Yep, been there done that got mixed results.

In theory, you shouldn't see any difference but it really depends on your service provider for ESX - unless it is all being done in house? Mainly because you will want to make sure that there is nothing else with high throughput being run off the same physical host. Also, backup/restore procedures etc for ESX

I remember running some performance software on the virtual machines for processor calculations, memory performance etc to compare against the physical server that was being migrated and there was no noticeable difference to be concerned of. Just to get that concern out of your head, maybe it is useful for you to run a similar set of tests.

Problem you are going to find is that you can't actually provide a real enough test on the virtual machine to simulate the true load it will get when all interfaces/users are pointed at it - just like migrating to another physical machine. There is an element of JDI (Just Do It).

On one VMWare migration we found much worse performance than the previous physical host (despite same server specifications) from the OPC interfaces...still not resolved.

I remember OSI did (or were going to do) a white paper on Physical vs Virtual PI, not sure if it ever materialised. They do not seem concerned by such a migration. For now, he is there TechSupport article with some details: http://techsupport.osisoft.com/support+solution/2/3ab2d3b65a0c45f69f5ee02dad8a2fed.htm