Let's imagine for a second, a completely non-virtual world. One in which hypervisors, guest-to-host ratios, tenancy, memory tuning and all those other many attributes that have redefined the IT business in the last half-decade had never even been thought of. How much storage would that world need?
I put it to everyone I talked to at SNW Europe (the must-attend storage conference, hosted in Germany) this year that virtualisation makes not one jot of difference to the overall ravening thirst for data and, hence, places to keep it.
De-duplication appears to comprise some a priori guesses about patterns of data.
Sure people commissioning big SANs for the express purpose of hosting formerly-physical machines in a handy fashion that users can't tell apart from the real thing, but ,while those VMs would be saving on CPU cycles, raw electricity, aircon, physical footprint and monopolisation of under-used equipment, they comprise precisely the same volume of bytes as their physical forebears.
Some pretty standard themes kept re-emerging during each meeting I had at SNW. They do like their archiving processes; de-duplicators. US vendors, it would seem like search and disclosure because of the litigious environment they find themselves in. Brits, on the other hand, like business continuity, because they live in fear of disaster and associated data loss.
Incidentally, chatting with Druva's Lisa Busby, about really backing up the vast sprawl of platforms common in US businesses, it turns out that tools to tackle the sprawl and it's impact on the duty of record-keeping and legal pre-discovery do not inherently overlap with the British need for resilience.
De-duplication appears to comprise some a priori guesses about patterns of data (I am imagining comments in the code of the de-duplicator firmware along the lines of "oh no not ANOTHER Microsoft Office Dev Team Easter Egg"), running on CPU hardware which spends most of its life running at 100 per
cent.
