Success Rate

    1 Introduction

    2G3G interoperability is a common and frequent problem in 3G optimization work.

    There already have many ways to optimize this problem. This document shows you a

    different situation that the current 2G network cannot reselect to 3G then find a new

    way by analyzing user behavior and Mathematical deduction and bring up a new index

    Inter-RAT HO to Traffic Ratio to solve this problem at last.

    2 Description

    3G Operator network has been put into commercial. The Inter-RAT handover succe

    ss rate of is Unstable, the value is between 60% to 85%. Both of the two RNCs are t

    he same and all the parameters are correct.

    After continually optimization work, the Inter-RAT handover success rate grow up t

    o 80%-90% but still unstable seems cannot be improved anymore.

    3 Analysis

    The NUM of Inter-RAT handover is so little that some accidental failures will have

    great effect on the KPI. At the same time we did DT for the Inter-RAT handover, the

    detail information about the DT as below:

    1>3G to 2G handover test

    Date : 28-Apr afternoon.

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    2/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 2

    Equipment : ZTE T7(test phone) & Nokia E62(compare to each other).

    Method :

    First, we let the two phones under 3G service.

    Then, select the phone to auto network selection mode.

    Then, one phone call the other and start to drive out of Bulawayo in four direction. ke

    ep the connection until it handover to 2G or drop.

    Result:

    All the road out of Bulawayo have handover to 2G successfully.

    Time

    Source 3G cell Target 2G cell

    CID,PSC

    CELL

    CID

    CELL

    15:37:54

    12042,8

    4

    Belmont PO_

    2

    320

    07 ILANDA3

    15:56:07

    12071,1

    01 Montrose_1

    320

    02 MONTROSE3

    16:16:24

    12171,1

    70

    Nkulumane_

    1

    106

    14 NKULUMANE1

    16:43:02

    12141,1

    24

    Entumbane_

    1

    106

    13 LOBENGULA1

    17:02:50

    12081,1

    02 Ilanda_1

    202

    40

    KHUMALONRT2

    Location

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    3/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 3

    2G to 3G reselection test

    Date : 28-Apr afternoon.

    Equipment : ZTE T7(test phone) & Nokia E62(compare to each other).

    Method :

    First, we let the two phones under 2G service.

    Then, select the phone to auto network selection mode.

    Then, keep two phones in idle mode and drive around the dense urban of Bulawayo, w

    here the 3G signal is very good.

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    4/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 4

    Result: We drive about two hours, but the two phones never reselect to 3G.

    Conclusion:

    1>The function and parameter settings of 3G to 2G handover is OK.

    2>2G cannot reselect to 3G, where it should do.

    (usually, where if 3G Ec/Io>-12db, UE should reselect to 3G automatically)

    The test result shows the 3G to 2G handover is OK, but 2G cannot reselect to 3G.

    2>User behavior analysis and Mathematical deduction

    After Communication we find Operator forbid 2G reselect to 3G, because they are

    afraid the poor 3G coverage will cause a lot of complains and they think this 2G strate

    gy will have no effect on 3G inter-RAT HO success rate, so they insist 3G part to solv

    e the problem alone.

    According to this situation, I decide to do a theoretical analysis and send this repo

    rt in written form then push Operator to solve. The user behavior analysis and Mathem

    atical deduction as below:

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    5/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 5

    figure2 Mathematical deduction about the less samples cause a low 3G to 2G HO

    Success Rate

    Explain: For the succeed ones, they will stay in 2G and will never come back in c

    ase they restart the phone or reselect to 3G manually. I have asked many people that

    very few of them do like this, so that the Success NUM will decrease a lot and the To

    tal NUM will decrease as well.

    For the Failure ones, they will stay in 3G and keep on making failures. So the Fail

    NUM will keep more or less the same.

    So ,the situation will become:

    The 3G->2G HO success rate will surely decrease a lot.

    3>Bring up a new index for further analysis

    From the analysis above, we can see the total samples will decrease a lot, but ho

    w much exactly? We need a reliable estimate . How to estimate? The best way is refer

    to the normal commercial network.

    Here I bring up a new index called Inter-RAT HO to Traffic Ratio . The formul

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    6/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 6

    a is: Inter-RAT HO to Traffic Ratio= NUM of Inter-RAT HO/Traffic. This index r

    eflect how many Inter-RAT HO will happened while one Erl traffic was come out(or one

    MB). Its easy to understand that if the 3G coverage is good this value will be smaller,

    otherwise this value will be bigger, theres a case need to pay attention , this index s

    hould calculate separately by CS & PS service.

    Now I use the existing date to calculate the value:

    1>Operator A 3G network Inter-RAT HO to Traffic Ratio

    Begin time

    Inter-RAT HO to T

    raffic Ratio (CS,Er

    l)

    Inter-RAT HO to T

    raffic Ratio(PS,M

    B)

    2011-01-10 18.4057971 1.83565972

    2>Operator B Inter-RAT HO to Traffic Ratio

    Begin time

    Inter-RAT HO to T

    raffic Ratio(CS,Erl)

    Inter-RAT HO to T

    raffic Ratio(PS,M

    B)

    2011-09-10 5.422366819 1.329927001

    2011-09-11 6.829227666 1.732457781

    2011-09-12 6.077752884 1.683087374

    2011-09-13 6.129878094 1.73063497

    2011-09-14 6.039713496 1.926980134

    2011-09-15 6.493291552 2.077293214

    2011-09-16 6.444757408 1.997815984

    3>Current Operator 3G network Inter-RAT HO to Traffic Ratio

    Begin ti

    me

    Number

    of attem

    pted out

    going CS

    inter-RA

    T hando

    vers

    Cell T

    raffic

    Volu

    me, C

    S(Erl)

    Inter-RAT

    HO to Tr

    affic Ratio

    (CS,Erl)

    Numb

    er of

    attem

    pted

    outgoi

    ng PS

    inter

    -RAT

    hando

    vers

    Total PS

    Traffic MB

    Inter-RAT

    HO to Tr

    affic Ratio

    (PS,MB )

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    7/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 7

    2012-05-

    14454

    499.6

    608

    0.90861640

    51163

    10984.809

    24

    0.10587348

    2

    2012-05-

    15446

    486.3

    928

    0.91695436

    31231

    10330.947

    0.11915655

    2012-05-

    16518

    491.0

    356

    1.05491333

    1222

    10020.576

    44

    0.12194907

    2

    2012-05-

    17434

    500.2

    817

    0.86751124

    41288

    10412.226

    78

    0.12370072

    5

    2012-05-

    18465

    475.5

    364

    0.97784312

    61388

    9370.5292

    16

    0.14812397

    1

    2012-05-

    19750

    515.2

    625

    1.45556876

    31683

    10472.070

    1

    0.16071321

    2012-05-

    20 476

    507.8

    369

    0.93730880

    9 1377

    10478.549

    78

    0.13141131

    4

    Because the radio environment of Operator A have a great difference with Current

    Operator So the Operator B date is more useful to reference but the coverage of Ope

    rator B is much more better than Current Operator, according to the formula the value

    of Current Operator will be bigger than Operator B

    Conservative estimate:Inter-RAT HO to Traffic Ratio(CS,Erl)>5,Inter-RAT HO

    to Traffic Ratio(PS,MB )>1,

    That is: the NUM of Inter-RAT CS HO should >2500, the NUM of Inter-RAT PS HO

    should >10000; So, the Inter-RAT HO success rate should more than 95%.

    S olution and Result

    According to the plan, on 14:00,6JUN Telecel modified one BSCs 2G->3G reselecti

    on parameter, detail is: Qsearch_I =7 FDD_Qoffset =0 FDD_Qmin =7

    Means:

    While in GSM model the UE(support both 2G and 3G) will always search for 3G si

    gnal.

    When the Ec/Io of 3G signal is better than-12db, the UE will start to reselect from

    2G to 3G.

    Because the result is very good in the following 2 days after the parameter modifi

    cation, Current Operator did parameter modification to the other BSC. After checking K

    PIs for several days we find after the modification the Inter-RAT HO success rate have

    grown up a lot and kept on stable, other KPIs like Traffic volume, NUM of access also

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    8/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 8

    increase a lot. Current Operator Radio manager is very satisfied with the result.

    The KPIs changes after 2G parameter modification as below:

    figure3 NUM of inter-RAT HO changes

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    9/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 9

    figure4 Inter-RAT HO success rate changes

    figure5 NUM of Access changes

    figure6 Traffic changes

  • 7/29/2019 Less Samples Cause a Low 3G to 2G HO Success Rate

    10/10

    Internal Use Only

    All Rights reserved, No Spreading abroad without Permission of ZTE 10

    After the KPIs became normal the Inter-RAT HO to Traffic Ratio turn to:

    Begin ti

    me

    Number

    of attem

    pted out

    going CS

    inter-RA

    T hando

    vers

    Cell T

    raffic

    Volu

    me, C

    S(Erl)

    Inter-RAT

    HO to Tr

    affic Ratio

    (CS,Erl)

    Numb

    er of

    attem

    pted

    outgoi

    ng PS

    inter

    -RAT

    hando

    vers

    Total PS

    Traffic MB

    Inter-RAT

    HO to Tr

    affic Ratio

    (PS,MB)

    2012-06-

    1812639

    1213.

    6667

    10.4138969

    9

    44454

    16959.742

    96

    2.62114821

    6

    2012-06-

    1913219

    1208.

    4864

    10.9384764

    3

    45696

    16325.896

    86

    2.79898864

    9

    2012-06-

    2012810

    1175.

    8292

    10.8944394

    3

    48127

    17455.973

    21

    2.75705052

    2012-06-

    2112971

    1206.

    15

    10.7540521

    5

    47125

    18305.446

    48

    2.57437042

    3

    2012-06-

    2214508

    1279.

    5175

    11.3386491

    4

    47976

    18195.402

    32

    2.63671004

    2

    2012-06-

    2316803

    1394.

    7425

    12.0473850

    9

    54943

    17206.769

    29

    3.19310377

    6

    2012-06-

    2414060

    1172.

    0358

    11.9962205

    9

    47366

    16965.943

    072.7918283

    5 Conclusion

    This document find a new way by analyzing user behavior and Mathematical deduc

    tion and bring up a new index Inter-RAT HO to Traffic Ratio to solve the 23G inter

    operability problem at last..