DNS Based Load Balancers
brandon at rd.bbc.co.uk
Wed Jul 5 12:07:54 UTC 2006
> GSLB based on DNS have one significant shortcoming that moone here has yet
> mentioned: they are performing their magic on the location of the
> _nameserver_ that issued the query.
> this can be VERY different to that of the ACTUAL location of the client.
Systems that infer stuff make errors, at least DNS GSLB fails working
I've seen problems with DRM that assumed the IP requesting http would
be the same as that requesting rtsp, neither of which were the actual
client after some ISP caching.
> for example, Akamai always sends to off to a serverfarm in Northern
> California, because that's where my DNS query is originating from.
> that is almost the exact opposite side of the planet from where I'm coming
That's more a failure of your systems than Akamai,
depending on something the other side of the planet
for something best done locally
> irony is that there is an akamai cluster about 10 feet away from where my
> [subsequent] http requests originate from...
yes, like a free ride when you've already paid
I don't see any point arguing over DNS GSLB. It exists, it can work,
and no surprise some products aren't well designed.
We've used a DNS GSLB system since 1997, it does the job we designed it
to, it has the edge cases we expected, we're happy with it.
Don't try using one without adult supervision
The hardware SLB products we've tried have caused more down time
than our DNS LB
More information about the NANOG