Jonathan brings up a good point:

>Using VTL: A slow network client will hold onto a VTL tape. The other images on that tape are then not

>available for duplication which means if there was a data center disaster those images

>would not have been copied to tape yet. This impacts your RPO.

This is a very valid valid point, but it will only come into play if you’re duping while you’re backing up.  Do many people do that? (I’m not sure.) 

Also, this effect can be minimized by having your virtual tapes be relatively small. Unless you’re doing what Stuart is talking about in his post (direct tape export from VTL to tape), then you should be able to make your tapes any size you want them to be.  Also, if you’ve got historically slow clients, you can put them all in one pool, so they only affect each other.

I’m not saying this isn’t a problem.  I’m just saying here’s how to work around it if you need VTLs, such as for large clients that you want to use block disk for (and you also want dedupe).  As I stated in my previous post, the only way to get both TODAY is to use VTLs.