Question on HSUPA TTI

[firstmaxim]

HSUPA uses TTI of 2 ms as well as 10 ms. Could someone suggest applications for these two TTI durations.

[ritgail]

HSUPA uses TTI of 2 ms as well as 10 ms. Could someone suggest applications for these two TTI durations.

Hi,

This two TTI are there because of the fact that the UE categories, some of which supports only 10ms TTI. Google HSUPA UE categories. Push repo if u like..

[rishbi3]

Hi,

Changing TTI from 10 to 2 ms reduces Latency.

[paulus]

Hi..

As far as I know, this function default is 2 ms.

The TTI will be only 3 timeslots long (2 ms), compared to 15 timeslots (10 ms) employed by the other physical channels. When a shorter TTI is used, the UE can inform the network every 2 ms if the transmission failed. In the old scheme, 10 ms would have to pass before a failure could be reported. Shorter frames also mean that the system can respond more quickly to changing channel conditions, and re-assign capacity amongst users.

Br,

paul

[boring]

10-ms TTI has larger coverage than 2-ms TTI, so if UE is located at cell edge then 10-ms will be automatically chosen by RNC

10-ms TTI provides peak uplink datarate of about 1800 kbps, which means most of the services will be acceptable
2-ms TTI provides even higher rates but also lower latency (e.g. suitable for VoIP services), but always you need to be careful that 2-ms coverage is smaller than that of 10-ms

Reps highly appreciated!!!

[fahmi]

10-ms TTI has larger coverage than 2-ms TTI, so if UE is located at cell edge then 10-ms will be automatically chosen by RNC

10-ms TTI provides peak uplink datarate of about 1800 kbps, which means most of the services will be acceptable
2-ms TTI provides even higher rates but also lower latency (e.g. suitable for VoIP services), but always you need to be careful that 2-ms coverage is smaller than that of 10-ms

Reps highly appreciated!!!

Hi boring,
Could please explain the point that 10-ms TTI has larger coverage than 2-ms TTI and especially why ??

Thaks a lot.

[firstmaxim]

Do you have any doc to justify that 10 ms TTI gives a higher coverage. Thanks for the post.

[brianm]

How does the change from 10ms to 2ms affect coverage???

[fahmi]

Hi All,

In fact, we have a compromise in our choice of TTI :


In order to be able to adapt quickly to the changing conditions in the radio link a communications system must have shorter TTIs.
In order to benefit more from the effect of interleaving and to increase the efficiency of error-correction and compression techniques a system must, in general, have longer TTIs.

Let's consider a UE quite far from the NodeB with channel quality not good enough, in such a case, when using 2m TTI there will be a lot of retransmission of same packets until the UE receives and decodes the right one.

But, with a 10 ms TTI, the execution of channel coding and especially interleaving will be more effective (with more space in a 10 ms TTI transmission packet)

That's why it seems that a 10 ms TTI will ensure better coverage.

How do you find this ??
Waiting for your comments.

BR,
Fahmi.

[boring]

Hi guys
yeap this is basically the case. the longer the TTI, the better the link performance BLER for same EbNo (remember in R99 we have bearers with 20 or even 40 ms TTI)
and as rightly you mention this is primarily due to interleaving working better.
cheers
b

[giegala]

Just to add:

2m TTI will require more power to transmit same amount of data than 10ms TTI. Because UE will power limited at cell edge, 10ms TTI gives better performance.

[firstmaxim]

Fahmi,

Agree with your statements. But, is there any reasons why this 2ms and 10ms TTI option is available only in HSUPA and not in HSDPA?

[fahmi]

Fahmi,

Agree with your statements. But, is there any reasons why this 2ms and 10ms TTI option is available only in HSUPA and not in HSDPA?

Hi,
Not exactly, 2ms TTI was introduced for the first time with HSDPA R'5

Whats New in HSDPA:

3 Slot TTI (2ms)
CQI: Channel Quality Indication
AMC: Adaptive Modulation and Coding
Constellation Re-Arranging (16 QAM)
H-ARQ: Hybrid Automatic Repeat Request
N-channel SAW: N-Channel Stop and Wait Protocol
New MAC (High Speed) instance

[firstmaxim]

Fahmi, i am trying to find out why this 10 ms option is not introduced by the standard for HSDPA. For, HSUPA, both 10 ms and 2ms TTI options are available.

[boring]

10-ms is not needed in HSDPA because the NodeB has sufficient power to provide service across the entire cell area

if for some reason, NodeB was not able to allocate large amounts of power to a signle user (TTI), then the 10-ms would be needed. but of course, by means of HSDPA design this does not happen

[firstmaxim]

I agree that the TTI sizes are designed to factor in the Node B/UE power asymmetry.