On Wed, 5 Jul 2000, Dmitri Krioukov wrote:
the major disadvantage of the foundry (bgp) solution is longer prefix injection.
the major problem with the dns-based solutions is that they're not topology-aware (-> suboptimal routing). attempts to make dns smart lead to rather awkward reverse pinging configurations and proprietary protocols running between load balancers. (there was also rfc2052 by paul vixie but it required modification of dns clients.)
I don't think anyone sane would advocate a reverse pinging setup, but rather returning a DNS answer based on which DNS server you hit, and you hit the DNS server thats "closest" to you. I guess it depends what you value, short web traffic and instantanious failover regardless of the disruption of existing connections due to a route flap or metric change (remember this could be the fault of ANYONE's network, not necessarily yours), or guarenteed connections with possible non-instantanious failover. It also depends on the density of your server farms (one on the east coast, one on the west coast, it would take a fair amount of problems for the routes of the "majority" of people on either coast to change to the other). This is grossly generalized load balancing at its finest, take from it what you will. -- Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/humble PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)