On 23 Nov 2016, at 15:23, Ca By <[email protected]<mailto:[email protected]>> 
wrote:

<snip>

That said, given HOMENET's charter to be the ideal network we always wanted 
without the technical debt, i suggest HOMENET take a strong stance and reject 
"crunchy core, soft middle" security approach.  Meaning, assuming that some 
other device is going to do security for you and you can leave a default 
password telnet open.... that idea needs to die.

We need to make sure that HOMENET does not have a diagram that says "security 
done here" with an arrow pointed at the gateway.  HOMENET needs to specifically 
mandate all nodes have sane security, and part of that is ripping off the 
band-aid / security blanket of "stateful firewall"... the popular notion that 
stateful firewall does anything meaningful is very damaging to ecosystem... 
mostly because it makes security the responsibility of some other node.... 
which is not ok.

Part of the “problem” is that the Homenet security architecture is not yet 
documented. It was somewhat punted during the discussions towards RFC 7368, 
with Section 3.6 mentioning RFC 6092 and RFC 4864, without being prescriptive - 
https://tools.ietf.org/html/rfc7368#section-3.6.

I have my doubts that any attempt to flesh that out further now would reach 
consensus, but given we’ve now moved on quite a way, e.g. knowing we have HNCP, 
Babel, etc, and having witnessed Mirai, it might be worth a try. Something for 
the chairs…?

Tim

_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to