Re: RE: AD request / L2 Triggers Charter Statement/ Updated Internet-Draft on Fast Router Discovery

"Alper E. YEGIN" <[email protected]> Mon, 1 Jul 2002 16:25:24 -0700
Newsgroups gmane.ietf.pilc
Message-ID <027f01c22156$941bd030$8b6015ac@AlperVAIO>
Jihoon,

> 1. You said, "this protocol is limited to nodes on the same IP subnet.
Therefore protocol messages would not normally go beyond routers."
> Is this limitation necessary? In some case, L2 trigger message needs to go
beyond one IP subnet. For example, old AP and new AP may be in different IP
subnet.
>
> +------+                   +------+
>  |  AR   | ------------  |  AR   |
> +------+                   +------+
>     |                                 |
>     |                                 |
> +------+                   +------+
>  |  AP   |                     |  AP   |
> +------+                   +------+
>      |
>      |
> +------+
>  |  MN  | ---------->
> +------+

In the current usage scenarios of the L2 triggers, we have only
identified a need to communicate these triggers from APs to
ARs on the same link. If you can identify a need to notify
ARs on other links, then this is worth considering.


>
> In IAPP protocol, APs in different IP subnet can exchange UDP messages. Is
it necessary to limit this protocol within just one IP subnet? It may be
better to relax this restriction.
>
> 2. In I-D, you indicated a carrier is meant, like defining a protocol
"based on" UDP".
>
> Is it only for IPv4 or does this include IPv6 too? In IPv6, it may be
better to use ICMP or Extension Header to carry L2 trigger. For example, in
Mobile IPv4, Binding update is using UDP. But in Mobile IPv6, Binding Update
is using Mobility Header. Would you explain why you choose UDP instead of
ICMP or Extension Header?
>

Actually it doesn't matter whether UDP is carried over IPv4 or IPv6.
Either is fine.

I think this is a simple client-server protocol. Some relevant information
is generated on a node, and consumer is on another IP node. I don't think
this justifies itself for a new ICMP type.. A UDP port should be sufficient.
Since we are not interested in piggybacking this data with existing traffic
from AP to ARs, I think extensions headers are not needed either...

UDP is the most straightforward choice, and I don't see a need
to use ICMP or extension headers... Do you?

alper

**********************Confidentiality Note**********************
Privileged/Confidential Information may be contained in this e-mail and any
attachment to it and may be covered by existing non-disclosure or
confidentiality agreements.
If you are not the addressee (or authorized to receive for the addressee),
you may not use, copy or disclose to anyone any information contained in
this e-mail.  If you have received this e-mail in error, please notify the
sender immediately by reply e-mail and delete it from your system.
Thank you very much.
*************************************************************



_______________________________________________
pilc mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/pilc
http://www.ietf.org/html.charters/pilc-charter.html
http://pilc.grc.nasa.gov/