Sign in with
Sign up | Sign in

Watt-ergate: The SSD Power Consumption Conspiracy Exposed

Watt-ergate: The SSD Power Consumption Conspiracy Exposed
By
Brought to you by What's this

SSD Power Consumption: Free Yourself From Fuzzy Math & Misinformation

When we review a solid state drive, we run it ragged through a number of tests. Of course, we report on throughput under a variety of conditions and workloads, but desktop enthusiasts have the luxury of focusing on the drag strip stats—your 0 to 60, your quarter mile. They want to know an SSD’s 4 KB random performance. They want to know its sequential performance. The rest? Superfluous details, as far as they’re concerned.

If only everyone else could be so lucky. In the mobile arena, how many watts a given SSD needs during idle and busy times actually counts for something. Of course, it would be great to close our eyes and pretend that power consumption doesn’t matter. While you’re at it, tell yourself that a notebook’s weight is negligible, one LCD is the same as another, and any other fiction you’d like to believe. When battery life is a limiting factor, it’s important to weigh an SSD’s power consumption with the rest of its test results as you make your buying decision.

The issue is arguably even more crucial at the enterprise level. Although the general consensus is that SSDs are more energy-efficient than hard drives, at least in terms of I/O instructions per second, how do we compare one SSD against another in a data center environment? Top-tier, enterprise-level SSDs are engineered to run 24x7 for five years. (The SSD in your desktop suddenly seems lazy by comparison.) So, assume the difference in power consumption between SSD A and SSD B is one watt, and then multiply it out. Over the course of those five years, that one measly watt results in a difference of nearly 44,000 W. When you put a dozen drives in a server and 5,000 of those servers in a data center, you’re sitting at 262,800 kWh. Even in areas with relatively low energy costs, a quarter of a million kilowatt-hours is more than chump change.

Thus it seems like a foregone conclusion to hitch your wagon to the SSD with the lowest power consumption. If you’re buying a drive to slide into your notebook, maybe that’s sound thinking. If you’re in charge of upgrading an entire data center, though, you could be making a costly mistake.