How NOT to Build a Storage System: One Piece at a Time

Over the last month or two, I have had the opportunity to chat with about 50 members of the press and analyst community as well as just as many end users of enterprise data storage. We have been talking about all of the new flash storage companies in our market and what will separate the winners from the losers. We have seen so many flash in the pan (pun intended) startup storage companies that have dropped a dozen or so SSDs in an off-the-shelf enclosure and hung a “Flash Storage Vendor” sign in their window. The barriers to entry into this level of the market are almost zero. Problem is (and most of us that have been in this market for some time know this), that in storage, it takes a platform, not just a product, to have a sustainable data storage system business.

This question reminds me of an old Johnny Cash song: “One Piece at a Time“. The story told in the song is about a GM factory worker who is on the Cadillac assembly line lamenting about how he’d love to be able to have one of those beauties. Unable to afford one of his own, he decides to steal “one piece at a time” over several years and build his own Cadillac in his garage at home. Sounded like a great plan, but once things started coming together, it got really ugly really fast.


This song came to my mind when I was talking to a member of the EMEA press about how difficult it is to add foundational pieces of an architecture after a product has been in the market for a few years. Block only protocols on your array? Adding NFS and CIFS/SMB will be a big challenge, and gateways are just a poorly fit car door that will fall off in time. No in-line deduplication today? Enjoy post-process dedupe jobs at 2AM for the rest of the service life of that array – kind of like adding a quart of oil into your junker every 500 miles.

This was a huge part of what made me decide to join Tegile almost three years ago. I saw that our founders had put together an excellent foundation that could be built upon. Having the foresight to put foundational elements into our platform puts Tegile in a great position to excel in areas of the market that are just nascent today. I am sure you can think of several of them right now. There are so many pillars that have been driven down to bedrock in our platform – it is amazing. I can’t wait to see how Tegile’s roadmap comes together over the next few years.

Roadmap aside, there is one other “strategic imperative” built into our platform. Tegile believes that there will always be a subset of an enterprise data that is very performance biased and there is another subset of enterprise data that is capacity biased. It is our job to address both of these requirements and make the task of balancing the tension between these two data subsets as transparent to the customer as possible. This strategic imperative says that today, eMLC SSDs are best in class for performance and 7,200RPM SAS drives are best in class for capacity. We know that best in class technologies will change, and that’s good for everybody. Having the foundational architecture to incorporate those new technologies into our product line is the beauty of IntelliFlash.

So, as you evaluate alternatives in modernizing your data storage, or any other technology for that matter, strap on your headphones and listen to that old classic “One Piece at a Time” and look for a platform to build upon, not just a product.

Leave a Reply

Your email address will not be published. Required fields are marked *