[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [DNA] Issue 6: DNAHostLinkIDList
- To: Sathya Narayanan <sathya@research.panasonic.com>
- Subject: Re: [DNA] Issue 6: DNAHostLinkIDList
- From: Greg Daley <Greg.Daley@eng.monash.edu.au>
- Date: Mon, 05 Feb 2007 14:15:54 +1100
- Cc: JinHyeock Choi <jinchoe@gmail.com>, Dna <dna@eng.monash.edu.au>
- In-reply-to: <45BF6396.6050202@Research.Panasonic.COM>
- References: <45537A0D.3020009@Research.Panasonic.COM><92e919fb0611162105x39e4fa03u8f9a07c4ec7f27e6@mail.gmail.com><455DDE50.7010204@Research.Panasonic.COM><92e919fb0611200137r174a212apadc4ca2bacbf262f@mail.gmail.com><456F85D1.30509@Research.Panasonic.COM><92e919fb0612010333ob58c56eo827a74128e6126bd@mail.gmail.com><45704479.8060209@Research.Panasonic.COM><45BA5EC0.5070104@Research.Panasonic.COM><92e919fb0701281841r912fe16q6db795dd5ded99ca@mail.gmail.com><45BF6396.6050202@Research.Panasonic.COM>
- Sender: owner-dna@ecselists.eng.monash.edu.au
Hi Sathya,
If there's no configured prefix on that interface of the router, it
shouldn't advertise prefixes (unless it learns one through Router
Discovery).
In that case it will not route to or from a non-link local prefix either.
Then it's not a problem as there's no prefix on link.
Advertising no prefixes (as complete) is the same as advertising the host's
existing prefixes are invalid, and provides strong link-change detection.
Completeness could be determined in the same way as usual (by solicitation),
and this issue can be avoided if any of the routers advertises a prefix,
as I indicated.
Greg
----- Original Message -----
From: Sathya Narayanan <sathya@research.panasonic.com>
Date: Wednesday, January 31, 2007 2:30 am
Subject: Re: [DNA] Issue 6: DNAHostLinkIDList
To: JinHyeock Choi <jinchoe@gmail.com>
Cc: Dna <dna@eng.monash.edu.au>
> JinHyeock -
>
> You are right. I didn't realize that before. Thanks for correcting
> me. I
> feel a lot more comfortable now.
>
> Now, is it ok the REQUIRE the sysadmin to configure the routers with
> atleast one prefix even if stateful address configuration is used. I
> guess it is ok. I don't know what others think.
>
> Greg -
>
> Your idea to use the configured non-link local prefix is a good
> one, but
> it still leaves the scenario where the router is configured only with
> link-local addresses on it local interface. AFAIK, thats going to
> be the
> most common scenario. So, if we require the routers to advertise
> atleastone prefix, it MUST be configured on the router even when
> statefuladdress configuration is used.
>
> Sathya
>
> > Dear Sathya
> >
> > Thanks for your trouble advancing the work.
> >
> >> I still haven't closed Issue 2:, changes to support Stateful
> >> autoconfiguration scenario. The current text depends on DHCP
> CONFIRM to
> >> address this scenario. I am still uncomfortable requiring
> configuration>> of each router on the link with the smallest prefix
> - a sysadmin then
> >> becomes responsible for keeping track of the smallest prefix and
> >> updating the configuration each time.
> >
> >
> > Sysamin doesn't have to keep track of the smallest prefix and
> > configure it to each router manually.
> >
> > Sysamin only have to configure a router to advertise at least one
> > prefix. Then, after a while, the routers on the same link, upon
> > hearing the prefixes from each other, will pick the smallest prefix
> > and advertise it in either PIO or LPIO.
> >
> > Thanks for your kind consideration.
> >
> > Best Regards
> >
> > JinHyeock
>
>
>