From: Scott F. Robohn (sfr@xxxxxxxx)
Date: Tue May 02 2000 - 23:15:54 GMT-3
Mike,
First, good luck next week! We're pulling for you.
On CLIP: Technically, the CLIP-with-PVC's reference on the
doc cd is a bit misleading, IMHO.
RFC1577 (updated in RFC2225) defines a couple of things: (a)
the ATMARPServer mechanism and (b) ATM Inverse ARP.
'Full-blown CLIP' uses SVCs. You use the ATMARPServer to
resolve an IP address to an ATM NSAP address. Then you use
signaling to establish an SVC to that ATM destination
address.
ATM Inverse ARP is commonly used for PVCs so that IP
addresses can be dynamically learned instead of using
map-lists. But there's no ATMARPServer involved here, so
it's not really CLIP.
So that's what you get from a guy who's taught CATM one too
many times so far this year ;)
Back to something that DOESN'T involve ATM,
Scott
Mike@ccci.com wrote:
>
> Joe,
>
> Good example of CLIP w/SVCs. BTW, according to the Cisco docs CLIP can be
> used with PVCs and inverse ARP. The configuration is a bit different than
> the one you provided of course.
>
> Details Here:
> http://127.0.0.1:8080/cc/td/doc/product/software/ios113ed/113ed_cr/wan_c/wca
> tmnpm.htm#xtocid587824
>
> Note that even though works with ATM Inverse ARP (i.e., a dynamic L3 to L2
> address mapping) , other protocols, most notably AppleTalk, do not and
> require a static mapping via the map-list and map-group commands.
>
> HTH,
>
> Mike McSpedon
> Halifax Next Week...
>
> At 01:54 PM 4/29/00 -0700, Joe Martin wrote:
> >John,
> > Doing PVCs is not CLIP. CLIP uses SVCs. You will need to define two
> >PVCs though - One for UNI signalling(qsaal) and one for ILMI NSAP
> >registration(ilmi). You can do pure PVCs without a LS1010, but you need one
> >to do CLIP or LANE to gain the SVC functionality. The ESI addresses you see
> >are arbitrary but unique numbers(ex. - R1 use something like
> >1111.1111.1111.00 and R2 use 2222.2222.2222.00). The interface for a LS1010
> >or LS1010 blade uses IOS so it looks just like a router. You access a
> >LS1010 blade from a 5500 by doing a show module and typing "session x" where
> >x=module of the LS1010. This also holds true for things like RSMs and
> >MFSCs. Heres a CLIP sample with some notes to work from:
> >
> >Int ATM x/x
> >Atm pvc 1 0 5 qsaal <--turns on UNI 3.x signalling
> >Atm pvc 2 0 16 ilmi <--turns on ILMI
> >Shut/no shut <--doing this on the interfaces forces a restart of ilmi and
> >signalling. Give it a minute.
> >Show atm ilmi-status <-- cut the prefix(highlight/edit/cut) The prefix is
> >everything up to the ESI - The first
> > 13 bytes.
> >Int ATM x/x
> >Atm esi-address 1111.1111.1111.00 <--remember to make unique per client and
> >server
> >Atm arp-server self <-- put on arp server. It doesn't matter which router is
> >the arp server
> >Atm arp-server nsap xx.xxxxxxxxxxxxxxxxxxxxxxxxxx.1111.1111.1111.00 - prefix
> >from cut <--Put on CLIP
> >
> >clients. This assumes that R1
> >
> >is the arp server since the esi
> >
> >used is 1111.1111.1111.00
> >Ip address x.x.x.x y.y.y.y <--put each client in the same subnet.
> >
> >Hope this helps,
> > JOE
> >
> >----- Original Message -----
> >From: John Conzone
> >To: ccielab
> >Sent: Saturday, April 29, 2000 12:51 PM
> >Subject: ATM PVC's. Lightstream Switches
> >
> >
> > Oaky, some basic ATM questions that I hope some of you can clear up.
> > I've configured router ATM interfaces using map lists pointing to
> >another IP device. PVC's. Is the considred CLIP. I'm thinking no, cause I
> >saw you needed to configure an ARP server for CLIP. In frame relay, the
> >switch needs to know how to route DCLI, like when you set up a router as a
> >Frame Switch. Do you need to set up the LightStream module the same way if
> >you use it to connect two or more ATM devices?
> > What is the Light Stream interface look like. How do you acces a Light
> >Stream module in a 5000? NASP addressing. Do you get one automatically from
> >the switch if you use ILIMI? I see in some configs hard coded ESI
> >assignments.
> > I keep reading document after document, including Caslow, but can't see
m
> >to cleat up the confusion!
> >
> >
> >
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:23:27 GMT-3