Ask any project manager if it’s possible to deliver something that is fast, good, and cheap, and they’ll laugh. The phenomenon known as the Iron Triangle limits just about everything in the world from meeting all three conflicting requirements. Yet, for the last two decades, enterprise storage array vendors have been trying to deliver just this. How’s that working out?
Today’s enterprise storage arrays are truly “Jack of all trades” devices. They exist at the center of the network, delivering reliability, high-availability, data services, performance, and plenty of capacity all in the same box.
Here’s the Rack Endgame series:
The Fat Middle
Take a look at the latest announcements from companies like EMC and Nimble Storage and you’ll find boxes that deliver both scalable capacity and remarkable levels of performance, at least compared to previous devices. They achieve this thanks to advanced engineering and liberal application of flash memory, the latest panacea of storage architecture. And they do a remarkably good job: so good, in fact, that today’s Nimble or CLARiiON is competitive with the highest performance offerings of just a few years ago.
Sitting at the middle of the capacity/performance spectrum is also good product marketing, since this is the fat middle of the storage market. Today, IT buyers want a storage array that delivers everything to every application, and these “fat middle” arrays are exactly what the doctor ordered. With centralized network storage forming the core of modern IT architecture, especially thanks to the success of VMware, these arrays are today’s gold standard.
Networked storage was important back in the 1990’s, but today it has become a necessity. With VMware taking over the data center, storage buyers have no choice but to use SAN or NAS. For too long, all the cool features of VMware vSphere required shared storage, and this has meant Fibre Channel, iSCSI, or NFS. Therefore, it was natural for such shared storage systems to become the core of the data center.
Smart Marketing
Logically, this is where companies focused their development effort. This is why so much of today’s storage development is on centralized SAN and NAS systems that would be at home in the 1990’s if not for the exceptional advancements in scale, performance, and data services. Look at today’s exciting announcements, companies like Data Gravity, and consider what they’re building: A better CLARiiON with cool, innovative features.
It’s a smart strategy to give customers what they want and compete with the best selling products on the market. That’s one reason why enterprise storage architecture has changed so little over the decades. Networked shared storage has been a good idea since the 1990’s, so why not just keep making it better?
And don’t get me wrong: It is better! Today’s systems are remarkably fast and represent a great value for buyers implementing a traditional centralized storage architecture in the data center. The storage state of the art has evolved from fixed, monolithic RAID arrays to today’s more-scalable clustered modular solutions. Automatic storage tiering (between flash and disk) is table stakes, as is data reduction technology and integration with higher-level systems like VMware and Windows Server.
These systems are so good that cutting-edge solutions like all-flash arrays and object stores are finding it difficult to compete with mainstream monsters. After all, why should buyers care about picky technical details when a Nimble performs so much like a Pure? These systems compete for the same buyers, rendering comparisons like Gartner’s Magic Quadrant for Solid State (or all-flash) Arrays irrelevant. Mainstream storage is as mainstream storage does, regardless of whether it’s all-flash or not.
Stephen’s Stance
It has taken extreme engineering resources to mix scalability and performance in systems like these, and my hat is off to the engineers have accomplished it. It’s also incredibly savvy of these companies to focus on the middle of the market instead of driving the edges outward. But perhaps things are changing, and we might see these “fat middle” storage arrays in a different light very soon. See my next entry in this series, “Virtualized and Distributed Storage: This Time For Sure!“.
Note: These are topics I discuss in my public speaking engagements, including my Truth in Storage seminar with Truth in IT. Check out the schedule for Truth in Storage and come listen to the whole story!
Disclaimer: I work with EMC, Nimble Storage, Data Gravity, and most other companies in enterprise storage with Foskett Services and Tech Field Day. I don’t think I’m biased, but you can draw your own conclusions.
What about a REAL cross-vendor comparison. Looks you like YOUR boxes.
What about the move away from mainframe era storage paradigms? Centralized storage still very much resembles the 90s, but Microsoft’s Storage Spaces (NFS compatible) seems like a really interesting approach to eliminating the centralized storage tax (http://www.stackaccel.com/the-enterprise-storage-market-in-2014-storage-for-the-rest-of-us/), with the potential to erode the margins enjoyed by the likes of EMC.
Netapp is part of the conversation as well. They didn’t need mentioning, I guess.
I’m afraid I don’t follow. What are you trying to say?
I couldn’t possibly mention every company in this space. It represents just about the entire storage industry. I hope you understand!
If you’re going to talk about the “Fat Middle”, Tegile should definitely be a part of the discussion. Their arrays can scale within the “Fat Middle” between performance and storage needs and they are compatible with Fibre Channel, iSCSI, and NFS.
Netapp? Never heard of her.