IBM's SAN Volume Controller effectively equals Hitachi Data Systems' USP. Discuss.
The popular IBM SVC is a virtualising storage controller that runs storage management applications and resides in the SAN fabric, attached to a fabric's director. It is widely used, very fast, and reliable.
The popular HDS USP is a virtualising storage controller that runs storage management applications and resides at the storage edge of a SAN fabric. It is widely used, very fast, and reliable.
Both devices have accessing server applications come at them through the fabric. The locational diifference between them is, I'd argue, essentially trivial. Yes, the USP can only 'see' storage arrays connected to it and not any other SAN storage devices connected directly to the SAN.
Yes, the SVC can, in theory, 'see' all storage devices connected to the SAN and so could, it wished, work with a tape library. In these days of VTLs I think such a direct tape-to-SAN connection is weakening in its appeal.
Yes, the USP can 'see' files as well as blocks. It's early days for this to be used in practise and there's no reason, theoretically, why a file virtualisation function couldn't be added to the SVC. We might muse about a hypothetical combination of F5 (Acopia as was) ARX file virtualising function and SVC block virtualising functions.
To my mind the core block virtualising functionality of SVC and USP are similar enough to render discussion about the ideal location of storage intelligence in a SAN fabric to be redundant.
Of course there are detailed differences in implementation and replication and so forth, as there always are when comparing different manufacturers' products doing essentially the same thing. Consider differences in de-duplication technology as an example.
But, I think, arguing about whether one or the other virtualising SAN storage product is better than the other because of its location in the SAN fabric is sterile and unproductive. It's angels on a pinhead territory.
The same is logically true for SAN storage virtualisation products embedded in SAN director enclosures, such as EMC's InVista.
InVista isn't a slow-seller because of its location in the SAN fabric. I'd submit it's a slow-seller because both IBM and HDS have much stronger and better developed SAN storage virtualisation offerings which do a whole lot more for customers.
Unlike real estate the three most important things for a virtualising SAN storage product are not location, location and location. Let's forget about SAN location and concentrate on three far more important things: functionality, performance and price.