Copan's approach demonstrates why a good data archiving strategy can translate into energy savings
Back in February, I banged off sort of a Valentine's blog to Nexsan about some of the work it was doing around MAID (Massive Array of Idle Disks), which is a concept that's being applied to enterprise storage architectures in order to eek out greater power efficiency. After all, storage devices are probably the second biggest eater of energy in a data center, and this isn't an equation that's likely to change soon.
To remind you all, MAID is a way of powering down hard drives within enterprise storage systems when they're not being called upon to do something. This white paper from StorageIO Group kinda details the various benefits that a management strategy based on MAID might provide. The concept is somewhat controversial, because of the havoc that an unplanned shut-down can wreak on a drive head. SCREEEEEEEEEECH. At the same time, though, MAID underscores the very real positive impact that a thoughtful data archiving strategy can have on an enterprise data center. The simple fact is that not all data is equal and not all of it needs to be available instantaneously or updated more than its original creation event. Why not store such data on devices that can sit idle until they're called into service?
Naturally, the bonafide pioneer in MAID technology, Copan Systems, has some evolving thoughts of its own on this technology, which first came into being about six years ago through work done with the University of Colorado. Plus, they've had some news this month, which prompted my recent discussion with CTO Chris Santilli.
First, the news: Copan has updated its Revolution 300 Series Platform with some new drive configurations that bring its maximum capacity up to 896 terabytes (holy cow, Batman!) of raw capacity, or 8 petabytes of de-duplicated data in a system footprint that is less than a single-square-meter footprint. The company's press materials claim a 60 percent to 85 percent power savings over traditional storage systems. Here's the whole skinny on the product series for those of you who are way more technical than I.
Copan intends this product as a solution for "persistent data" that needs to be saved but that doesn't have to be at your virtual fingertips. Anecdotally, Santilli holds up the following two developments as key capabilities of the update that are directly applicable to the energy-efficiency discussion. (There are more, but I direct you to read them in the product materials because I am bound to misinterpret some data point.): - First off, Copan claims the Revolution 300 Series Platform offers more power efficiency per gigabyte, since the system is designed to only power up disk drives that are needed to deliver data. They're not on all the time. The higher density architecture basically means fewer power suppliers, fewer processors and fewer fans. - A 40-drive VTL cache option that can support more than 1,000 concurrent data streams. That means that more data is available quickly. Santilli cites the example of a company with media servers containing video, which takes up an enormous amount of disk space. Most of the time, this content sits idle, but when it's called up, it needs to stream in the most efficient way possible.
For those of you who wish to run some calculations on how Copan technology would affect your architecture, here's the company's power and cooling savings calculator.