Load Balancing Question...

[David R. Conrad]

Kevin,

> If you reduce the TTL on the RRset then the effect of intermediate caching
> servers is minimized, the downside being that you increase the query traffic and
> just generally be a bad net citizen when you do this. 

Right, although this doesn't seem to bother many of the folks who are selling
load balancing "solutions".  At least not yet...

> But it may be a perfectly
> acceptable approach on a private net or internet. 

Agreed.  The problem I see is that people who understand less about the
implications of the technology will implement a server side load balancing
solution, test it locally, see that it works as advertised, then assume it
will work globally.

But perhaps I'm just being paranoid.

> I don't know anyone
> who would trust their IBM mainframe to be directly Internet-connected.

Eh?  IBM mainframes make wonderful servers and I know lots connected to the
Internet (umdd.umd.edu, for example, which has been on the net since at least
'86).  This is why tn3270 exists...

> Actually, a bigger problem with my hack was that the weighting 
> information wouldn't have any way to propagate to slaves. 

As you mention, this can be handled by using some out of band mechanism to
inform the slaves of the weighting, so I don't see this as significant as the
caching issue.

Rgds,
-drc