From: Jack Heney (jheneyccie@xxxxxxxxxxx)
Date: Tue Oct 03 2000 - 20:34:10 GMT-3
Thanks for the explanation (it certainly wasn't too crappy)...I'm sorry to
keep firing questions off, but the more I learn, the more questions it
creates....Three more quick things....1.) Does the term MAC frame refer to a
frame that is 802.5 without an LLC header, or is it some sort of ethernet
framing (802.3 or ethertype or something)?
2.)
>LLC broadcast frames are free the traverse the entire bridged domain
>whereas MAC frames are not allowed to leave the ring (or switchport if your
>a switch)
Why would a switch filter a broadcast? I've always interpreted "switch" to
represent a device that forwards broadcast and unkown frames.
3.) I understand your explanation about frames without a RIF no being
forwarded to other trcrfs if you are using SRB at the trcrf level, but this
is the default behavior of the 3900 (SRB at the trcrf level, not
SRT)....What types of frames would be generated without RIF's? Are these
the MAC frames you are talking about, or is the LLC/MAC distinction
completely separate from the RIF/no RIF distinction?
Thanks again for all your help...Jack
>From: "Steve McNutt" <lpd@jacksonville.net>
>Reply-To: "Steve McNutt" <lpd@jacksonville.net>
>To: <ccielab@groupstudy.com>
>Subject: RE: Token Ring Vlans
>Date: Tue, 3 Oct 2000 19:02:38 -0400
>
>I know this is going to be a crappy explaination. token ring is a bith to
>explain, and token ring switching is bitch*2 hehe.
>
>In case you don't know this, there are two kind of frames in token ring.
>MAC frames and LLC frames. MAC frames perform ring managment functions.
>LLC frames carry user data. the Token Ring state machine makes extensive
>use of MAC frames to do it's thing, and the amount of MAC traffic you have
>increases with the number of hosts on the ring. A ring with 200 hosts has
>quite a bit of mac broadcast traffic being generated, esp when people boot
>up their machines in the morning.
>
>now to the answers:
>
>1. LLC broadcast frames are free the traverse the entire bridged domain
>whereas MAC frames are not allowed to leave the ring (or switchport if your
>a switch). It's possible to have LLC broadcast frames that are constrained
>to a single TRCRF by the switch (using a functional group for multicast
>comes to mind), but normally you don't see that kind of thing. Of course
>you can have more LLC broadcast traffic than MAC broadcast traffic on a
>ring
>at any given point in time. I was just trying to point out that multiple
>TRCRFs on a TRBRF don't really buy you anything performance-wise because
>the
>MAC traffic is constrained at the port level, and most LLC traffic
>traverses
>ALL trcrf's in the trbrf VLAN.
>
>2. with srt, LLC broadcasts will span all of your trcrf's. with srb at
>the
>trcrf level, I do belive (although I've never tried it) that frames without
>a RIF will not be forwarded to any other trcrfs. Honestly I don't think
>this would be very good practice. Can't think of a good reason not to run
>SRT on all of your token ring vlans. If you want to constrain LLC
>broadcast
>traffic, use a router.
>
>-----Original Message-----
>From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
>Jack Heney
>Sent: Tuesday, October 03, 2000 5:40 PM
>To: Steve.McNutt@ahlcorp.com; ccielab@groupstudy.com
>Subject: RE: Token Ring Vlans
>
>
>Okay...It's becoming clearer but I have 2 more questions....Why is the
>majority of your ring level broadcast traffic MAC frames? Also, It says
>that you can configure the 3900 for either SRB or SRT between the trcrf and
>the trbrf...If I choose SRT for two trcrf's that share a parent trbrf, does
>that mean that these broadcasts will span multiple trcrf's...For that
>matter, if I choose SRB for both trcrf's, will that prevent the broadcasts
>from being bridged from one trcrf to another through the trbrf?
>Thanks again,
>Jack
>
>
> >From: "McNutt, Steve" <Steve.McNutt@ahlcorp.com>
> >Reply-To: "McNutt, Steve" <Steve.McNutt@ahlcorp.com>
> >To: "'ccielab@groupstudy.com'" <ccielab@groupstudy.com>
> >Subject: RE: Token Ring Vlans
> >Date: Tue, 3 Oct 2000 15:48:22 -0400
> >
> >Hey, that was a pretty good explanation.. except for the Cisco
> >markettecture
> >about the usefulness of TRCRFs. ;-)
> >
> >In reality, MAC frames (which are comprise the bulk of your ring level
> >broadcast traffic) do not span the TRCRF or even enter the backplane of
>the
> >switch (switchports in the same trcrf are transparently bridged). the
> >microsegmentation effect occurs regardless of whether two switchports are
> >on
> >the same or different trcrfs.
> >
> >My experience has been that you use multiple TRCRFs because TRCRFs cannot
> >span switches(with the execption of the default trcrf), whereas TRBRFs
>can.
> >So if you have multiple switches that you want to be on the same ip
>subnet,
> >you need to create a TRCRF for each switch and then assign it to the same
> >parent TRBRF.
> >
> >
> >-----Original Message-----
> >From: Asbjorn Hojmark [mailto:Asbjorn@Hojmark.ORG]
> >Sent: Tuesday, October 03, 2000 2:57 PM
> >To: 'Jack Heney'
> >Cc: ccielab@groupstudy.com
> >Subject: RE: Token Ring Vlans
> >
> > > If so, what advantage is gained by creating these vlans?
> >
> >It's like micro-segmentation for token-ring. Remember that
> >stations on a ring share a token and that some frames are
> >local to a ring while others are global for the bridged
> >domain.
> >
> > > Also, can rings that don't share the same parent trbrf
> > > communicate without some sort of routing occurring
> >
> >No.
> >
> >HTH,
> >-A
> >--
> >Heroes: Vint Cerf & Bob Kahn, Leonard Kleinrock, Robert Metcalfe
> >Links : http://www.hojmark.org/networking/
> >
> >
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:25:23 GMT-3