Re: [Int-area] Some thoughts on draft-yong-intarea-inter-sites-over-tunnels

[Joe Touch <[e-mail]>]

Fred,

The question is not (yet) whether there are problems. The question is
whether a new solution is required where existing mechanisms suffice.

Your doc has only one paragraph that describes anything close to a use
case, but doesn't explain why AERO is *needed* to solve that case.

I.e., I've repeatedly looked and still don't find the answers there.

Joe


On 12/12/2016 8:00 AM, Templin, Fred L wrote:
>
> Hi Joe,
>
>  
>
> Whatever. You seem to keep implying that there are problems, but I can
> assure
>
> you there are none. Why not have a look at the document, because I
> think you
>
> will find the answers to your questions there.
>
>  
>
> Thanks - Fred
>
>  
>
> *From:*Joe Touch [mailto:[e-mail]]
> *Sent:* Friday, December 09, 2016 8:53 PM
> *To:* Templin, Fred L <[e-mail]>; Lucy yong
> <[e-mail]>; Brian E Carpenter
> <[e-mail]>; [e-mail]
> *Subject:* Re: [Int-area] Some thoughts on
> draft-yong-intarea-inter-sites-over-tunnels
>
>  
>
> Fred,
>
>  
>
> On 12/9/2016 4:25 PM, Templin, Fred L wrote:
>
>     Hi Joe,
>
>      
>
>     I read your document and, for the applications I am concerned
>     with, I still
>
>     think what I am doing is the better approach. One thing that you
>     may not
>
>     have gathered is that the AERO interface does not maintain a
>     replicated copy
>
>     of the entire IP forwarding table;
>
> I wasn't assuming it did - therein lies the problem.
>
>
>     it only keeps neighbor cache entries for its
>
>     currently active sets of neighbors. For AERO Clients, this would
>     include the
>
>     default router(s) and any peers that it has recently received
>     Redirects from.
>
>     For AERO Servers, the neighbor cache would include entries for the
>     current
>
>     list of associated Clients. So, the AERO interface is not a
>     full-blown IP router;
>
>     it is a neighbor discovery engine for its active set of neighbors.
>
>
> But it would need to have the full-blown IP forwarding capabilities to
> determine which next IP address is intended for a given packet handed
> to it by the master IP forwarding table.
>
>
>      So, unlike a dynamic routing protocol the AERO interface uses
>     IPv6 Neighbor
>
>     Unreachability Detection (NUD) instead of routing protocol
>     keepalives to
>
>     maintain reachability. There is also no routing protocol control
>     messaging
>
>     going out over the underlying data links – it is simply data
>     packets plus
>
>     occasional NUD messages.
>
> That only describes how the table is populated. There's the further
> issue of how the table is indexed, which is a full-blown forwarding
> lookup (with policy information as well).
>
>
>      I noticed that your document was from 1997, which is the same year I
>
>     started with SRI International. I think that was right around the time
>
>     you and I first met.
>
>
> Not sure - it was presented in early 1997 at the GBN workshop at
> Infocom, but also at a few DARPA PI meetings before.
>
> FWIW, I didn't think we met until the IETF, which was in Dec in DC
> that same year.
>
> Joe
>