Thanks for writing up the document.
As a general comment:
I think it should be added that there is a huge difference in building
enterprise routers and home routers in terms of budgets for development
/ support and margins.
This should somehow reflect in the complexity of the specifications and
reference software we choose or produce in this WG.
A few more specific ones.
wrt. Link State Algorithm
"This information can be used by other applications outside the routing
protocol itself. Additionally the flooding algorithm has been found as
an efficient method for other applications to distribute node-specific
application data, although some care must be taken with this use so as
not to disrupt the fundamental routing function."
Is this relevant for the homenet case, since this additional information
(Prefix Assignment, Naming, Security, etc.) is exchanged using HNCP already?
wrt. Algorithm Comparison
"Loop-Avoiding Distance Vector [...] scales badly in extremely dense
deployments, where a single node has thousands of direct neighbours"
Are thousands of direct neighbors relevant for the homenet usecase?
In general if we are talking about 100-200 routers here we've built OLSR
and BATMAN mesh networks in that range with WRT54GLs and similar devices
back then.
wrt. Link Metrics
Do the protocols assign any metrics by default (e.g. based on link speed
/ delay / reliability)? For Babel it is mentioned that it does so for
WiFi however is there a generic case? Does IS-IS do something similar?
I mean in the end what good does a 16-bit or 24-bit metric space do us
in a more or less unmanged environment if all hops "cost" the same by
default unless otherwise configured?
wrt. Security Features
"We should just state whether each protocol supports auth or encryption,
and whether it supports symmetric or something more exciting."
I agree. To my knowledge it seemed that none of the protocols supported
encryption or asymmetric authentication. While we can manage PSKs in
HNCP to solve the authentication issue we might want to mention the
potential insight that someone might gain from sniffing the unencrypted
traffic of either routing protocol.
wrt. IS-IS
"all major router vendors implement IS-IS" "As all major routing vendors
have (proprietary) IS-IS implementations"
Is this really true for home router vendors? I don't doubt it is for
enterprise routers but really the only routing protocol I've seen in
home router firmwares so far is RIP and even that only in very few
cases. Somewhat related: is source-specific routing in general deployed
anywhere in the enterprise (or similar) space already?
Cheers,
Steven
_______________________________________________
homenet mailing list
homenet@ietf.org
https://www.ietf.org/mailman/listinfo/homenet