Load balancing both the SP and IdP

[Cantor, Scott]

On 10/23/13 9:57 AM, "Peter Schober" <peter.schober at univie.ac.at> wrote:
>
>You mean "not clustering at all", just to be clear?

I would call it active/passive rather than not at all. You still need a
load balancer if you don't want to wait for DNS timeouts to restore
service, so I think that's still more than nothing.

>Univie has been doing "that" (i.e., nothing, other than have hot
>spares waiting) for years (making Terracotta so absurd today, when
>most downtimes seemed to be caused by the very thing that was put in
>place to avoid them, among other things).

That's a given of course. I'm just thankful Paul explored Infinispan and
determined it's not workable before we wasted more time on "clustering"
technology. Never again.

-- Scott