Pre-scan for wireless channel selection

Abstract

A system and method for pre-scanning for a wireless channel selection is provided. The methodology is executed by a first fixed location wireless device to improve initial channel selection in a wireless network where a plurality of fixed location wireless devices is contemporaneously initialized. On a predetermined channel, there is a transmission of at least one message including an identifier of the first device. On the predetermined channel, there is a transmission of at least one message indicative of readiness to transition to a subsequent processing step. The methodology also commences scanning for channel selection purposes.

Description

FIELD OF THE INVENTION

This invention relates generally to the field of wireless communications, and more particularly to synchronization and transmission power management of fixed location wireless devices to facilitate channel selection.

BACKGROUND OF THE INVENTION

Initial channel selection by fixed location devices such as access points in a wireless network such as a WLAN can be problematic. Many access points are configured to attempt to begin operation on a particular channel, regardless of vendor. If there is a loss of power to the WLAN and all access point s initially attempt to begin operation on a single channel then there is likely to be interference and poor performance. Even if the WLAN is configured to migrate access points to other channels over time, such as under the control of a server, significant delay may be experienced before the WLAN settles into a relatively stable configuration.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 illustrates access points in a WLAN.

FIG. 2 is a state diagram that illustrates a pre-scan procedure for improving channel selection.

FIG. 3 illustrates a portion of a pre-scan frame.

DETAILED DESCRIPTION

Referring to FIG. 1, initial channel selection occurs when multiple access points (100-114) in a constellation, i.e., with at least some overlap in range, are contemporaneously initialized. This may occur because of power restoration following a failure, because of a switch or server reboot, or for other reasons. The invention is directed to improving channel selection by fixed wireless devices such as access points under such conditions. Because mobile stations do not become associated with access points until channel selection has completed, there are no stations depicted in FIG. 1.

As described in co-pending U.S. patent application Ser. No. 10/780,841, METHOD FOR SCANNING RADIO FREQUENCY CHANNELS, filed Feb. 18, 2004, which is incorporated by reference, initial channel selection includes a plurality of sequential steps (intervals) which are executed by the access points (100-114). From the perspective of access point (104), for example, in the scan interval the access point scans potential radio frequency channels and listens for messages from other access points on those radio frequency channels. Messages received from other access points during channel selection, e.g., access points (100, 102, 106, 108), are used by the access point (104) to construct a scan table having an entry for each access point from which a message is received during the scan interval. The access point also maintains a channel map having an entry for each of the radio frequency channels on which a message was received. Each entry contains the device ID of the access point that sent a message on the corresponding channel. The access point (104) sorts the channel map into a triplet channel map, wherein each successive group of three entries is associated with three successive channels, and wherein the average power is stored in the entry for each channel. The access point then selects, from the triplet channel map, the channel with the lowest center average power. The access point then transmits preclaim messages on the selected channel during a preclaim interval, and listens for messages on the selected channel. The scan table is updated based on each message received on the selected channel during the preclaim interval. The access point also calculates an adjacency vector sum which represents the sum of all average power levels on all channels. During a claim interval, the access point transmits claim messages on the selected channel. The claim messages including the adjacency vector sum. The access point also listens for, and records, messages from other access points on the selected channel, including adjacency vector sums. The access point maintains a claim table having an entry for each device ID that sent a message on the selected channel during the claim interval, and evaluates the claim table at the end of the claim interval. The evaluation includes the following steps:

• • (1) if the claim table has no entries, causing the access point to commence communications with other devices via the selected channel;
  • (2) if the claim table has entries, then checking to see if the selected channel was occupied at the beginning of the claim interval, and if the selected channel was not occupied at the beginning of the claim interval, causing the access point to return to scanning channels during a scan interval;
  • (3) if the selected channel was occupied at the beginning of the claim interval, checking to see if all the claim table entries contain power levels that are less than a power level that was recorded on the selected channel before the claim interval, and if all the claim table entries contain power levels that are less than the power level that was recorded on the selected channel before the claim interval, then causing the access point to commence communications with other devices via the selected channel;
  • (4) if any claim table entry contains a power level that is greater than the power level that was recorded on the selected channel before the claim interval, then comparing the adjacency vector sum to an adjacency vector that was received in one of the messages, and if the adjacency vector sum is greater than the adjacency vector received, causing the apparatus to commence communications with other devices via the selected channel, otherwise causing the apparatus to return to scanning channels during a scan interval.

The above-described procedure functions quite well where the access points are identical and the WLAN operates in “clean air,” i.e., without interference from other devices. However, if the access points are not identical then differences in reboot speed and software execution speed can result in asynchronous operation. This can happen because different access points utilize different operating systems and different hardware. In extreme cases, one access point may complete channel selection before another access point has even begun channel selection. Further, interference caused by other devices may degrade the exchange of messages on which the procedure is partially based. The result is often a non-ideal channel selection.

Referring to FIGS. 1 through 3, a pre-scan interval is employed to synchronize the access points (100-114) and facilitate power control. The pre-scan interval precedes the scan interval in the procedure described above. All of the access points begin the pre-scan interval in State 1 on a predetermined channel. Further operation will be described from the perspective of access point (104) for clarity. In State 1, access point (104) is operable to transmit pre-scan frames (300) at full power. The pre-scan frames include a device ID of the transmitting device, e.g., an AP ID such as a MAC address, an advertisement of scan cycle time, and a sync mark bit. The access point also listens for pre-scan frames transmitted by other access points. When a pre-scan frame is received, the received power level and AP ID are noted and used to construct a table. The access point normally remains in State 1 until a minimum number of frame samples have been received from each access point (100, 102, 106, 108) that has been heard, and a minimum pre-scan time has elapsed. For example, the access point (104) may remain in State 1 until twenty samples per access point are received, and at least X seconds have elapsed. However, in extreme circumstances where a max pre-scan time has elapsed, or where a countdown message is received from another access point, thereby indicating that the other access point has moved to State 3, the access point (104) proceeds directly to State 3. Otherwise, under the normal circumstances already described, the access point moves to State 2.

State 2 provides synchronization of transition of the access points between State 1 and State 3. When the access point is in State 2, it continues to transmit pre-scan frames. However, the pre-scan frames now have the sync-mark bit set in order to indicate that the access point is in State 2. The access point continues to listen for pre-scan frames from other access points to determine whether those access points have transitioned to State 2, i.e., by reading the sync-mark bit in the received frames. The access point transitions to State 3 when one of three conditions occurs. The first condition is when all detected access points have transmitted pre-scan frames with the sync-mark bit set, i.e., when all detected access points have signaled readiness to move to State 3. The second condition is when the max pre-scan time has elapsed. The third condition is when countdown messages are received from another access point, thereby indicating that some access points has moved to State 3.

State 3 provides further synchronization. In State 3 the access point begins a countdown to transition to the scan interval, e.g., a 50 ms timer. While counting down, the access point continues to transmit pre-scan frames. However, the pre-scan frames now advertise the state of the countdown timer, e.g., 10 ms until transition to scan interval. The access point continues to listen for pre-scan frames from other access points. If differences in the countdown state are detected, e.g., because of clock speed, the access points re-synchronize. Resynchronization can be accomplished by adopting the countdown state of the slowest clock, i.e., the access point advertising the greatest amount of time remaining until transition to the scan interval. As already discussed, access points in State 1 and State 2 that detect the countdown will immediately transition to State 3, thereby providing further synchronization. When the countdown has elapsed, e.g., when the timer reaches 0, the access point transitions to the scan interval and proceeds with channel selection as already discussed above.

The access points may be further synchronized on the basis of scan cycle time, as advertised in the pre-scan frame (300). As discussed, the scan interval follows the pre-scan interval. In particular, an access point executes pre-scan once, followed by one or more scans. The scan cycle time is the total time required for an access point to go through one channel selection cycle, i.e., from scan to preclaim to claim. Since different access points may be built using different platforms that differ in terms of CPU, OS etc., the different access points can have different scan cycle times. In order for access points that need more than one scan cycle to select a channel, a procedure is executed to cause the access points to synchronize at the end of each scan cycle so that all access points re-enter the scan state contemporaneously. As already described, access points advertise their scan cycle time in the pre-scan frames during the pre-scan interval. In order to synchronize on the basis of scan cycle time, each receiving access point is operative in response to the advertised scan cycle time received from other access points to compare the advertised scan cycle time with the scan cycle time of the receiving access point. If the scan cycle time of the receiving access point is longer than the advertised scan cycle time the receiving access point continues to use its own scan cycle time, and to advertise that scan cycle time in subsequent pre-scan frames. However, if the received scan cycle time is longer then the access point adopts the longer scan cycle time as its own, and advertises that longer scan cycle time in subsequent pre-scan frames. Therefore, all access points eventually converge on the longest scan cycle time, as can be verified by examining the pre-scan frames from all the access points.

While the invention is described through the above exemplary embodiments, it will be understood by those of ordinary skill in the art that modification to and variation of the illustrated embodiments may be made without departing from the inventive concepts herein disclosed. Moreover, while the preferred embodiments are described in connection with various illustrative structures, one skilled in the art will recognize that the system may be embodied using a variety of specific structures. Accordingly, the invention should not be viewed as limited except by the scope and spirit of the appended claims.

Claims

1. A method executed by a first fixed location wireless device for improving initial channel selection in a wireless network where a plurality of fixed location wireless devices are contemporaneously initialized, comprising the steps of: transmitting, on a predetermined channel, at least one message including an identifier of the first device;transmitting, on the predetermined channel, at least one message indicative of readiness to transition to a subsequent step, wherein transmitting at least one message indicative of readiness to transition to a subsequent step is not executed until (a) a minimum number of messages are received from each identified device, and (b) a predetermined minimum amount of time has elapsed, and further including: delaying transition to commencing scanning by a time indicated by a countdown timer;transmitting, on the predetermined channel, an indication of countdown timer state; andreceiving messages including device identifiers from other fixed location wireless devices; andcommencing scanning for channel selection purposes after the message including the identifier of the first device and the message indicative of readiness have been transmitted.

2. The method of claim 1 wherein delaying transition to commencing scanning by a time indicated by a countdown timer is executed immediately if either (a) a predetermined maximum amount of time has elapsed, or (b) another fixed location wireless device has started a countdown timer to commencing scanning.

3. The method of claim 1 wherein, following transmitting the at least one message indicative of readiness to transition to a subsequent step, delaying transition to commencing scanning by a time indicated by a countdown timer is executed immediately if either (a) each identified device indicates readiness to begin a countdown timer to commencing scanning, or (b) a predetermined maximum amount of time has elapsed, or (c) another fixed location wireless device has started a countdown timer to commencing scanning.

4. The method of claim 1 including the further step of adopting, as countdown timer state, a countdown timer state indicative of greatest amount of time remaining to transition to commencing scanning from countdown timer state advertisements from all detected fixed location wireless devices.

5. A first fixed location wireless device operable to facilitate initial channel selection in a wireless network where a plurality of fixed location wireless devices are contemporaneously initialized, comprising: a countdown timer for delaying transition to commencing scanning by a time indicated by the countdown timer;a transmitter operable to transmit, on a predetermined channel: at least one message including an identifier of the first device,an indication of countdown timer state, andat least one message indicative of readiness to transition to a subsequent step, wherein the transmitter does not transmit the at least one message indicative of readiness to transition to a subsequent step until (a) a minimum number of messages are received from each identified device, and (b) a predetermined minimum amount of time has elapsed; anda receiver operable to receive messages including device identifiers from other fixed location wireless devices, and commence scanning for channel selection purposes after the message including the identifier of the first device and the message indicative of readiness have been transmitted.

6. The device of claim 5 further including circuitry operable to trigger commencement of the subsequent step if either (a) a predetermined maximum amount of time has elapsed, or (b) another fixed location wireless device has started a countdown timer to commencing scanning.

7. The device of claim 5 further including circuitry operable following transmission of the at least one message indicative of readiness to transition to a subsequent step, to trigger commencement of the subsequent step if either if either (a) each identified device indicates readiness to begin a countdown timer to commencing scanning, or (b) a predetermined maximum amount of time has elapsed, or (c) another fixed location wireless device has started a countdown timer to commencing scanning.

8. The device of claim 5 further including circuitry operable to update the countdown timer state to be a countdown timer state indicative of greatest amount of time remaining to transition to commencing scanning from countdown timer state advertisements from all detected fixed location wireless devices.