[ns] 802.11 Carrier sense

[shshah@cairo.cs.uiuc.edu]

On Thu, 18 Apr 2002, Sorav Bansal wrote:

> 
> 
> Hi,
>     I do not think that should happen though. The is_idle function should
> return false at that point if another node in the vicinity has started
> transmitting before my backoff period ends.
>     Also I could not understand, which rate are you increasing and why it
> should make the is_idle function return false at that point.

I tried to increase the CBR sending rate of all the nodes. (Decreased the 
interval_ parameter.) This is so that I could ensure that another node in 
the vicinity IS transmitting before my backoff period ends. All the nodes 
in my topology can hear each other. The transmission range is the default 
(250m?) and my nodes are all within 100m of each other.

Therefore, I tried to increase the sending rate to try to ensure that
is_idle() DOES become false, but however much I increase the rate, 
is_idle() never becomes false... Maybe I'm doing something trivially wrong 
which is screwing things up.

-Samarth.

 > 
> --sorav
> 
> 
> 
>                                                                                                 
>                     shshah@cairo.c                                                              
>                     s.uiuc.edu           To:     Sorav Bansal/India/Contr/IBM@IBMIN             
>                     Sent by:             cc:     ns-users@ISI.EDU, <ns-users-admin@ISI.EDU>     
>                     ns-users-admin       Subject:     Re: [ns] 802.11 Carrier sense             
>                     @ISI.EDU                                                                    
>                                                                                                 
>                                                                                                 
>                     04/17/02 12:50                                                              
>                     AM                                                                          
>                     Please respond                                                              
>                     to shshah                                                                   
>                                                                                                 
>                                                                                                 
> 
> 
> 
> 
> On Tue, 16 Apr 2002, Sorav Bansal wrote:
> 
> >
> >
> > The call to is_idle in the function check_pktRTS() is the call to is_idle
> > before an RTS is to be dispatched.
> 
> Thanks. I'm unable to generate a scenario, however, in which the 'if (!
> is_idle())' check turns out to be true. The channel always seems to be
> idle and the condition always seems to be false, however high a rate I
> send at.
> 
> Thus, if I set the backoff interval differently (not using inc_cw() as is
> done in check_pktRTS(), but using a function of my own), this code is not
> executed at all since the if (! is_idle()) is always false. Hence, my
> purpose is still not served. Can you give me any input on this?
> 
> Thanks and regards,
> Samarth.
> 
> >
> > --sorav
> >
> >
> >
> >
> 
> >                     shshah@cairo.c
> 
> >                     s.uiuc.edu           To:     Sorav
> Bansal/India/Contr/IBM@IBMIN
> >                                          cc:     ns-users@ISI.EDU,
> <ns-users-admin@ISI.EDU>
> >                     04/15/02 09:33       Subject:     Re: [ns] 802.11
> Carrier sense
> >                     PM
> 
> >                     Please respond
> 
> >                     to shshah
> 
> >
> 
> >
> 
> >
> >
> >
> > On Mon, 15 Apr 2002, Sorav Bansal wrote:
> >
> > >
> > >
> > > Hi,
> > >     The function is_idle() in mac-802_11.cc checks if the medium is
> idle
> > or
> > > not (carrier sensing).
> >
> > Yes, I am aware of this, but I would like to know which particular call
> to
> > is_idle() is the *first* carrier sense before the *first* RTS is sent.
> > is_idle() is checked multiple times in mac-802_11.cc, I am only
> interested
> > in the particular above case. Could you give me this info? Thanks in
> > advance.
> >
> > Regards,
> > Samarth.
> >
> > The medium is not idle if a packet is being
> > > received. And all packets with signal strength above the carrier sense
> > > threshold are received by the MAC layer (except that the error field is
> > set
> > > to 1 for packet below the receiving threshold). Hope that answers your
> > > query.
> > > --sorav
> > >
> > >
> > >
> > >
> >
> > >                     shshah@cairo.c
> >
> > >                     s.uiuc.edu           To:     ns-users@ISI.EDU
> >
> > >                     Sent by:             cc:
> >
> > >                     ns-users-admin       Subject:     [ns] 802.11
> Carrier
> > sense
> > >                     @ISI.EDU
> >
> > >
> >
> > >
> >
> > >                     04/15/02 01:12
> >
> > >                     PM
> >
> > >                     Please respond
> >
> > >                     to shshah
> >
> > >
> >
> > >
> >
> > >
> > >
> > >
> > >
> > > Hi,
> > >
> > > I'm unable to discern exactly when (i.e. where in the 802.11 code) the
> > > first carrier sense before the first RTS for a packet occurs. If the
> > > carrier is found busy at this carrier sense, I would like to be able to
> > > modify the backoff interval after the deferral that follows the busy
> > > carrier sense. Rather than have the station backoff cw_ slot times, I'd
> > > like to have it backoff some other number of slot times.
> > >
> > > I do not wish to modify the backoff procedure that is followed after a
> > > collision, however. I want the cw_ doubling, backoff = cw_ number of
> > > slots, etc. to remain as is, in the case of collision.
> > >
> > > Could anyone kindly provide me some guidance? Thanks in advance.
> > >
> > > Regards,
> > > Samarth.
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> >
> >
> >
> >
> >
> 
> 
> 
> 
> 
> 
> 
> 
>