[311] in Hesiod

home help back first fref pref prev next nref lref last post

Re: Info please

daemon@ATHENA.MIT.EDU (Theodore Y. Ts'o)
Fri Jul 19 18:59:00 1996

Date: Fri, 19 Jul 1996 18:58:41 -0400
From: "Theodore Y. Ts'o" <tytso@MIT.EDU>
To: amit@jpmorgan.com
Cc: asafier@csc.com, "Theodore Y. Ts'o" <tytso@MIT.EDU>, hesiod@MIT.EDU
In-Reply-To: Amit Bhatiani's message of Fri, 19 Jul 1996 17:27:59 -0400,
	<31EFFDDF.1C7B@jpmorgan.com>

   Date: Fri, 19 Jul 1996 17:27:59 -0400
   From: Amit Bhatiani <bhatiani_amit@jpmorgan.com>

   This is not even a DNS issue, but part of a larger issue which goes by
   the name of high availability. The basic idea is that you connect your
   primary and secondary machines two/three ways. One is the normal network
   that they see each other on now, and you could possibly have a another
   serial PPP link (if you don't have another network interface and would
   not like to buy cards for the machines). The secondary machine is in
   some form of single user mode. I say some form because you want to have
   enough of the IP stack up to talk on the second PPP interface. This can
   be done without bringing the main network interface up.
    The secondary polls (ping or something similar) the primary every so
   often (30 seconds) and decides if the primary is up or down. If it is in
   fact down, the primary machines runs a script that brings up the main
   network interface with the same IP address and hostname as the primary. 

If you want real high availability, then you should have automatic
failover of your rounter of your entire network path to the rest of the
outside world.  This then isn't a DNS issue, but a matter of having your
router be multi-homed and having multiple points of connection to the
Internet.

   Ted is right in that there is no way in DNS to solve this problem, but
   that's because DNS is a name service, not a fault-tolerant system..:-)

That's because the DNS assumes that the network will provide whatever
fault-tolerance you're looking for.  And providing high-quality fault
tolernate in the IP protocol is a well understood problem ---- the
solutions may be rather expensive, but they're well-understood.

							- Ted

home help back first fref pref prev next nref lref last post