Monday, August 2, 2010

[Geopriv] 802.11 measurements

Gabor has provided some excellent feedback on the measurement definition in the -measurements draft.

I plan to make the following changes, with the possible exception of one final item, which I will explain below.

1. Change wap to ap in line with 802.11 terminology.

2. Add location, as reported by the AP.

3. Add a "verified" tag to the BSSID. It is possible for a device to gain an assurance that the BSSID is correct. This Boolean attribute simply labels the BSSID as good. This doesn't help with the spoofing problem, but it can be used to filter bad data. In particular, this helps with the attack that has an attacker deploy a number of access points in order to alter the measurements that a device gets. (Come to think of it, this sort of attack probably needs text too.)

4. Remove round trip delay, which is difficult (or nigh impossible) to measure when scheduling delays are considered. The approach that is being taken in 802.11 is to synchronize clocks and measure the one way flight time. Either node can measure this value, but since it's the same underlying value, there will only be a single flight time (as opposed to an rtd in both directions).

5. The potentially controversial suggestion is the removal of channel. It's possible that this value changes over time, rendering it effectively useless. However, in a lot of cases, the channel remains static. Therefore, I am inclined to retain channel, even though it might not provide any assistance in some deployments.

--Martin
_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv