Julien Abeille (jabeille) a écrit :
Hi Alex,
Regarding routers with one interface, it is true that regarding usual
practice and deployments, it is new.
I agree.
However conceptually, a router is usually defined as "a node that
forwards IP packets not explicitly addressed to itself".
Silence on the assumed number of interfaces isn't a favoring argument
for one-interface routers either.
A router putting a packet back on the same interface it came from sounds
bad routing problem in the first place...
To the best of my knowledge, this is what RFCs assume when they use
the router concept.
But many RFCs specify behaviour with respect to a particular
interface.
Hence we should not run into problems nor need to modify standards to
support the one interface scenario. Do you see scenarios /
implementations where having one interface will break something?
I'm not sure what detail is needed to support the view that
one-interface routers are particularly new beasts...
IP-repeating packets on the same interface is noise in the air. It
breaks the availability of free timeslots to neighboring nodes.
Waste of memory; common routing tables (and ND-related structures too)
have a field "interface"; if all entries had the same content in that
field then why would the field be needed at all?
I will try to make a separate topic (again) on this...
Alex
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan
