Re: Re: [ns] Problem about the implemention of mac802_11

[Joe Macker <macker@itd.nrl.navy.mil>]

At 11:39 AM 8/29/2001 +0800, ��Ӿ wrote:
>I have just examined the code of mac_802_11.cc and find that
>the implementation of ns is somewhat reasonable. I have mistaken
>that there are a backoff process after difs. Besides, In IEEE802.11
>standard 9.2,it says that "The CSMA/CA protocol is designed to 
>reduce the collision probably ... Just after the medium becomes
> idle following a busy medium...This is the situation that 
>necessitates a random backoff procedure ...". When the medium is
>idle for some time, it's not necessary to backoff. 

When I read my 1999 version of the spec, its actually not that simple.  

There are rules (even for first time transmissions) that define if the backoff procedure is to be used after DIFS. Although not always, the backoff process is heavily applied in DCF after DIFS.

I am not implying anything about the implemented code, just that there is often a backoff process even after DIFS deferral and its important. Read 9.2.4.