Storage Optimization


Who’s Really Melting the Ice Cap?

Posted in Featured,Storage by storageoptimization on May 21, 2008
Jon William Toigo’s blog “Drunken Data,” which has fun with its headlines, has a post titled “Climate Change or Silly Season?” In it, he references an article that ran in Macworld UK stating that Apple Computer–that darling of uberyuppies and designers–has been rated as a contributor to global warming.
Credit Flickr User Tom\'s Caps
Toigo’s response: “While I agree that the company generates a lot of hot air, the truth is that storage hardware, not PCs/MACs/servers, is the big power pig. Behind it all is a total mismanagement of data. Think about naming your files better and deploying archive technology the next time you see that video of a chunk of ice breaking off from a glacier.”
In truth, servers (and computers in general) give off a lot more heat per unit of rack space than storage. Processors that are running full out generate a lot of heat, and consume a lot of power. At the same time, both individuals and corporations have a high ratio of storage to servers, so if you add it all up, it might be the case that a data center uses as much power for storage as it does for servers.
That being said, I don’t think “naming your files better” is going to turn out to be the answer. Some combination of thin provisioning (waste less free space) and storage optimization (store things efficiently, along the lines that virtual machines use CPU efficiently) is the direction that things are headed.
The key thing to keep in mind is: are the servers and storage being used efficiently? In the server arena, virtualization has turned out to be the magic answer – allowing data centers to consolidate multiple logical servers on to one physical one to make sure each physical server is being used efficiently, and that a lot of idle servers aren’t wasting power, rackspace and cooling.
In short, I think that “storage optimization” is to making storage more efficient what “server virtualization” was to making servers more efficient.
Advertisements

3 Responses to 'Who’s Really Melting the Ice Cap?'

Subscribe to comments with RSS or TrackBack to 'Who’s Really Melting the Ice Cap?'.

  1. Jon Toigo said,

    Thanks for the comment. The fact is, though, based on interviews with many enterprise clients, that servers may generate more heat than storage, but there are more arrays being deployed. Dell says that storage has now exceeded servers as the main power pig in the data center in Plano. Storage capacity demand is driven by data accumulation and the preponderance of data accumulated today is junk.

    Based on 10000 storage assessments performed by Sun and normalized by Randy Chalfant and company, only 30 percent of the data stored on any spindle you own is business relevant. Another 40 percent belongs in an archive because it has busines value but is never accessed. The balance of the capacity of every spindle out there is a mixture of reserved not allocated space, orphan data whose owner or server no longer exists at the company, and contraband data (MP3s, Brittney Spears vids, jpeg images, etc.) With archive, SRM, and data hygiene, you could return up to 70% of capacity back to productive use, reducing the need to plug more hardware into the wall.

    I rest my case.

  2. Carter said,

    Jon, I totally agree with you. Servers are hotter, but there is simply a lot more storage, so it adds up. What I think, though, is that asking for better hygiene is too tall an order. To fix up this situation, a solution has to be really simple. Good storage optimization, in the way that I mean it – which is automatically identifying files by type and content, and apply various optimization (ie, shrinking) methods to them by policy, you can get 90% of your capacity back, not just 70%. And you can do it with way less storage admin intervention. That’s why I liken this kind of optimization for storage to what virtualization did for servers – it’s just dead simple to see where the value is and why you’d do it.

  3. Stephen2615 said,

    As I am going to visit the high arctic (Svalbard) in about 6 weeks time and have a concern about global warming, I just don’t really agree with some of the statements in this post. I had a meeting today with EMC in relation to archiving and they basically said the same thing. I would like to know where they get their info from. This is not the first time EMC have said something along those lines. The seem infatuated with mp3s and cell phone photographs. HDS don’t bother talking numbers. They know better.

    Perhaps I might be the only person in the world where a high percentage of my data is structured data. We enforce policies where all work related documents are stored in a document management database. Sure I have the ocassional pdf on my home drive but when perhaps 5000 people use about 1.5 TB of disk for home drives, you have to wonder about its effectiveness. It stops the rubbish from accumulating. I don’t care what is on the C Drive on anyone’s PC.

    We don’t have any mp3, mp4, divx, avi, etc file on our filers. We also allow PST files. Out of say 500 TB of data, around 3 is flat based file system type drives. We have other non sturctured data but it falls into a broader structured base such a SAS data which is used constantly.

    There is no way in the world that our arrays pump out the heat or chew the energy that our HP P and C class blades do. I don’t need hot and cold rows to put my arrays. I dont need special cooling systems. Disks don’t get hot compared to multi core CPU’s and when you have 5000 of them in datacenter in a very small space, it all adds up.

    One fully maxed USP in five cabinets holds roughly 300 TB. Also, not all servers use SAN. Perhaps all these metrics are for SAN enabled servers?

    I think that serious storage users have systems in place purely to stop wasting huge amounts of money. Why spend x amount of dollars on a TB with a True Copy/Shadow Image licence to store naked pics of Britney on it? If I had 20 TB of cheap SATA storage and no policy in place, I would happily share those latest bittorrent Battlestar Galactica files instead of using USB drives.

    I agree that we have data that needs to be archived but database archiving is not easy. Its expensive and most DBA’s would rather lose their left hand than trust something like IBM Optim. Its their nature.

    My previous job was also very similar. Almost no unstructured data.

    So where are all those companies that have these problems? Are they serious users of SAN? Do they have many milions of dollars invested in storage? Is their data very valuable not just to them, but to the broader population?

    Stephen


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: