[c-nsp] 2950 Questions

[Paul Stewart]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Thanks... that's the info I was *really* looking for...:)  While our
intent is plain vanilla it's good to know about the stuff mentioned
below...

Paul


Andre Beck wrote:
| On Thu, Mar 24, 2005 at 01:12:01AM +1000, Virgil wrote:
|
|>On 24/3/05 12:33 AM, "Justin M. Streiner" <streiner at cluebyfour.org> wrote:
|>
|>
|>>The switch doesn't know what SIP is, and it doesn't care - it just
|>>forwards packets :-)
|>
|>Being pedantic and all, but a switch (layer 2 device) forwards *frames*.
|>Routers (layer 3 device) forward packets.
|
|
| Then again, the 2950 has "mls qos" capabilities. I'm not currently
| up to date on which subset exactly, but that could well mean:
|
| * The switch could access 802.1Q/D frame priority information (aka
|   Ethernet CoS) and take it into consideration in queuing decisions
| * The "switch" could access the TOS field of IPv4 packet carrying
|   frames to get alternate CoS information (either pure IP precedence
|   or even DSCP)
| * The "switch" could classify IPv4 packet bearing frames on a subset
|   of extended IP access list semantics to further control QoS decisions
| * The "switch" could even *manipulate* the TOS field of said IPv4
|   bearing frames, enforcing some policy or trust scheme. For instance,
|   a 29xx/355x with "mls qos" (default is "no mls qos") but nothing else
|   configured will *zero* the DSCP.
|
| So yes, it's basically a frame bouncing device, but it has bells and
| whistles that can take IP into account, primarily for QoS reasons.
| With the protocol of interest beeing SIP, having a look on that stuff
| might not be the worst of ideas. As long as "mls qos" is not engaged,
| it will stay just a bridge^Wswitch, though.
|
| Andre.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (MingW32)

iD8DBQFCQuD9qMetgU57IuQRApuwAJ4s5QEY+xGdTIvW6o2/UxbET7mBSwCfaVnn
k3RM+vxWiOxK+muYh9iOJAc=
=+rWv
-----END PGP SIGNATURE-----