Neil Barton

Monday, March 02, 2009

Storage Service Prices

A Compass newsletter last week "has observed annual unit price decreases of 25 percent to 30 percent for storage hardware, 30 percent to 35 percent for storage services, and 5 percent to 10 percent for storage software".

That's a remarkable claim.

Fair enough on hardware: it has been well documented for a number of years that the number of Gbytes you can store on a disk drive is growing, while the cost of manufacturing a disk drive is in slight decline. That inevitably leads to a reduction in unit prices, if you price storage "per Gbyte", which most people do.

The storage services claim is more problematic. Storage services pricing is driven by two factors: the amount of effort needed to manage the customer's environment, and the cost per day of that effort. Labour costs are certainly not declining by 35% a year, they are rising (very fast in some geographies). So to achieve 35% year storage service reductions, it you would have to cut the effort needed to manage 1 Tbyte by a third, each and every year, indefinitely. That's a level of productivity improvement which would be astonishing, especially if you take the growing view that it's not just the number of Gbytes which drives storage effort, but the number of logical volumes to manage.

I speculate that behind the headline in the Compass report there is something else hidden. One possibility is that there is a one-off shift from on-shore labour to off-shore labour for storage management. This could have have the effect of reducing the average prices on Compass's database by 30% from last year to this. But it's not a sustainable model for forward pricing of an outsourcing deal. You can only off-shore once.

There is another possibility. Storage management is very susceptible to economies of scale. It is much cheaper per Tbyte to manage 500 Tbytes than it is to manage 50 Tbytes. It could be that, by accident, the data in the Compass study happens to have a lot of smaller installations last year, and a lot of larger installations this year. So volume discounts would be partly the reason for what appears superficially to be a unit price decline.

I could be wrong about both of these, of course. I haven't seen the underlying data on which Compass base the claim, or what adjustments they have applied to the data. But the claim doesn't jive with the known facts, which is why I question it.

As always, this is my personal view, not that of my current employer (or any previous employers!)


Post a Comment

<< Home