Yes Invista does lun virtualization and presents to hosts by design.

Although I am not sure if I have understood you properly. Could you please rephrase your question in a different way?

LUN has to be virtualized before you present to host.

How does Invista virtualize a lun:
All LUNs present to hosts with Invista uniform LUN address. Invista may change backend physical storage without changing LUN addresses present to hosts.

lets see if the question in my mind similar to his...

imagine a host running perfectly fine with emc disks and production IO.
now i want to migrate the storage given to it using invista and without stopping a single i/o. is that completely possible?

or if, my host is configured for symmetrix disks and i want to migrate them (to mabe other vendor's system). is it possible to do without affecting prodction i/o?

Yes Invista can virtualize existing LUN's.

Can it do it without downtime?
Depends, if the host supports PowerPath and is on the EMC Support Matrix for PowerPath Migration Enabler and at the appropriate versions of PP to support it, the answer is yes. Otherwise No.

If not on the ESM or at the correct PP level, then you will incur a brief outage to encapsulate the LUN (import the LUN as a storage element into Invista, create a virtual volume with max size, i.e a 1 to 1 map of the entire physical LUN to virtual LUN that can be presented to the host as an Invista lun, and present to the host, do some host device/mapping), and your ready to go.

From there you can then move the virtual LUN from the 1 to 1 encapsulated LUN to a meta or another array via data mobility jobs without the host every know it has happened.

nope, not so far...but now i understand it a little bit more...

how about moving from emc to non-emc array? will we be limited to migrating onto devices/arrays supported by this powerpath only?