Newbie Question

[Simon Waters]

Kris McElroy wrote:
> 
> Would this work as a failover or am I missing the picture?

You can't rely on applications to requery other IP's if the
first one fails alas.

However I'm assured web browsers are better at doing this than
other applications.

It isn't an ideal fail-over mechanism relying on client
behaviour, but try testing it, see if it good enough for you.

Typically big sites are using the multiple IP addresses for load
balancing, and may have clever hardware sorting this list so
geographically close servers are preferred, removing overworked
or down servers, and other clever tricks. Which explains why
some busy sites have low TTL on their resource records.