<   Previous Post  Kids these days
Storage virtualizati...  Next Post:   >

Comments (3)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 localhost commented Trackback

I agree - both Hu and Mr. T are almost irrationally fanatical about their "must be in the controller" position.<div>&nbsp;</div> Oddly, Hitachi's engineers haven't been able to deliver thin provisioning for virtualized external storage yet. <div>&nbsp;</div> I can't decide if that's becuase a) it's hard to do, or b) it's "impossible" because in this configuration, the USP-V is acting as a network virtualization device? When inserted in the network between the hosts and the actual storage arrays, the USPV is really just a very big (and expensive) I/O switching device, more similar than different to both SVC and Invista (no offense BarryW).<div>&nbsp;</div> I chuckle at the oxymoron everytime I hear them say "in the controller!"

2 localhost commented Permalink

Funnily enough I filed a patent a year or two back for 'pre-emptive migration of data' both up and down through the tiers... it ended up as a publish as it was covered by mainframe prior art! But with a little knowledge of access statistics - some higher level software could easily do block level ILM using the migration features in SVC. Another great reason why sitting in the data path and watching the actual I/O data in the SAN is a good idea :)<div>&nbsp;</div>

3 localhost commented Trackback

Is thin provisioning/over allocation a feature of IBM SVC? Or will it be in the near future?