Forum

I know in a ps poll frame the duration/id field is used to transmit the association identity of the sender. How are these association identities allocated? Does the AP assign them? Does the host NIC use a part of its MAC address? I know each one must be unique so it seems like they must be assigned by a central authority like the AP.

Thanks,
Brad

The Association ID is distributed by the AP to the STA during association, in the Association Response (or ReAssociation Response) frame.

It does not use any part of the MAC for allocation.

Test questions for you (and anyone else):

Why use an AID to uniquely identify each STA when there is already a MAC address for that?

What is the value range of an AID?

GT

>> Why use an AID to uniquely identify each STA when there is already a MAC address for that?

Fewer bits?

>> What is the value range of an AID?

Since it's 14 bits, I'm guessing 0 - 16383?

Anyone? Bueller?

Fewer bits is correct!

What frames carry an AID that would benefit from the fewer bits (16 vs 48)?

You are correct that it is 14 bits for the AID (the two most significant are set to 1 1 indicating an AID. However, the entire range of those 14 bits is not used for an AID. AID's have a possible value of 1 - 2007. 2008 - 16,383 are currently reserved.


GT

Thanks GT!

Hi bmbreer

???¡é?¡é?????¡­?¡°Anyone? Bueller????¡é?¡é????????

That was funny!! Bye the way, did you know that the speaker of that phrase [ Ben Stein ] was a political speech writer for Nixon ?

http://en.wikipedia.org/wiki/Ben_Stein

As Grant was saying, there can be up to 2008 unique Association Id???¡é?¡é?????¡é???¡és [ AID 0 being reserved for broadcast/multicast ]. That could possibly mean up to 251 bytes [ ignoring practical limits of numbers on AP???¡é?¡é?????¡é???¡és ]. 251 bytes is a fair amount of space, considering the fact that there may be a bunch of stations who do not have any data buffered.

If you do a protocol decode of a Tim element, you will not [ as far as I have seen ] be able to see the ???¡é?¡é?????¡­?¡°layout???¡é?¡é???????? of the entire true bitmap. In other words, AID 1, 2, 3???¡é?¡é???????|.2007. To save space, ???¡é?¡é?????¡­?¡°they???¡é?¡é???????? came up with a clever idea to ???¡é?¡é?????¡­?¡°compress???¡é?¡é???????? the true bit map to what is called a ???¡é?¡é?????¡­?¡°virtual bitmap???¡é?¡é????????. It???¡é?¡é?????¡é???¡és a little bit like in data and video compression where you have 010000000000000???¡é?¡é???????|???¡é?¡é???????| and instead of sending the whole bunch of zeros [ which takes up bandwidth ], instead, you insert a code that shows how many zeros in a row have been sent.

In a typical decode, you will see three important pieces of information:

1. The value of the length field
2. The value of the bitmap offset field
3. the value of the partial virtual bitmap field.

We start off with the full TIM virtual bit map which has two rows:

Top row gives the number of the AID [ 1-2007 ] and the bottom row is the ???¡é?¡é?????¡­?¡°flag???¡é?¡é???????? row. Whenever there is data for a station, the flag is ???¡é?¡é?????¡­?¡°1???¡é?¡é???????? and when there is no data, the flag is zero.

e.g.

AID 1 2 3 4 5 ???¡é?¡é???????|???¡é?¡é???????|???¡é?¡é???????|. 16 17 18 19 20 21 ???¡é?¡é???????|???¡é?¡é???????|???¡é?¡é???????|2007

FLAG 0 0 0 0 0 ???¡é?¡é???????|???¡é?¡é???????|???¡é?¡é???????|. 1 1 0 0 0 0 ???¡é?¡é???????|???¡é?¡é???????|???¡é?¡é???????|. 0

In this example, we only have data pending for AID???¡é?¡é?????¡é???¡és 16 and 17 [ data buffered at the access point ].

The partial bitmap gets rid of all the zeros by combining them in a special way [ almost like compression ]. Using the 3 pieces of information above [ length field etc ], an algorithm in the STA works out what AID???¡é?¡é????????s have zeros and what AID???¡é?¡é????????S have one???¡é?¡é?????¡é???¡és.

If you can dig up a copy of the IEEE802.11-2007 standard and go to Annex L, you will see an interesting little section called:


???¡é?¡é?????¡­?¡°An example of encoding a TIM virtual bit map???¡é?¡é????????

Over the last three months I???¡é?¡é?????¡é???¡éve read the whole standard and believe I should be punished for doing that.

For those who cannot sleep at night, some notes on how the algorithm works:

1. All zeros prior to the partial virtual bitmap [ PVP ] are summarized by a derived value called X.

2. All zeros subsequent [ after ] the partial virtual bitmap are summarized by a derived value called Y.

3. To calculate X and Y, you need to calculate two other values called N1 and N2.

4. N1 = bitmap offset*2

5. N2 = [ length ???¡é?¡é?????¡é?€?? 4 ] + N1

6. X = [ N1*8 ] ???¡é?¡é?????¡é?€?? 1

7. Y = [ N2 + 1] *8

Dave

For some more detailed information on the whole Bitmap procedure look up the following on Google:


???¡é?¡é?????¡­?¡°Efficient TIM element supporting multiple BSSIDs???¡é?¡é????????

By the way, don???¡é?¡é?????¡é???¡ét ever feel sorry for the IEEE 802.11 committees who have to make up all these ???¡é?¡é?????¡­?¡°orrible things to confuse us with. Take a look at where they proposed to hold meetings in 2002 !!

1.7. Future Meetings
1.7.1. May ???¡é?¡é?????¡é?€?? Sydney, AU
1.7.2. July ???¡é?¡é?????¡é?€?? Vancouver, BC
1.7.3. September ???¡é?¡é?????¡é?€?? Hyatt Monterey, CA
1.7.4. November, Koloa, Hawaii
1.7.5. January 2003 ???¡é?¡é?????¡é?€?? Ft Lauderdale, FL
1.7.5.1. Straw Poll ???¡é?¡é?????¡é?€?? In favor of Ft Lauderdale: 117:0:5
1.7.5.2. Any volunteers for alternate venues? No
1.7.6. May 2003 ???¡é?¡é?????¡é?€?? Sophia Antipolis, FR
1.7.6.1. Any volunteers for other venues? No
1.7.6.2. Show of hands for interest in France: 95:8:6
1.8. Financial summary
1.8.1. Will be presented on Wednesday.
1.9. ExCom

Huge shock ???¡é?¡é?????¡é?€?? they voted unanimously for the above meeting locations???¡é?¡é???????|.


:'(

It's just not fair. They get to go to exotic locations [ yes Devin, in my books after a childhood of Scottish summers Fort Lauderdale is exotic ! ]and think up ways to torture us with..... ah well...

Dave

I look at that list and see

Sydney: 22 hours hours in a plane.
Vancouver: 7 hours in a plane.
Hawaii: 11 hours in a plane.
France: 8-9 hours on a plane
Monterey: 6 hours on a plane
Fort Lauderdale: ummm I don't think so.

But then people tell me I tend to be a bit crotchity.