From: Joe (joe.clyde@utah.edu)
Date: Wed Nov 26 2008 - 15:42:39 ARST
I have a couple of questions regarding QOS and matching "voice"
traffic. I will ask the more specific ones but will also throw out some more
general ones for the group as well. I should note that this question came
about while doing an Internet Expert Lab, Vol 2, lab 10 Section7.2 but
think the question is fairly general and hopefully is useful for the actual
lab.
First, in NBAR, what exactly is "match protocol vofr"? Obviously vofr is
voice over frame relay, but how is that traffic actually defined? Is it
simply a UDP range of ports (16384 - 32767) traversing a frame-relay link,
or is there some other type of identifier.
That leads to my second question, on matching "voice" traffic, what is
the difference, if any, between let's say an ACL that matches
"permit udp any any range 16384 - 32767"
or
"match ip rtp 16384 16383"
or
"match protocol rtp audio"
Or...you get the idea, multiple way.
Ultimately there a quite a few ways to match voice, and many of the
practice labs I've done (not just IE) generically ask for "voice" and
invariably I'll answer it one way while the lab guide/ proctor guide will
answer it another. Further more, something like "vofr" would seem to make
sense for something like frame-relay fragmentation...or seems more specific
anyway for a QOS policy dealing specifically with a frame-relay cloud...but
then again I don't really know what constitutes "vofr".
I am probably over thinking this, but if there is a "best practice" for
matching "voice" traffic, I would love to know. Generally I prefer using
NBAR, where allowed, over say, ACL's (less chance of type-O's, or calling
the wrong ACL on accident, et) but is there a difference? If so, what are
key words or scenarios that would lead you to one over another? Thanks for
the help guys.
Blogs and organic groups at http://www.ccie.net
This archive was generated by hypermail 2.1.4 : Mon Dec 01 2008 - 08:18:32 ARST