>
> What's keeping OpenSRS from having a fully meshed/redundant network?
> Typically, if you have enough money to purchase a single 650x and a colo,
you have
> enough money to purchase 2 switches and run HSRP.
>
> Considering the critical nature of OpenSRS, I would highly suggest some
> hardware/geographic redundancy.
Chip -
Redundancy is definitely part of our architecture. Sometimes things come
your way that are out of your control. This was certainly one of those
situations. As was mentioned earlier, we're still looking into the issue.
Geo redundancy, while nice, is not always appropriate for all applications
at all cost levels. LibertyRMS has featured this from day one. On the other
hand, we're not sure that the likelihood of Toronto falling into a chasm is
enough to warrant such a drastic measure at this stage.
Your point is well taken however. And, as most of you are aware by now,
those that need to know read this list on a regular basis.
-rwr
