Original posted at https://www.freeitdata.com/.
Buzzwords suck. That’s right, I said it. Hybrid, All-flash, Converged, Hyper-converged, Data-aware, VM-aware, Software-defined, Object Storage, BigData, Scale-out – we get it, but why is it all so confusing? Or better yet, why is it all so similar?
In Spite of the Buzz - A Win for the Customer
Despite the overuse of buzzwords and re-classifications, the storage industry has actually seen dramatic improvements over the last 3-4 years. All of these changes are great for the customer. More efficient technologies & more competition have lead to lower operating cost, better pricing, and better solutions. But what good is saving all that time and money implementing and using these products, if it takes just as long to weed through all the jargon to find the right one?
Technology-Defined-Storage
There are plenty of options and a ton of overlap, but each of these unique solutions has a place in today’s modern data centers. Let’s take a moment and weed through the buzzwords to get a better understanding of where each one fits best.
-
Performance Centric – The top tier. The absolute must have NOW situations. There are several solutions that will fall into this section. Typically, the all-flash technologies that promise 300k IOPS. These types of solutions are best suited for your production database or VDI environments, but are often limited by budget and capacity.
-
Me Too (Hybrid) – The two trick ponies. These solutions typically combine a flash element with spinning disc to create a “best of both worlds” scenario. You can get speed and capacity at a reduced cost when compared to the performance tier. They also aren’t the cheapest, or the fastest.
-
Designed for the virtually minded – Sliding all my money to the middle of the table betting on “virtualization black”. Very specific and very good at what they do, these solutions focus on how they function within virtualized spaces.
-
Cheap-n-deep (Capacity Play) – Yes, they still have a place in our hearts. The “box of drives” are still very useful inside today’s modern data centers. Perfect for backup/archiving and large amounts of unstructured data. Typically these boxes won’t overwhelm you with performance or features, but they fit the wallet nicely.
-
All-in-one (Converged + Hyper-Converged) – I like Legos too. Like building blocks, these solutions allow you to group compute, storage and networking into one device and stack them as needed. This makes it super simple to manage your environment and allocate assets, but it’s an all or nothing type buy.
Below we have dissected a few of the industries leaders to look at IOPS, Capacity, cost and how they compare. This is a great snapshot, but doesn’t by any means tell the entire story.



Software Makes All the Difference
It’s all in the software. Obviously, reliant to some extent upon the hardware, the software really determines how the important stuff is handled...the data. De-dupe, compression, hot data, cold data, these features all play a big role in the IOPS and capacity capabilities of each solution. Not to mention the reporting and administration capabilities provided by these unique and elegant software platforms. Many of the hardware components inside each of these “boxes” are virtually the same. They are manufactured by the same companies, assembled in the same manner, with same CPUs, the same RAM. The software layered on top of this hardware really defines its capabilities.
Fit-Defined-Storage
In a perfect world, we just look at the speeds and feeds, features and functionalities and find the best technology to fix the problem, but there are many other business considerations when evaluating data center technologies. Installation, integration, usability, performance, the list goes on and on, but budget is often the biggest one.
Uh oh. More buzzwords - TCO, cost per GB, cost per IOP. All just ways to assess whether those features are worth the money. We couldn't just look at cost as a single determinant on picking the right solution, but looking at it relative to the performance and capacity is one easy way to determine “value.” Here is a look at the same data above by cost per GB.

How about looking at it by cost per GB, per IOP.

One size doesn’t fit all. At least not when you factor in more than just size. The process of evaluating solutions can become long, complex, and costly.
Scale-out, IOPS, TCO, cost per GB…all of these things together with budget, timing, integration & ease of use factor into finding the right fit. By themselves they’re just features, much like buzzwords.
Disclosure: My company has a business relationship with this vendor other than being a customer: We are vendor-agnostic implementers.