[Submitter's Last Name]
Stanwood
[Submitter's First Name]
Ken
[Membership Status: (M)ember; (O)bserver; (N)either]
M
[Starting Page #]
112
[Starting Line #]
52
[(T)echnical for Content-Related Material; (E)ditorial for typos,
grammar, etc., (TBD) for a comment on a section marked "TBD"]
T
[Detailed Description of Proposed Insertion, Deletion, Change]
A dcoument titled "Combining Burst FDD and FSDD modes in 802.16.1" is being submitted
to further explain this change.
This comment affects section 2.6.3 through page 114, line 28.
The Burst FDD and FSDD variants of PHY mode B can be combined into a single variant that
always starts a frame with TDM. TDM continues as long as it is feasible to allocate
to either full-duplex terminals or terminals which are half-duplex yet can receive
before they transmit in the current frame. After the feasibility of TDM is exhausted
for the frame, the frame can continue with TDMA. Figures 52 and 53 would be replaced
with a single figure showing the combined mode. This figure was given to the editors
in "FDD_TDM_TDMA DL Subframe.vsd".
Depending upon the BS scheduling of the downlink, a system could still choose to operate
the downlink TDM only, TDMA only or mixed TDM and TDMA.
Here are the specific edits for this section for this section:
/////start of replacement text for section 2.6.3
In line 55, change "(TDD, Burst FDD and FSDD)" to "(TDD and Burst FDD)".
In line 57, change "has two optional forms as shown in Figure 39 and Figure 53" with
"allows TDM and TDMA to co-exist in the same frame as shown in Figure 52".
In line 58, change "In both cases it starts" with "The frame starts".
On page 113 change lines 1 through 22 with:
"All downlink subframes start with a TDMA section. They may optionally contain a TDMA
section. This structure is shown in Figure 52. Whether an interval is TDM or TDMA is
indicated by the DL IUC in the DL-MAP message in the Frame Control Header.
The TDM section is sorted by modulation withthe modulation changed indicated in the
DL-MAP. They may optionally be sorted by CPE. Note that any one of the differently
modulated TDM sections may be absent.
In the TDMA portion of the downlink subframe, downlink data destined to each individual
CPE is grouped into one (preferable) or more bursts, each starting with a short (e.g.,
12 symbos) preamble. The DL-MAP also contains a map of these bursts.
In TDD systems, the TDMA section would typicall never be used, and no DL IUCs in the
DL-MAP would be present except those indicating the TDM transitions. The TDD case would
also contain a Tx/Rx Transition Gap (TRTG).
In FDD systems, the goal is to keep the TDMA section as short as possible. Since full-
duplex FDD terminals and half-dulpex FDD terminals which receive before they transmit
can receive the TDM portion of the downlink. In general, TDMA bursts will be present in
a frame only in those instances where a half-duplex terminal is required to receive
before it transmits. This allows the efficiency (due to reduced overhead) of TDM to be
used as much as possible while allowing co-existing TDMA to increase statistical
multiplexing of half-duplex terminals.
Relpace Figure 52 and Figure 53 with "FDD_TDM_TDMA DL Subframe.vsd"
////end of replacement text fro section 2.6.3
The change also affects the DL-MAP format described in pages 77-79 since the map would
need to allow TDM style MAP elements followed by TDMA style MAP elements. This would be
achieved by specifically defining the DL IUCs (something that needs to be done anyway),
to distinguish between TDM and TDMA. Since TDM is broadcast, TDM intervals would be
addressed to the broadcast channel ID. Alternatively, terminal/connection IDs can be
completely removed in the DL-MAP by requiring all terminals to listen to all intervals
with an IUC compatible (at or more robust than) with their negotiated DL modulation/FEC.
Sections 3.2.1 and 3.2.2 on page 269 become combined in a similar fashion.
Figures 120, 122, and 123 on pages 271-272 become combined into a similar picture and
their associated text in section 3.3 on pages 271 and 272 must be changed to reflect
the merge.
[Reason for Edit]
This has the advantage that TDD, FDD TDM, and FSDD all share the same DL maps. It also
has the advantage that you don't need to differentiate between Burst FDD and FSDD - they
become one. The efficiency of FDD TDM and the statistical multiplexing gains of FSDD
can be realized simultaneously.
The change in no way affects systems that use PHY mode A.