On Fri, 22 Mar 2002, Yamasaki Toshi wrote:
> > - pppext might be workable for PPP connections, and is a valid mechanism
> > especially if there is no simple alternative (e.g. if the only alternative
> > is full statefull DHCPv6).
> 
> No. PPPext solves the issue only when you use PPP. That means you need
> another solution when you don't use PPP, and that makes it difficult to
> achieve zero-configuration enviroment.

Which is why I said 'PPP connections'. :-)

> > - dhcpv6 will work ok, I think, but is a bit complex for being usable
> > in every scenario
> 
> Yes, that's why Ralf proposed the DHCP subset or DHCP-lite named "DNCP".

Which may still be too heavy.

> > - APD w/ ICMPv6 is ok but can be simplified (e.g. removing prefix return
> > message)
> 
> I think APD is simple enough, and DNCP is also simple enough. Technically
> there is almost no difference between DNCP and APD.

True, but APD has IMO unnecessary components of stateful nature, like 
prefix return message.

> > - RA (Lutchansky) is a nice hack, kinda works with hacking.
> 
> It assumes only P-to-P enviroment. Again you need another solution for
> non-P-to-P enviroment. Our goal is to achieve auto-configuration, right?
> Then, how dose a CPE know whether it is conneted to P-to-P link or P-to-MP
> link automatically?

Prefix Delegation need not be (IMO) completely automatic: it's completely
ok to have to add a 'request-prefix interface eth0', 'request-prefix
interface any', or whatever in the configuration of the CPE.

Via that, detecting the link type is irrelevant: we don't necessary end up 
with one mechanism already.

> > Note that IMO, configuring these with "stateless DHCPv6" but obtaining the
> > prefix with e.g. APD makes perfect sense -- there might be reluctance to
> > implement full DHCP, or even the minimal set of "stateful" DHCP.
> 
> Both APD and DNCP are "stateful".

see below.

> I'm providing commercial IPv6 services today, and will start IPv6 over
> DSL/FTTH services soon. You should understand that "stateful" is one of the
> most important requirements of commercial ISPs.
> 
> How do you charge and/or solve abuse problems while you don't know who used
> which prefixes and when?

I don't think that anyone is assuming 100% automatic prefix generation 
like EUI64 IID generation: rather, I assume that in (at least almost) 
every case, the prefixes will be configured statically somewhere, along 
the lines "identifier/customer: prefix".

With stateful above I meant protocols which expect one to request a prefix 
and return it back, release it, return it back if no longer used, etc. -- 
consider router advertisements which aren't stateful in that manner.

But these are what the requirements document is for..

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords

--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to