PDA

View Full Version : Help for HSDPA service drops.



kadir
2010-01-30, 10:18 PM
Hi all;

I have problem about HSDPA, in our network, when the users is connected to internet, if they dont download anythink, the modem disconnects and this happens at busyhour.What are the possible reasons?

Secondly, our RAB establishment failure rate is high, does it effect the drop rate or not ? This also happens in busy hour at high loaded node-b.

could you show me the way at least, please help...??

BR.

kadir
2010-01-31, 07:24 AM
no comment ?

wolverine
2010-01-31, 08:45 AM
Hi Kadir,

When you say "the modem disconnects", do you mean a PDP Context Deactivation?

If you are talking about the RRC connection being released, then quite a few devices (esp. smartphones) will send a Signalling Connection Release Indication when there is no activity for 5-10s. This is termed "Fast Dormancy", and your counters should take this into account and not count it as an abnormal release.

If neither of the above is helpful, then it might help if you provide some further information about your problem.

cococrunch
2010-02-05, 04:52 PM
Hi Kadir,

RAB Estab failures are likely to occur on high loaded NodeBs. May I ask which capacity is congested, i.e. CE, Iub, Codes, UL/DL Power, etc.?? And what is the failure cause of the RAB Estab?

If you look at the trend per hour of a CONGESTED site with high RAB Estab fails, you will notice that failures occur on the busy hours only (6:00PM - 8:00PM in our network). This is one quick sign to look at congested sites.

Thanks.

kadir
2010-02-06, 03:29 AM
Hi friend;

when i look at the stats, Our congestion is on UL CE. can i ask, how many CE is used on HSPA service per user? or is it needed for HSPA? in our licence the CE elements are max. in UL/DL. If it is because of the congestion i think we need new Node-B, whitout this we can not solve this problem.

Thank you, so much.


Hi Kadir,

RAB Estab failures are likely to occur on high loaded NodeBs. May I ask which capacity is congested, i.e. CE, Iub, Codes, UL/DL Power, etc.?? And what is the failure cause of the RAB Estab?

If you look at the trend per hour of a CONGESTED site with high RAB Estab fails, you will notice that failures occur on the busy hours only (6:00PM - 8:00PM in our network). This is one quick sign to look at congested sites.

Thanks.

cococrunch
2010-02-07, 02:37 PM
Hi Kadir,

CE usage is shared in HSPA depending on Baseband dimensioning. In NSN, CE requirement are as follows.

1. 16 users per BTS --> 32 CE required
2. 16 users per cell --> 96 CE required
3. 48 users per BTS --> 80 CE required
4. 48 users per cell --> 240 CE required


This is only for HSPA downlink, associated HSDPA UL and HSUPA are not included. HSUPA is not enabled in our network yet because of heavy CE usage(more often than not UL CE util is greater than DL CE Util). I think its better to have RU10(FSME) first before enabling HSUPA.

If not RU10, you can try to activate the feature HSDPA 16 kbps UL Return Channel. This enables you to have 1 CE(16 kbps) usage on the associated HSDPA UL rather than 4 CE(64 kbps), that is 75% off on the associated HSDPA UL CE usage.

Thanks

maasricht
2010-02-07, 03:59 PM
Hi Kadir,

CE usage is shared in HSPA depending on Baseband dimensioning. In NSN, CE requirement are as follows.

1. 16 users per BTS --> 32 CE required
2. 16 users per cell --> 96 CE required
3. 48 users per BTS --> 80 CE required
4. 48 users per cell --> 240 CE required


This is only for HSPA downlink, associated HSDPA UL and HSUPA are not included. HSUPA is not enabled in our network yet because of heavy CE usage(more often than not UL CE util is greater than DL CE Util). I think its better to have RU10(FSME) first before enabling HSUPA.

If not RU10, you can try to activate the feature HSDPA 16 kbps UL Return Channel. This enables you to have 1 CE(16 kbps) usage on the associated HSDPA UL rather than 4 CE(64 kbps), that is 75% off on the associated HSDPA UL CE usage.

Thanks

I want to mention CE for Flexi Rel2 on RU10.


1. 16 users per BTS --> 32 CE required
2. 16 users per cell --> 96 CE required (3 cells)
3. 48 users per BTS --> 72 CE required
4. 48 users / 64 users per cell --> 216 CE required

If use FSMD/ FSME, no needed to reserve CE for CCCH.

mirza
2010-02-07, 04:13 PM
Hi all;

I have problem about HSDPA, in our network, when the users is connected to internet, if they dont download anythink, the modem disconnects and this happens at busyhour.What are the possible reasons?

Secondly, our RAB establishment failure rate is high, does it effect the drop rate or not ? This also happens in busy hour at high loaded node-b.

could you show me the way at least, please help...??

BR.

Hi

can you tell me if possibel , you work for which carrier / network ?

thanks

mirza

gprastomo
2010-02-07, 04:52 PM
Hi all;

I have problem about HSDPA, in our network, when the users is connected to internet, if they dont download anythink, the modem disconnects and this happens at busyhour.What are the possible reasons?

Secondly, our RAB establishment failure rate is high, does it effect the drop rate or not ? This also happens in busy hour at high loaded node-b.

could you show me the way at least, please help...??

BR.

Hi,

FIrst of all, the PDP context deactivation could be caused by several reason :
1. if there is no activity for some period, the RRC state will move to FACH then CELL PCH then idle mode (disconnect), this one is methode to make the Radio Resource more efficient, usually it takes ~1 hour depend on your parameter setting, check this behaviour
2. according to your information, the RAB estab fail is so high, this one of indication,let say users try to make a call then fail so they will try to make a call again more than 1 in the same time so the Radio Load both UL&DL will overloaded (imagine 100 users try to make call in the same time) and the RACH power is higher than dedicated channel so the possibility of having overload on the radio side is so high, during this time the existing user which already connect to the network will be affected then by Radio REsource procedure the existing call will be downgraded even will be dropped (include HSDPA), then the new Call will be rejected. So try to investigate why the Rab Estab fail is so high, is it due to UL/DL Admission control, IuB Blocking, CE blocking, IuPS/CS blocking see on the stats then try to solve it.

3. You have to breakdown on the counter which on of the most drop on that particular cell, is it due to radio, capacity etc ?

try above first the we can see on the next step

regard

kadir
2010-02-07, 05:41 PM
Hi

can you tell me if possibel , you work for which carrier / network ?

thanks

mirza

hi,

i work for huawei wcdma network

thanks

Andry
2010-02-07, 06:20 PM
This Congestion, they can be two cases:
- A shortage of channel element;
- Low Forward Power.
Requires Optimization Power Reductions Parameters and build the network.

I hope you can help my information.

kadir
2010-02-07, 07:14 PM
Hi,

FIrst of all, the PDP context deactivation could be caused by several reason :
1. if there is no activity for some period, the RRC state will move to FACH then CELL PCH then idle mode (disconnect), this one is methode to make the Radio Resource more efficient, usually it takes ~1 hour depend on your parameter setting, check this behaviour
2. according to your information, the RAB estab fail is so high, this one of indication,let say users try to make a call then fail so they will try to make a call again more than 1 in the same time so the Radio Load both UL&DL will overloaded (imagine 100 users try to make call in the same time) and the RACH power is higher than dedicated channel so the possibility of having overload on the radio side is so high, during this time the existing user which already connect to the network will be affected then by Radio REsource procedure the existing call will be downgraded even will be dropped (include HSDPA), then the new Call will be rejected. So try to investigate why the Rab Estab fail is so high, is it due to UL/DL Admission control, IuB Blocking, CE blocking, IuPS/CS blocking see on the stats then try to solve it.

3. You have to breakdown on the counter which on of the most drop on that particular cell, is it due to radio, capacity etc ?

try above first the we can see on the next step

regard

thanks for your kindly reply my friend,

i will check counters that you suggested, then we can continue the brain storm :)

kadir
2010-02-07, 07:19 PM
Hi Kadir,

CE usage is shared in HSPA depending on Baseband dimensioning. In NSN, CE requirement are as follows.

1. 16 users per BTS --> 32 CE required
2. 16 users per cell --> 96 CE required
3. 48 users per BTS --> 80 CE required
4. 48 users per cell --> 240 CE required


This is only for HSPA downlink, associated HSDPA UL and HSUPA are not included. HSUPA is not enabled in our network yet because of heavy CE usage(more often than not UL CE util is greater than DL CE Util). I think its better to have RU10(FSME) first before enabling HSUPA.

If not RU10, you can try to activate the feature HSDPA 16 kbps UL Return Channel. This enables you to have 1 CE(16 kbps) usage on the associated HSDPA UL rather than 4 CE(64 kbps), that is 75% off on the associated HSDPA UL CE usage.

Thanks

First of all, thank you for your valuable info.

our network is huawei, and when i check the counter we have UL CE congestion, %99 percent of the RAB fail due to UL CE congestion. In our network HSDPA feature active but our modem does not support the HSUPA feature yet, so, why we have UL CE congesiton?

Thanks,

gprastomo
2010-02-07, 07:43 PM
First of all, thank you for your valuable info.

our network is huawei, and when i check the counter we have UL CE congestion, %99 percent of the RAB fail due to UL CE congestion. In our network HSDPA feature active but our modem does not support the HSUPA feature yet, so, why we have UL CE congesiton?

Thanks,

This problem i have ever heard on Huawei Network, UL CE suddenly have high utilization.
What im affraid is there is problem with SW or HW.
But even you dont have HSUPA yet then try to see how many HSDPA user on that cell, and because if there is huge number of HSDPA user usually there will be problem on UL CE usage if your RAB combination is 64k UL and HSDPA on DL. THen you can try to activate 16 kbps UL HSPDA return channel as previous comment. ANd try to reduce the SHO Overhead. But If there is problem on SW or HW, then try to restart it first :)

br

Ctap
2010-02-08, 12:24 PM
First of all, thank you for your valuable info.

our network is huawei, and when i check the counter we have UL CE congestion, %99 percent of the RAB fail due to UL CE congestion. In our network HSDPA feature active but our modem does not support the HSUPA feature yet, so, why we have UL CE congesiton?

Thanks,


Hi!
If your network not use HSUPA try to check that HSUPA turned off (In fact that turned on HSUPA use 60 UL CE permanent)

DSP CELL: DSPT=BYCELL, CELLID=CELL;

In this case must be some thing like that

HSUPA State Information
-----------------------
HSUPA Operate State = HSUPA Disabled
HSUPA Disabled Cause = HSUPA Not Actived

LST BRDHSUPA: SRN=0, SN=0;

Once more thing its try to rearrange license on NodeB in your network. One BBU3806 without extension can use up to 192 UL CE.

kadir
2010-02-09, 03:52 PM
This problem i have ever heard on Huawei Network, UL CE suddenly have high utilization.
What im affraid is there is problem with SW or HW.
But even you dont have HSUPA yet then try to see how many HSDPA user on that cell, and because if there is huge number of HSDPA user usually there will be problem on UL CE usage if your RAB combination is 64k UL and HSDPA on DL. THen you can try to activate 16 kbps UL HSPDA return channel as previous comment. ANd try to reduce the SHO Overhead. But If there is problem on SW or HW, then try to restart it first :)

br

Hi friend;

How can i make HSDPA UL RAB combination to 16kbps, i mean which parameter is related with it .?

thanks for your kindly reply.

gprastomo
2010-02-09, 04:06 PM
Hi friend;

How can i make HSDPA UL RAB combination to 16kbps, i mean which parameter is related with it .?

thanks for your kindly reply.

THis one is feature, and at that time just suggest to the team and the NSN team discuss internally to activate the feature. In NSN it should be HSDPA16KBPSReturnChannel

br

Ctap
2010-02-10, 10:51 AM
Hi friend;

How can i make HSDPA UL RAB combination to 16kbps, i mean which parameter is related with it .?

thanks for your kindly reply.

Hi, one question - you have this situation on whole network?

venom
2010-02-10, 01:12 PM
Frequent Disconnections is not only because of congestion, also check for Iub, Check for Connection Manager / CoCo if it is made correctly, check for CQI (Channel Quality indicator), Also check for PRx Noise if it is within specs... there could be many reasons

In HSDPA, only Hard handover takes places, also check for handover settings......... !

kadir
2010-02-10, 03:39 PM
Hi, one question - you have this situation on whole network?

hi;

the problem is not in whole network, mostly on high loaded node-b in busy hour.

marceloastengo
2010-02-11, 03:17 AM
what is high load in node b

kadir
2010-02-11, 04:48 AM
what is high load in node b

i mean congested node-b in busy hour.

koderman
2010-02-11, 05:42 AM
I think, need write log (drivetest) file with this problem and trace file from RNC. And analyze both information. For my NSN WCDMA net we have problem like this with RAS06. But after RU10 this problem go awey... In my situation its problem was because of very loaded Node B (Flexy BTS) and problems with BTS processors... Huawei equipment may some problem like this... Post logfile and trace file for the forum pro-user analisis...

Ctap
2010-02-11, 09:09 AM
i mean congested node-b in busy hour.
HI!
How many UL CE license on this NodeB, what type of HW/SW use on your network. How many HSDPA subscribers on cell of this NodeB in BusyHour? Check maximum VS_HSDPA_UE_Mean_Cell in the cell. Check HSUPA (on or off on cell???).
In our network we have the same problems, but we have poor resource on IuPS i-face.

kadir
2010-02-11, 04:04 PM
HI!
How many UL CE license on this NodeB, what type of HW/SW use on your network. How many HSDPA subscribers on cell of this NodeB in BusyHour? Check maximum VS_HSDPA_UE_Mean_Cell in the cell. Check HSUPA (on or off on cell???).
In our network we have the same problems, but we have poor resource on IuPS i-face.

Hi;

in our network, HW CE capabilities are in UL=192 DL=256 and SW licence is the same. HSUPA feature is active and max user 25-30 in busy.

thanks my friend,

kadir
2010-02-11, 06:09 PM
HI!
How many UL CE license on this NodeB, what type of HW/SW use on your network. How many HSDPA subscribers on cell of this NodeB in BusyHour? Check maximum VS_HSDPA_UE_Mean_Cell in the cell. Check HSUPA (on or off on cell???).
In our network we have the same problems, but we have poor resource on IuPS i-face.


I wonder, like below guy said that if we disable the HSUPA feature in whole network, can we earn extra channel element from it ? and especially in CE congested cell, can we make improvement alittle for it?

thanks

Ctap
2010-02-15, 09:40 AM
I wonder, like below guy said that if we disable the HSUPA feature in whole network, can we earn extra channel element from it ? and especially in CE congested cell, can we make improvement alittle for it?

thanks

Hi! You can try to turn off HSUPA only for this NodeB, this must free 60 UL CE on Uplink.

nupong
2010-02-16, 01:20 AM
Hi;

in our network, HW CE capabilities are in UL=192 DL=256 and SW licence is the same. HSUPA feature is active and max user 25-30 in busy.

thanks my friend,


Hi! You can try to turn off HSUPA only for this NodeB, this must free 60 UL CE on Uplink.

Could you please explain why HSUPA would reserve 60 UL CE? How about HSUPA "ON" and no HSUPA user, it i also reserve 60 UL CE?
Any relationship to 25-30 users in BH?


I'm just wondering whether we should check #HSUPA users (maybe HSUPA RAB). If less HSUPA users, only little improvement I think.

Moreover, If I'm not wrong, the network we are discussing having no HSUPA, rite?

I would like to make an apology for my (maybe) non-sense question as I'm newbie in HSPA:L

Ctap
2010-02-16, 10:46 AM
Could you please explain why HSUPA would reserve 60 UL CE? How about HSUPA "ON" and no HSUPA user, it i also reserve 60 UL CE?
Any relationship to 25-30 users in BH?


I'm just wondering whether we should check #HSUPA users (maybe HSUPA RAB). If less HSUPA users, only little improvement I think.

Moreover, If I'm not wrong, the network we are discussing having no HSUPA, rite?

I would like to make an apology for my (maybe) non-sense question as I'm newbie in HSPA:L


Hi!
In our practice. We have UL CE congestion, and one of recommendation is be - turn off HSUPA, after that we free about 60 UL CE.

nupong
2010-02-17, 04:15 PM
Hi!
In our practice. We have UL CE congestion, and one of recommendation is be - turn off HSUPA, after that we free about 60 UL CE.

I'm still surprised. What make use of HSUPA technology if we disable them.

kadir
2010-02-17, 05:05 PM
I'm still surprised. What make use of HSUPA technology if we disable them.

in our network there are no HSUPA supported modem, so, if we disable HSUPA feature, the users will use R99 tech. like before, they dont sense the difference.

nupong
2010-02-18, 07:46 AM
in our network there are no HSUPA supported modem, so, if we disable HSUPA feature, the users will use R99 tech. like before, they dont sense the difference.

So, this solution is applicable where there is no HSUPA usage only.

Any solution for HSUPA in real network?

Ctap
2010-02-18, 07:02 PM
So, this solution is applicable where there is no HSUPA usage only.

Any solution for HSUPA in real network?

Hi!
For HSUPA network i recommended if you have problem with UL CE congestion, try to add license, try to add BBU and second carrier.
May be help DCCC algorithm to downgrade minimum bitrate. But if all this method is not do effect, build new site. :rolleyes:

nabil el haddani
2010-02-18, 10:45 PM
Link is dead!!

kadir
2010-02-19, 06:20 AM
Link is dead!!

which link is dead my friend, there is no any link in this topic?

mmadan
2010-02-19, 10:23 PM
Hi Kadir,

My two cents - there are several UE implementation on the market that in case if service request fails repetitively, PDP Context will be deactivated locally.

One of the reasons of failing Service request type data is because of RAB establishment failures mentioned earlier by you.

So yes - there is a high probability that end user will experience so called "modem disconnection" during peak hour.

I'm not very familiar with Huawei therefore I have this question - does the pool of CE for UL and DL are managed separately or it is a common pool that it is dynamically alocated based on usage?

BR

cococrunch
2010-02-20, 02:56 AM
Hi Kadir,

I think what you can try to do is maximize the bit rate per SF on the UL since you're problem is on the uplink CE congestion and you dont have HSUPA yet.

Try to see first on the counters what SF is HSDPA Return Channel using in your network in majority.

For example, if you're HSDPA Return Channel uses UL SF 32, you try to maximize UL SF 64 so it will use SF 64 instead of SF 32. This way, it might improve you're accessibility since more users can access while using lower CE consumption. SF 64 has also lower SIR target than SF 32.

In our network theres only little improvement but it might have a different effect on yours. :)

kadir
2010-02-22, 04:48 PM
Hi Kadir,

My two cents - there are several UE implementation on the market that in case if service request fails repetitively, PDP Context will be deactivated locally.

One of the reasons of failing Service request type data is because of RAB establishment failures mentioned earlier by you.

So yes - there is a high probability that end user will experience so called "modem disconnection" during peak hour.

I'm not very familiar with Huawei therefore I have this question - does the pool of CE for UL and DL are managed separately or it is a common pool that it is dynamically alocated based on usage?

BR

Hi friend, i have a huawei doc about Admission control and one of the chapter that node-b credit usage and dimensioning, i could share it with you but i dont have soft-copy of it.

there are one paragraf that, " if the UL capacity credit and DL capacity credit are separate, the maintenance on the local cell (and local cell group, if any) is performed in the UL and DL, respectively."

and in an other pragraph, "if the UL Capacity and DL capacity credit are not separate, the maintenance only on the global capacity credit is performed for the local cell (and local cell group, if any)."

these two paragraph shows us a there are separate and not separate mode in huawei. maybe there is a feature which activate this, or, we can do it this with CE licence, i mean, with the licence there are UL and DL CE pool and we separate it to node-bs, may be there is a licence that has only one CE pool for UL and DL. I am working on this now, when i find the exact solution about it i will share with you immdietly.

kadir.

tukangoptim
2010-02-23, 02:34 AM
hi pak kadir, for congestion its easy lah.. just ask the project to increase credit available and save our analysis for another problem :D, AFAIK , huawei is one of the most generous vevndor when it comes to license and capacity. in my current project VSLCULCreditAvailableShared mostly 512, 768 , and even 1152. :p
for "disconnect problem", i share the view with gito. this may be caused by inactivity timer fach, ura pch bla bla bla. :o
can check using lst PSINACTTIMER, and lst UESTATETRANSTIMER for state transition. :time:

kadir
2010-02-24, 09:16 PM
hi pak kadir, for congestion its easy lah.. just ask the project to increase credit available and save our analysis for another problem :D, AFAIK , huawei is one of the most generous vevndor when it comes to license and capacity. in my current project VSLCULCreditAvailableShared mostly 512, 768 , and even 1152. :p
for "disconnect problem", i share the view with gito. this may be caused by inactivity timer fach, ura pch bla bla bla. :o
can check using lst PSINACTTIMER, and lst UESTATETRANSTIMER for state transition. :time:

Hi friend, here is the our timer settings that you asked.

List UE State-Transition Algorithms
-----------------------------------
Cell Reselection Timer[s] = 180
Cell Reselection Counter Threshold = 9
BE DCH To FACH Transition Timer[s] = 5
BE DCH To F/RACH Or F/RACH To PCH 4b Threshold[byte] = D64
BE DCH To F/RACH 4b Time To Trigger[ms] = D5000
BE DCH To F/RACH 4b Pending Time After Trigger[ms] = D1000
BE FACH To PCH Transition Timer[s] = 65535
BE FACH To PCH 4b Time To Trigger[ms] = D5000
BE FACH To PCH 4b Pending Time After Trigger[ms] = D1000
BE F/RACH To DCH 4a Threshold[byte] = D1024
BE F/RACH To DCH 4a Time To Trigger[ms] = D240
BE HS-DSCH To FACH Transition Timer[s] = 5
BE HS-DSCH To FACH 4b Threshold[byte] = D64
BE HS-DSCH To FACH 4b Time To Trigger[ms] = D5000
BE HS-DSCH To FACH 4b Pending Time After Trigger[ms] = D1000
BE FACH To HS-DSCH 4a Threshold[byte] = D1024
BE FACH To HS-DSCH 4a Time To Trigger[ms] = D240
Realtime Traff DCH or HSDPA To FACH Transition Timer[s] = 180
Realtime Traff DCH or HSDPA To FACH 4b Threshold[byte] = D64
Realtime Traff DCH or HSDPA To FACH 4b Time To trigger[ms] = D5000
Realtime Traff DCH or HSDPA To FACH 4b Pending Time[ms] = D16000
Realtime Traff FACH To DCH Or HSPA 4a Threshold[byte] = D1024
Realtime Traff FACH To DCH Or HSPA 4a Time To Trigger[ms] = D240
E-DCH Throu Meas Period[10ms] = 30
BE E-DCH To FACH State Transformation Timer[s] = 5
E-DCH To FACH 4b Threshold[kbit/s] = 8
E-DCH To FACH 4b Period Amount To Trigger = 2
E-DCH To FACH 4b Pending Period Amount After Trigger = 4
BE FACH To E-DCH 4a Threshold[byte] = D1024
BE FACH To E-DCH 4a Time To Trigger[ms] = D240
(Number of results = 1)

--- END

+++ WCDMA-RNC 2010-02-24 14:49:33
O&M #269897
%%LST PSINACTTIMER: LstFormat=VERTICAL;%%
RETCODE = 0 Execution succeeded.

List PS User Inactive Detecting Timer
-------------------------------------
Conversational service T1[s] = 20
Conversational service T2[s] = 20
Streaming service T1[s] = 20
Streaming service T2[s] = 20
Interactive service T1[s] = 20
Interactive service T2[s] = 20
Background service T1[s] = 20
Background service T2[s] = 20
IMS signal T1[s] = 20
IMS signal T2[s] = 20
(Number of results = 1)


br.

tukangoptim
2010-02-28, 03:46 AM
now we have more to discuss..
hmm.. where are the experts ? :cool: :cool:

kadir
2010-03-04, 09:12 PM
Hi everybody;

I have solved the problem about my 3g network i mean the actions that before, we had discussed in the forum are very helpfull for me to give breath to 3g network. My RAB estb. faliure rate decrease from 17% to 6%.

i have just want to share this and thanks soo much to everybody.

Best Regard,

tukangoptim
2010-03-06, 12:37 PM
Hi everybody;

I have solved the problem about my 3g network i mean the actions that before, we had discussed in the forum are very helpfull for me to give breath to 3g network. My RAB estb. faliure rate decrease from 17% to 6%.

i have just want to share this and thanks soo much to everybody.

Best Regard,

so brother, what wass the root cause ?

kadir
2010-03-07, 02:52 AM
so brother, what wass the root cause ?

we had UL CE congestion, %99 percent of the rab failures are coming from the UL CE congestion problem, so, we have disabled the hsupa feature(we earned 66 CE) and also with FRC parameter we decreases the UL BE INITIAL BIT RATE from 64 to 32kbit/sec, these two action is not permanent solution ofcourse but it gave us time for the new node-b.

dk2000
2010-03-08, 11:36 PM
Hi Kadir,
Maybe can you reduce the CE for HSUPA, I think it better than you disabled HSUPA on this cell??

kadir
2010-03-08, 11:47 PM
Hi Kadir,
Maybe can you reduce the CE for HSUPA, I think it better than you disabled HSUPA on this cell??

Hi friend;

we dont have any HSUPA supported 3G Modem in our network so, we dont need hsupa recently. But ofcourse i want to know that how can i reduce the hsupa CE and shift it to desired place? could you inform us?

thankks

threeman
2010-03-26, 01:15 PM
Hi,

FIrst of all, the PDP context deactivation could be caused by several reason :
1. if there is no activity for some period, the RRC state will move to FACH then CELL PCH then idle mode (disconnect), this one is methode to make the Radio Resource more efficient, usually it takes ~1 hour depend on your parameter setting, check this behaviour
2. according to your information, the RAB estab fail is so high, this one of indication,let say users try to make a call then fail so they will try to make a call again more than 1 in the same time so the Radio Load both UL&DL will overloaded (imagine 100 users try to make call in the same time) and the RACH power is higher than dedicated channel so the possibility of having overload on the radio side is so high, during this time the existing user which already connect to the network will be affected then by Radio REsource procedure the existing call will be downgraded even will be dropped (include HSDPA), then the new Call will be rejected. So try to investigate why the Rab Estab fail is so high, is it due to UL/DL Admission control, IuB Blocking, CE blocking, IuPS/CS blocking see on the stats then try to solve it.

3. You have to breakdown on the counter which on of the most drop on that particular cell, is it due to radio, capacity etc ?

try above first the we can see on the next step

regard

GPrastomo ..
1. I agree with u. there is a parameter called "inactivity timer" to trigger downswitch (DCH --> FACH --> Idle). this timer is to release resource in air interface that is not used in several time so can be used by another user. but this timer does not release the PDP context. so user modem still connect/still have IP which is allocated by SGSN.
So i think for this case, You should make sure from TEMS/NEMO by drivetest "who initiate PDP context deactivation ". if it initiate from SGSN(downlink direction), i have experience there is mismatch timer parameter (TIAS TIAR). if it initiate from MS , is there any bad radio condition ?

2. i agree with gprastomo. we can breakdown cause of high rab estb fail. if related with capacity we can see lack of hardware, lack of power, lack of DL code, lack of admisiion connection, failed after admision --> iub blocking, hardware problem, transmisi problem.