HSUPA TTI2ms Disable

[paulus]

Hi guys..

Here I want to share and ask you regarding to the trial I made below.

Below are the results of de-activation TTI 2ms of HSUPA (SET UCORRMALGOSWITCH: MAP_PS_STREAM_ON_HSDPA_SWITCH-0;) :

34903

34904

34905

34906


From the result above, i conclude some points :
1. Deactivating TTI2ms will reduce PS drop. As we know, that TTI 10 ms is more resistant toward fading, signal fluctuation during data transaction/data access
rather than using TTI 2ms. But with TTI 2ms, will be able to do correction (faster) if RF issues occur (transmission fail or etc). In addition, using TTI10ms, will have consume more resource than TTI 2ms. But from M2000 counter, I can't see which point of resource are more consumable in TTI 10 ms, seems no issue in capacity resource. In the other hand, it improve the HSDPA success rate.
2. Based on theory, HSSCCH will be transmitted each TTI interval periode. If with TTI 2ms, means HSCCH will be transmitted as per 2ms, if TTI 10 ms, HSSCCH will be transmitted per 10 ms. HSSCCH is a code of 128kbps which carries the HSDPA user (header information, data carried, etc). Which means, if using 10ms TTI, the number of transmitted users are lower than that of 2ms TTI setting. But, as we can see from the result above, seems the user of HSDPA/HSUPA is still stable. So does the HSDPA/HSUPA Payload.
3. After deactivating this function, speed test for UPload was taken, but seems no difference before and after deactivation. No complain issue regarding to this action. But from result above, HSUPA throughput was slightly decreased.

I just wonder, as a part of optimization, is it illegal to disable this TTI 2ms? Coz, I also compared RNC setting for other operators, some are also disabled.

I'd appreciate for any comments & input for this case..

Thanks
BR//Paul

[dtvt2]

3. After deactivating this function, speed test for UPload was taken, but seems no difference before and after deactivation. No complain issue regarding to this action. But from result above, HSUPA throughput was slightly decreased.

I thinks When you deactivate 2ms HSUPA, UL speed must decrease. if not maybe you were assign to 10ms HUSPA. PLS check the threshold of Ec/No, RSCP used to assign 2ms HUSPA to UE. In my network UL speed above 2Mbps

[khurrambilal01]

We had also disabled the use of HSUPA 2ms TTI in high traffic cell and observed reduction in usage of CEs, HSDPA per user throughput increased.

The test you have performed to check Pre/Post HSUPA throughput --> In pre test might be the load/Radio conditions were not so good that you were being assigned to HSUPA 2ms TTI, so that's why after changes still you are experiencing the same throughput because you are using same TTI 10ms as before in actual field test.

[dfindme]

Hi Paulus,

I think you want to improve CE uplink utilization, for me its should illegal cause if you have capacity issue you must upgarde, and impact to throughput user experience.

but you can do reduce usage between 2msTTI and 10msTTI ratio.
and before it please monitor current usage 2msTTI and 10msTTI.


thanks

[RFoptimizer]

Hi Paulus,
Switch from 2ms TTI to 10msTTI, reduce load and then RTWP so The improvement come from the Cells with high load and RTWP, Have you checked RTWP after and before.
thanks

[sabila]

speed if TTI 2ms(HSUPA phase 2 up to 5.7Mbps ON should can reach at least 3/4 Mbps
if off speed decrease--TTI 10ms(HSUPA phase 1,max maybe at 1.4 Mbps)

and TTI 2m on impacted RTWP more higher

[ninjafine]

Hi, Paulus! Check also round trip time. Some CS gamers may be disappointed.

[paulus]

Hi Paulus,

I think you want to improve CE uplink utilization, for me its should illegal cause if you have capacity issue you must upgarde, and impact to throughput user experience.

but you can do reduce usage between 2msTTI and 10msTTI ratio.
and before it please monitor current usage 2msTTI and 10msTTI.


thanks

Hi dfindme,

How to change the ratio between 2ms TTI usage and 10ms TTI? If incase I want Indoor user to use 2ms TTI more than 10ms TTI in order to improve HSUPA thorughput,

what parameters should be changed?

Thanks
BR//Paul

[paulus]

We had also disabled the use of HSUPA 2ms TTI in high traffic cell and observed reduction in usage of CEs, HSDPA per user throughput increased.

The test you have performed to check Pre/Post HSUPA throughput --> In pre test might be the load/Radio conditions were not so good that you were being assigned to HSUPA 2ms TTI, so that's why after changes still you are experiencing the same throughput because you are using same TTI 10ms as before in actual field test.


Hi khurrambilal01,

Thanks for your share. So it's true that tti 2ms will consume more CE resource than 10 ms coz higher rate will consume more CE.

BR//PCS

[paulus]

Hi, Paulus! Check also round trip time. Some CS gamers may be disappointed.

Hi ninjafine,

what counter to check round trip time?

Thanks
BR//Paul

[electron]

Hi,

You can check PS latency KPIs.


Hi ninjafine,

what counter to check round trip time?

Thanks
BR//Paul

[electron]

Hi guys,


I have below observation based on my experience as well:

EUL 2ms using huge UL CE and some operator prefer to reduce maximum number of EUL 2ms in cells or totally disable it !

By disabling EUL 2ms definitely average EUL Throughput will reduce ( preferably have a trial in a cluster level to observe the impact )

Logically UL RSSI after activating EUL 2ms shall increase but i personally did not see the impact cell level


Please share how much is average EUL Throughput via drive test in your network to compare is it low or in normal condition ( please mention EUL 2ms is activated or deactivated ) ?



Cheers

[agenov]

Just one remark, this switch is not for deactivation of HSUPA 2ms TTI . This just allows ur streaming service to be mapped on HSDPA channel.

BR
Alex


Hi guys..

Here I want to share and ask you regarding to the trial I made below.

Below are the results of de-activation TTI 2ms of HSUPA (SET UCORRMALGOSWITCH: MAP_PS_STREAM_ON_HSDPA_SWITCH-0;) :

34903

34904

34905

34906


From the result above, i conclude some points :
1. Deactivating TTI2ms will reduce PS drop. As we know, that TTI 10 ms is more resistant toward fading, signal fluctuation during data transaction/data access
rather than using TTI 2ms. But with TTI 2ms, will be able to do correction (faster) if RF issues occur (transmission fail or etc). In addition, using TTI10ms, will have consume more resource than TTI 2ms. But from M2000 counter, I can't see which point of resource are more consumable in TTI 10 ms, seems no issue in capacity resource. In the other hand, it improve the HSDPA success rate.
2. Based on theory, HSSCCH will be transmitted each TTI interval periode. If with TTI 2ms, means HSCCH will be transmitted as per 2ms, if TTI 10 ms, HSSCCH will be transmitted per 10 ms. HSSCCH is a code of 128kbps which carries the HSDPA user (header information, data carried, etc). Which means, if using 10ms TTI, the number of transmitted users are lower than that of 2ms TTI setting. But, as we can see from the result above, seems the user of HSDPA/HSUPA is still stable. So does the HSDPA/HSUPA Payload.
3. After deactivating this function, speed test for UPload was taken, but seems no difference before and after deactivation. No complain issue regarding to this action. But from result above, HSUPA throughput was slightly decreased.

I just wonder, as a part of optimization, is it illegal to disable this TTI 2ms? Coz, I also compared RNC setting for other operators, some are also disabled.

I'd appreciate for any comments & input for this case..

Thanks
BR//Paul

[ninjafine]

Hi ninjafine,

what counter to check round trip time?

Thanks
BR//Paul

Hi, Paul!
I'd rather suggest not to use counters and use real drive-test measurements to check round trip time, since it depends on packets size, intervals between packets ant etc.

[gunpowder]

how you can check RTT from field
thanks

[giegala]

You have to use PING. Multiple ICMP message size would cause different RTT - due to RRC state switching and PDU segmentation. Mupltiple parallel PINGs can be used to trigger RRC state/ channel/ HSUPA TTI switching.

[dfindme]

u can use BEHSUPA2MSTTIRATETHS on UFRC.

[g.kashi@yahoo.com]

The New Doc:

http://www.4shared.com/zip/YTbOJLCnce/_For_Customer_M2000-CME_Produc.html
http://www.4shared.com/rar/_R6oPJ2Uce/Huawei_Trng_2__1_.html