Will flash save the data centre? Don't spread your wings yet, Vultan

Why horrible reality hobbles the gleaming, solid-state future


Analysis The all-flash data centre: does it exist? Will it ever?

Can we really imagine a data centre with no spinning disks and all the executing apps relying on solid-state storage? Such is the vision put forward by Violin Memory in its "Disk is Dead" campaign. But is this the future of the data centre?

Violin chief marketeer Amy Love blogs a colourful metaphor. “HDDs are being replaced as the primary IT data storage medium. The all-flash data centre is already a reality for many businesses in technology... Just as the Stratocaster and the Les Paul became the de facto standard for just about every rock band, flash storage platforms will be the heart and soul of tomorrow’s 24x7 data centre,” Love, rather naturally, enthuses.

Should we buy into this? Turns out more mundane facts about realities of supply will almost certainly prove to be the flies in this beautiful ointment.

The idea of an all-flash data centre (AFDC) can seem absurd when you look at different kinds of data and the best place to store it:

  • Primary data with fastest access needs
  • Immediate-access data, but not fastest
  • Reference dara
  • Back-up data
  • Archive data

Data is said to age and to flow through these stages in a data life cycle process. It’s also said to have different storage media needs based on how much of it exists and access time requirements on the one hand, and storage media speeds and costs on the other.

Flash media is the fastest to access and the most expensive to buy on a per-GB basis. Next fastest and second-most expensive are 15,000rpm disk drives, then 10K ones, on to 7,200rpm bulk capacity drives, which are now nudging up to 10TB. Then it is tape, with the lowest $/GB cost and slowest access speed.

This is all well understood and if we map data classes on to storage media types, we get this kind of table:

Data_class_and_media

Data classes and appropriate media. Red for best fit and sand for so-so fit. White for no fit at all.

This table prompts us to ask how Violin’s all-flash data centre customers are dealing with data that is ill-suited to flash storage – such as reference, back-up and archive data?

We should note that Violin's raw $/GB cost is higher than 15K disk, but after deduplication and compression are applied, the effective $/GB cost is similar to 15K disk, even below it, depending upon the data reduction ratio. Intuitively, these customers are storing any immediate-access data, like mail, on flash too.

They could get away with doing this by not storing any non-fastest-access data on site at all and using a remote data centre or cloud provider.

Shifting all your slow-access, capacity data to a different data centre – one full of disks – while saying you have an all-flash data centre approach is a big fudge. Pumping all that data out to a public cloud is not a fudge at all. You now actually and literally do have an all-flash data centre.

But such an approach isn’t feasible for many enterprises and organisations, perhaps the majority, because of cost and concerns about security and access reliability. It’s not unknown for, say, Microsoft’s Azure to have an outage.

Next page: Flashy new gear

Biting the hand that feeds IT © 1998–2021