Hi,
The ability to do regular, non-disruptive tests and then audit the results over time is absolutely critical. This should not be overstated.
You also get a much more complete recovery plan - what starts serially, what starts in parallel, when do I pause and call someone, when do I launch customized scripts, if necessary, etc... SRM also changes VM network settings if the IPs are not in place at both sides.
SRM also gets you a QOS guarantee on recovered VMs. Since you will only have half the servers at the time of recovery, deciding how the VMs are balanced in new resource pools and what VMs are suspended to make way is critical for making sure the VMs are able to do real work upon recovery.
And lastly, but most importantly, the big difference is replicated vs. shared storage. If you are using storage shared between both sites, but that is not replicated, you can very nicely handle single host failures, but when you lose a whole building - power, facilities, fire, etc.... you are dead in the water. SRM, along with help from storage replication makes both buildings completely self-sufficient.
One caveat here, is that you happen to have active/active shared storage, so that the storage appears RW at both sides, and if networking is stretched, this last point is moot for you. But, I mention it since most storage vendors out there do not provide active/active replication, thereby making a very sharp distinction between SRM and HA.
Hope this helps.
-Jay
