—
Dr. Joe Touch, temporal epistemologist
www.strayalpha.com
> On Feb 28, 2022, at 9:54 AM, Toerless Eckert <t...@cs.fau.de> wrote:
>
> On Fri, Feb 25, 2022 at 08:19:42PM -0800, to...@strayalpha.com wrote:
>> I disagree; a tunnel (done correctly) is isomorphic to a link. There’s no
>> difference between tunnels and what we already rely on as “L2”.
>
> I guess wrt. routing we (Internet Routing Architecture) started out with
> alot of simplifying assumptions about the properties of links. Primarily them
> being full-mesh-mp2mp or p2p. And then MPLS gave us arbitrary mesh mp2p LSPs,
> and then MANET came with arbitrary partial-connectivity meshes. I don't this
> we ever formalized any of this so that we could really confidently use
> link or tunnels and be clear about what options the reader will assume.
RFC3918 tried to formalize some of it; the rest is pending in
draft-ietf-intarea-tunnels.
>> The flaw is the OSI model assuming layer levels are absolute (they’re
>> relative) from all viewpoints (again, relative). There’s a strong
>> equivalence between a link, a tunnel, and a router (which, in essence,
>> emulates shared link). And, interestingly, forwarding can also be described
>> as recursive tunneling.
>
> I thought you just look whats on the wire and you know
> the level. If its a frame it's L2, if it's a packet, it's L3. If it's
> anything more complicated, it's a tunnel (sorry, couldn't resist ;-))
There is a base case to the recursion, i.e., where logical information meets
fermions and bosons (literally). But that tells you only that base layer; it
tells you nothing about the meaning of the headers you see inside, e.g., in
OSI, they would be 1,2,3,4,5,6,7, but in an IP tunnel, they could be
1,2,3,3,4,5,6,7, with GRE they could be 1,2,3,2,3,4,5,6,7, etc.
(i.e., referring to protocols by the layer they’re commonly associated with).
> Is there any IETF RFC that is actually any better than OSI in this respect ?
I hope the tunnels doc will be ;-)
> And yes. I wish there was better material about recursion as the core
> fundamental of building networks. Especially because so far we mostly
> have re-invented the wheel for every instance of recursion (frame,
> packets, tunnel headers..).
Part of the reason that doc is delayed is the “intro to computer networking
from first principles” I’m writing as we speak that does exactly that.
Joe
_______________________________________________
Int-area mailing list
Int-area@ietf.org
https://www.ietf.org/mailman/listinfo/int-area
