Bandwidth allocation and request mechanisms

Increasing (or decreasing) bandwidth requirements is necessary for all services except incompressible constant bit rate UGS connections. The bandwidth allocations of incompressible UGS connections do not change between connection periods. The requirements of compressible UGS connections, such as channelized T1, may increase or

decrease depending on the traffic condition. When an SS needs to ask for bandwidth on a connection with its scheduling service, it sends a message to the BS containing the immediate requirements. QoS for the connection was established at connection establishment and is looked up by the BS.

Requests

Requests refer to the mechanism used by SSs to inform to the BS regarding their need of uplink bandwidth allocation. A Request may come as a stand-alone bandwidth request header or it may come as a PiggyBack Request.

The Bandwidth Request message may be transmitted during uplink bandwidth allocation, except during initial ranging interval. Bandwidth Requests may be incremental or aggregate.

When the BS receives an incremental Bandwidth Request, it adds the quantity of bandwidth requested to its current bandwidth in the connection. Otherwise, if the BS receives an aggregate Bandwidth Request, it should replace its assigned bandwidth of the connection with the quantity of bandwidth requested. The Type field in the bandwidth request header indicates whether the request is incremental or aggregate. Since Piggybacked Bandwidth Requests do not have a type field, Piggybacked Bandwidth Requests should be always incremental.

Finally, when the BS receives a Bandwidth Request from any SS, it should make use of admission control scheme to check whether the request is permitted or not.

Polling

Polling is the process in which the BS allocates bandwidth to SS by sending request.

These allocations may be to an SS or to a group of SSs. The polling should be done on SS basis. Note that bandwidth is always requested by a CID and it is allocated to an SS.

According to the bandwidth congestion situation, there are three polling types, unicast, multicast, and broadcast.

First, we introduce the procedure of unicast polling. When an SS is polled, there is no explicit message that can be sent to poll SS. Rather, the BS actively allocate SS slots in UL-MAP (introduced in section 2.1.3) for sending bandwidth request. UGS connections do not need to be polled individually, because they need a much efficient way to get bandwidth unless they set the poll-me (PM) bit in the packet header. Unicast polling could only be used if the bandwidth is sufficient for polling whole individual SSs.

Figure 2.1 Unicast Polling

As illustrated in Figure 2.1, first the unicast polling procedure should check whether more BW is available for individual polling or not. If it is not enough, the multicast and broadcast polling algorithm should be initiated. Then we continue to make some check, for example, were SSs with expired polling interval? Did unpolled SSs PM bit set? After the necessary checking procedure, the SS would be polled.

In Figure 2.2, it shows that SS’s polling interval was already expired and BS allocated the bandwidth to SS. The allocated slots were defined in UL-MAP.

Then, we can realize that SS would transmit bandwidth request in allocated slots, just like Figure 2.3. It also provides contention period to a group SSs.

Finally, BW allocation algorithm would be used and uplink subframe map would be changed, as shown in Figure 2.4.

Figure 2.2 Additional BW allocation in unicast polling

Figure 2.3 Unicast Polling BW Request

Figure 2.4 Unicast Polling BW allocation for data

If available bandwidth is insufficient for polling whole individual SS, some SSs may be polled by multicast polling or broadcast polling. Certain CIDs are reserved for the two polling types, when some SSs share one multicast or broadcast polling, they should be assigned the same CID. An example is provided in Table 2.2.

UL-MAP IE fields Interval description

(UIUC = uplink interval usage code) CID (16 bits)

UIUC (4 bits)

Offset (12 bits)

Initial Ranging 0000 2 0

Multicast group 0xFFC5 Bandwidth Request 0xFFC5 1 405 Multicast group 0xFFDA Bandwidth Request 0xFFDA 1 605

Broadcast Bandwidth Request 0xFFFF 1 805

SS 5 Uplink Grant 0x007B 4 961

SS 21 Uplink Grant 0x01C9 7 1136

Table 2.2 Sample UL-MAP with multicast and broadcast IE

The information exchange procedure for multicast and broadcast polling is shown in

Figure 2.5. Different from individual polling that depends on contention period to select the slot through which to transmit its bandwidth request. In order to avoid collision, they use contention resolution algorithm to decrease the collision probability. If no grant has been received during the specified timeout duration, we could assume that the transmission was unsuccessful.

Based on Figure 2.5, we could discuss the procedure in more detail. Just like unicast polling, SS must be allocated the slots first when SS’s polling interval was expired, as shown in Figure 2.6.

Figure 2.5 Multicast and Broadcast Polling

After the SSs have been allocated a uplink duration, they could participate the contention period to get the transmission opportunity. If a certain SS has already obtained the slot and transmitted the request successfully, the BS would check the request in accordance with

admission control scheme. Otherwise, SSs must wait for backoff time until it gets TXOP. The BW request message must include the SS ID, connection ID, and the requested bandwidth as shown in Figure 2.7. Last, the data transmission could be transmitted as illustrated in Figure 2.8.

Figure 2.6 Multicast or Broadcast Polling BW allocation for BR

Figure 2.7 Multicast or Broadcast Polling BW Request

Figure 2.8 Multicast or Broadcast Polling BW allocation for data

Then, we briefly introduce how the contention resolution algorithm be implemented. The BS has more flexibility in controlling the contention resolution. The mandatory method of contention resolution that should be supported is based on a truncated binary exponential backoff, with the initial backoff window and the maximum backoff window controlled by the BS. The values are specified as part of the uplink channel descriptor (UCD) message which represents a power-of-two value. For example, a value of 4 indicates a window between 0 and 15; a value of 10 indicates a window between 0 and 1023.