From: Tom Lijnse (Tom.Lijnse@globalknowledge.nl)
Date: Wed Apr 21 2004 - 12:50:31 GMT-3
Hi Larry,
I noticed this thread fairly late, because I have taken a short vacation last week, so you may have already figured it out by now. I still thought the question was worth answering, since some of the documentation is fairly confusing on this issue.
The way I see it "RFC 2684: Multiprotocol encapsulation over AAL5" defines just that: How to transport IP Packets, IPX packets, Appletalk packets, etc or even bridged frames across ATM PVCs and SVCs by encapsulating them in ATM cells.
What this RFC does not specify is how a device should figure out to which PVC (VPI/VCI) or SVC (NSAP) to send its packets. The "Layer 3 to Layer 2 mapping" so to speak. The obvious solution would be to simply manually map IP next-hop addresses, IPX next-hop addresses and so on to the correct VPI/VCI pair for PVCs or NSAP addresses for SVCs. (Just like you can manually map IP, IPX, etc to DLCIs for Frame-Relay).
"RFC 2225: Classical IP and ARP over ATM" is an addition to RFC 2684 and defines methods to establish IP to VPI/VCI or IP to NSAP mappings automatically. For PVCs it defines inverse-arp as a way to map IP next-hop addresses to VPI/VCI pairs dynamically (just like Frame-Relay inverse-arp maps IP next-hop to DLCIs). For SVCs it defines how you can assign one of the devices the role of ARP-SERVER which serves as a repository for IP to NSAP mappings. All other devices register their IP/NSAP combination with the ARP server.
Very often configurations with static mappings are presented as "pure" RFC2684/RFC1483 configurations, while configurations using inverse-arp or arp-server are presented as RFC2225/RFC1577 configurations (even though these configurations also use the RFC2684/RFC1483 encapsulation method).
As for when to use manual versus dynamic mappings it really depends on the wording of the question. If the question leaves it up to you it's basically a trade-off between stability and ease-of-use/scalability. This is similar to the decision to use manual mapping versus inverse-arp on Frame-Relay or fixing port speed/duplex versus autonegotiation on ethernet.
Hope this helps,
Regards,
Tom Lijnse
CCIE #11031
Global Knowledge Netherlands
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
Larry Roberts
Sent: Friday, April 16, 2004 10:11 PM
To: security@groupstudy.com; ccielab@groupstudy.com
Subject: Struggling with ATM
I'm hoping that someone can help clear things up for me.
I'm reviewing the ATM connectivity chapter (7) of the CCIE Practical
Studies: Security book.
I'm struggling to find the differences between RFC 2684:Multiprotocol encap
over AAL5 vs RFC 2225: Classical IP and ARP over ATM.
In an IP only world ( of which the security lab is one) I don't see any
difference in the configuration except for the inclusion of an inarp timeout
value in the latter. ( at least in the book )
I'm trying to decipher how they could phrase a question that would lead you
to choose one over the other.
Can someone point out the differences that I am obviously missing so I can
come up with a decision matrix?
I don't want answers to the CCIE lab, but rather what lead you to choose one
method over the other?
Right now the only thing I can think of is PVC autodiscovery, and the
inclusion/exclusion of an ATMARP server.
Thanks
Larry
This archive was generated by hypermail 2.1.4 : Mon May 03 2004 - 19:48:51 GMT-3