redundant dns services [3rd party dns]

[Peter McGarvey]

I've just remembered that there is a method to do this in the DNS.  The
SRV record type does all that you ask, and more.  Alas, no clients
support it yet - although there is a rumor that someone is working on
adding SRV support to Mozilla.  Shame, it looks like a most excellent
feature, specified in RFC 2052 IIRC.

* Lou Kamenov <phayze at secureroot.org.uk> [2003-04-07 16:51:42 BST]:
> In some email I received from Peter McGarvey <fbsu-x at packet.org.uk> on Mon, 7 Apr 2003
> 15:40:25 +0100, wrote:
> 
> > * Lou Kamenov <phayze at secureroot.org.uk> [2003-04-07 14:16:58 BST]:
> > 
> > > I need a dns server that can do good round-robin which checks if the
> > > requested record is up or not(read online, accessible),·
> > 
> > I'm personally not aware of any DNS server which checks if something is
> > up before replying.  Even if your authoritative DNS server could do
> > this, the DNS caches wouldn't - thereby rendering such functionality
> > useless.
> 
> dynamic dns will do the job, but the cache would spoil the good thing, the 
> time between record-changed and record-refreshed-all-over-the-globe wont be less
> than 10-15 minutes, may be more - may be less..
> 
> 
> > 
> > > DNS have a good scalability for MX records(also it's in a way
> > > integrated with the SMTP protocol
> > 
> > When you query a DNS server for MX records, you'll get all the MX
> > entries (along with the relative priorities).  It's the job of the SMTP
> > server to use that data to work out which MX to use, 
> 
> That's right
> 
> >DNS plays no part  in this decision.
> 
> Exactly. If a web client/browser can make such a decision, it'll be great,
> but somehow I'm not aware of anything similar?
> 
> > 
> > > Anyway, the thing i`m looking for looks is something like:
> > >
> > > http://www.zoneedit.com/doc/faq.html#fo
> > 
> > Looks like some sort of Dynamic Round-Robin DNS.
> > 
> > It's possible I suppose.  However, this:
> > 
> > > The average failure detection time is 5 minutes. This time varies
> > > depending on the speed of your site and the nature of the failure.
> > > Recovery times are faster, averaging 3 minutes
> > 
> > does not make any claims about the time it takes for the DNS to
> > converge.  It's quite possible for the DNS caches to still be directing
> > people to the broken site for anything up to a day, so personally I'd
> > want to test such a thing before paying for it.
> 
> I've requested some more information on this one, if it works the way i thought it'd be
> useless, since it'll take some time for the records to travel around the globe :/
> 
> Thanks for the input Peter.
> 
> 
> cheers.
> 
> -- 
> 
> Lou Kamenov	AEYE R&D 	lou.kamenov at aeye.net	
> FreeBSD BGUG	http://www.freebsd-bg.org	lou at FreeBSD-bg.org	
> Secureroot UK	http://secureroot.org.uk	phayze at secureroot.org.uk
> Key Fingerprint - 936F F64A AD50 2D27 07E7  6629 F493 95AE A297 084A
> One advantage of talking to yourself is that you know at least
> somebody's listening. - Franklin P. Jones 

-- 
TTFN, FNORD

Peter McGarvey
Freelance FreeBSD Hacker
(will work for bandwidth)