Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [STDS-802-16] FW: [STDS-802-16] questions regarding 802.16e WMAN-OFDMA?



Dear all,

 

 

Thank you again Itzik, Zion, for you feedback.

 

I  have the following questions follow up :

 

<<<< (question):

If all 42 symbols can be used for DL in TDD, without TTG (RX-TX gap) the frame length will be 4.8 msec (42 * 115.2 usec) instead of 5 msec. Even in that case, the gap between frame will be RTG only as described above?  If so, wouldn’t it cause the network sync problem? Or, even though all the 42 symbols are used, there still will be TTG portion in the frame?  

 

<<<<<<(QUESTION)

 Basically, in WMAN-OFDMA DL, there are two types of mandatory subchannel allocations, PUSC & FUSC.

 In DL PUSC, one slot is defined as ‘2 symbols & 1 subchannel’ and in DL FUSC, it is defined as ‘1 symbol & 1 subchannel’.

 Thus, if ‘No. subchannels’ is indeed defined as “The no. subchannels parameter indicated the number of subchannels for a specific allocation”  not as per symbol,

then I have another question about it. In WMAN-OFDMA DL_MAP_IE from 802.16-2004/Cor1/D5, the number of subchannels is defined in 6 bits while the number of symbols is defined in 7 bits. The point that I am confused at with his definition is that the maximum size of burst is limited with the number of subchannels in FUSC (which is 64 subchannels). In FUSC, there will be 16 subchannels per one OFDM symbol, thus it means, the max size of DL burst will take only 4 symbols in total (assuming whole subchannels per symbol are used for only one user)?  To rephrase my question, can I ask the definition of ‘No. subchannels’ again just to make sure if it indicate the ‘number of total subchannels within one burst allocation ‘ or it is the ‘number of subchannels per OFDM symbol’? If it is the ‘number of subchannels per symbol’, then the burst size can be derived (in  # of slots) as (No.subchannels X No. OFDM symbols) in FUSC and  (No. Subchannels X (No. OFDM symbols/2)) in PUSC; or if it the ‘number of subchannels in the DL burst’, the the burst size (in # of slots) would be N0.subchannels in FUSC with the constraint of N0. OFDM symbols = No. subchannels and in PUSC with the constraint of No. OFDM symbols = No. Subchannels/2.    

 

 

 

 

 

 Thank you in advance for your feedback.

 

Best regards,

 

Youssef

 

 


From: Itzik Kitroser [mailto:itzikk@runcom.co.il]
Sent: Sunday, March 19, 2006 4:29 AM
To: youssefa@earthlink.net
Cc: 'Zion Hadad'
Subject: RE: [STDS-802-16] FW: [STDS-802-16] questions regarding 802.16e WMAN-OFDMA?

 

Hello Youssef,

Please find some answers from Zion and myself enclosed within the text.

 

Itzik.

 


From: Youssef Abdelilah [mailto:youssefa@earthlink.net]
Sent: Friday, March 17, 2006 8:30 PM
To: Zion Hadad; STDS-802-16@listserv.ieee.org
Subject: RE: [STDS-802-16] FW: [STDS-802-16] questions regarding 802.16e WMAN-OFDMA?

Thank you Zion,

 

Any input on  the other questions or anyone from 802.16 or 802.16e group. Thank you in advance

 

2.The largest and smallest time for RTG (Receiver - Transmitter Gap) between two consecutive frames.  

This is really depends on you specific implementation and system parameters (such as frame size and number of symbols within a frame). Probably you will not require more than one OFDM/OFDMA symbol as an upper bound. For example, if you look at the profile section, you will find requirements for subscriber side RTG (SSRTG) will not to exceed 50 us for TDD and 100 us for H-FDD.

<<<< (question):

If all 42 symbols can be used for DL in TDD, without TTG (RX-TX gap) the frame length will be 4.8 msec (42 * 115.2 usec) instead of 5 msec. Even in that case, the gap between frame will be RTG only as described above?  If so, wouldn’t it cause the network sync problem? Or, even though all the 42 symbols are used, there still will be TTG portion in the frame?  

 

 

 

3. Each SS (Subscriber Station) or MS (Mobile Station) have to process only specific DL bursts (i.e., not all DL bursts)?    

This depends on your operational mode. Basically the SS/MS should decode all DL bursts which is capable, this means less and equal to its modulation/coding rate parameters. In addition, in OFDMA, you can provide a list of CIDs for each DL burst allocation (i.e., DL-MAP IE) such that the mobile user will be able to decode only the relevant bursts. This can help the power consumption of the MS.

-- If so, what parameters that PHY can expect from the above layer and also, what are the range of the parameters?    

Please look at the above answer. 

4. In each DL_MAP_IE of WMAN-OFDMA, information of the size and the location of each logical DL burst are provided as following 4 parameters, 'OFDMA_Symbol Offset', 'subchannel offset', 'No. OFDMA symbols', and 'No. subchannels'. Though it's not clear, we assume that 'No. subchannels' indicates the number of subchannels corresponding to one symbol (i.e., not whole number of subchannels in a burst). What is the maximum value of those parameters for the system we described above?   

The no. subchannels parameter indicated the number of subchannels for a specific allocation. The combination of these parameters basically provide a rectangular allocation within the DL frame for a specific DL-Burst.

From the BS side, there is no limit, so theoretically, the BS can allocate one rectangle with specific profile. In practice this is a combination of requirements such as how many profiles to allocate (because we want to maximize the system throughput), power boosting, etc.

The problem that I see is that if we take the maximum as the largest number from the bits, one DL burst can take whole frame. Thus, I am wondering if there is any rule to regulate the burst size beyond those parameter expression.   

This is exactly the task of the scheduler and what makes one's system more efficient than other's :-)

 <<<<<<(QUESTION)

 Basically, in WMAN-OFDMA DL, there are two types of mendatory subchannel allocations, PUSC & FUSC.

 In DL PUSC, one slot is defined as ‘2 symbols & 1 subchannel’ and in DL FUSC, it is defined as ‘1 symbol & 1 subchannel’.

 Thus, if ‘No. subchannels’ is indeed defined as “The no. subchannels parameter indicated the number of subchannels for a specific allocation”  not as per symbol,

then I have another question about it. In WMAN-OFDMA DL_MAP_IE from 802.16-2004/Cor1/D5, the number of subchannels is defined in 6 bits while the number of symbols is defined in 7 bits. The point that I am confused at with his definition is that the maximum size of burst is limited with the number of subchannels in FUSC (which is 64 subchannels). In FUSC, there will be 16 subchannels per one OFDM symbol, thus it means, the max size of DL burst will take only 4 symbols in total (assuming whole subchannels per symbol are used for only one user)?  To rephrase my question, can I ask the definition of ‘No. subchannels’ again just to make sure if it indicate the ‘number of total subchannels within one burst allocation ‘ or it is the ‘number of subchannels per OFDM symbol’? If it is the ‘number of subchannels per symbol’, then the burst size can be derived (in  # of slots) as (No.subchannels X No. OFDM symbols) in FUSC and  (No. Subchannels X (No. OFDM symbols/2)) in PUSC; or if it the ‘number of subchannels in the DL burst’, the the burst size (in # of slots) would be N0.subchannels in FUSC with the constraint of N0. OFDM symbols = No. subchannels and in PUSC with the constraint of No. OFDM symbols = No. Subchannels/2.    

 

 5. (This question is somewhat related to the question 3 & 4) Is it possible that whole DL subframe can be assigned to a single user  Yes   

 ? If so, the whole information for the single user can be put into one single burst or have to be in multi-bursts  

This depends on specific requirements/possibilities of the scheduler and characteristics of the traffic for the user.

 

 

Regards,

 

Youssef

 

 


From: Zion Hadad [mailto:zionh@RUNCOM.CO.IL]
Sent: Thursday, March 16, 2006 11:54 PM
To: STDS-802-16@listserv.ieee.org
Subject: Re: [STDS-802-16] FW: [STDS-802-16] questions regarding 802.16e WMAN-OFDMA?

 

All the 42 in case of Broadcasting only (as far as the Phy)

 


From: Youssef Abdelilah [mailto:youssefa@EARTHLINK.NET]
Sent: Thursday, March 16, 2006 5:23 PM
To: STDS-802-16@LISTSERV.IEEE.ORG
Subject: [STDS-802-16] FW: [STDS-802-16] questions regarding 802.16e WMAN-OFDMA?

 

Dear all,

 

Target System: WMAN-OFDMA with 10 MHz bandwidth, frame period of 5 ms, TDD mode, 1024 fixed FFT points, and 42 symbols in a frame.

Questions:

1. The sizes for DL subframe and UL subframe in one TDD frame are supposed to be given to PHY layer from the above layer (it is written in the spec); what is the worst case for the DL subframe size. In other words, among 42 symbols in a frame what would be the largest number of symbols that can be assigned to DL subframe.

2. The largest and smallest time for RTG (Receiver - Transmitter Gap) between two consecutive frames.

3. Each SS (Subscriber Station) or MS (Mobile Station) have to process only specific DL bursts (i.e, not all DL bursts)?

-- If so, what parameters that PHY can expect from the above layer and also, what are the range of the parameters?

4. In each DL_MAP_IE of WMAN-OFDMA, information of the size and the location of each logical DL burst are provided as following 4 parameters, 'OFDMA_Symbol Offset', 'subchannel offset', 'No. OFDMA symbols', and 'No. subchannels'. Though it's not clear, we assume that 'No. subchannels' indicates the number of subchannels corresponding to one symbol (i.e., not whole number of subchannels in a burst). What is the maximum value of those parameters for the system we described above?

The problem that I see is that if we take the maximum as the largest number from the bits, one DL burst can take whole frame. Thus, I am wondering if there is any rule to regulate the burst size beyond those parameter expression.

5. (This question is somewhat related to the question 3 & 4) Is it possible that whole DL subframe can be assigned to a single user? If so, the whole information for the single user can be put into one single burst or have to be in multi-bursts?

 

Thank you,

 

Youssef Abdelilah