Re: voice based queuing

Linux Advanced Routing and Traffic Control

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



ganesh kumar godavari wrote:
> hello group,
> 
>    i want to know if there is any way i can determine using iptables if 
> the ip packet contains voice?
> 
>  i want to know this as i want to do some queuing for output packets and 
> the voice packets are given high preference next ftp,telnet,ssh.....
> 
> 
> i want to know if this is possible using iptables and tc. if so how. if 
> i can identify the packet to be voice then i can do the rest using tc.

I've followed (from afar, things are busy) your quest.

The issues are several.

1.  The ports used, except for setup, are dynamic, as the Microsoft page 
told you (although NetMeeting is hardly a good example of a compliant 
H.323 app).

2.  You can do stateful inspection of headers and have a good shot at 
finding RTP and RTCP packets, and then mark those.

Where you probably need to go is toward marking all from the H.323 
hosts.  You will need to spend some time with TOS/DiffServ, and 802.1p.

Also, be aware that if you do not have priority queuing from endpoint to 
endpoint (also known as E2E) you will not see an appreciable improvement 
in handling.

We are engaged in a couple of efforts to identify Voice and Video Over 
IP, and mark it for priority queuing on our campus.  Cisco has a nice 
simple strategy they're apparently trying to make a "standard" but it's 
at odds with diffserv marking.  If you go with the Cisco markings, and 
you enable your queue disciplines based on the markings, 
_throughout_the_entire_network_ then you may see some improvement.

Can you di this with LARTC, yes, I believe so.  Sorry I cannot be more 
certain, but while it's on my list to test, I'm still a bit busy...  The 
trick is getting the appropriate packets marked properly at the edge, 
rather than trying to mark them at the core.  And why, one might ask, is 
this so?  Well, we see a lot of congestion at the edge in wiring closet 
aggregation switches.  If these switches cannot handle priority queuing 
AND they experience periods (even short ones) of congestion, then most 
of the effort you're going to is for naught.

Gerry Creager
network Engineering
AATLT, Texas A&M University
College Station, Texas


_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

[Index of Archives]     [LARTC Home Page]     [Netfilter]     [Netfilter Development]     [Network Development]     [Bugtraq]     [GCC Help]     [Yosemite News]     [Linux Kernel]     [Fedora Users]
  Powered by Linux