Henry Ford was known as a man with a few words to say about a lot of things. One of my favorite quotes by Mr. Ford is about the variety of color options available for his cars.
“Any customer can have a car painted any color that he wants so long as it is black.” -Henry Ford
This quote came to mind when I read an article about a relatively young start-up called VAST Data claiming that they now identify as a software company. The reason this quote resonated with me relative to this news is this company has developed its solution to be very dependent on a particular hardware platform. So in the spirit of Henry Ford, VAST is telling its customers that their software will run on any hardware platform as long as it comes from the distributor they have selected. You see, saying you are a software company doesn’t magically make it true, just like saying a black car is blue doesn’t make it true.
I wrote a blog called “Betamax v. VHS,” where I talk about the challenges of betting on a technology platform that, as of yet, has not had widespread adoption. It was in response to Micron Technologies exiting the storage class memory business (SCM), where it was the sole manufacturer of the 3D Xpoint wafer co-developed by Micron and Intel. Micron had made a very calculated business decision to leave that business by saying, “Micron has now determined that there is insufficient market validation to justify the ongoing high levels of investments required to successfully commercialize 3D XPoint at scale, to address the evolving memory and storage needs of its customers.” While Micron would fulfill its commitment to Intel through the end of the year, it still left uncertainty regarding the Intel Optane line’s future that uses the 3D Xpoint technology. As I mentioned in my previous blog, I like the technology, but if it has not gained widespread adoption, even the best technology will die out sooner or later.
From an analyst’s perspective, I can fully understand why some companies choose to tie themselves to a particular hardware solution platform because it may shorten time to market. They choose the “easy button.” However, to “bet the farm” on a platform that has yet to find validation in enterprise data centers, places customers in a very sticky predicament. Customers then have to align their own strategy to this “bet,” and that’s when they end up spiraling in “vendor-lock-in.” Customers no longer have a choice due to the limitations of a hardware-centric approach. Oligarchs like NetApp and Dell EMC are perfect examples of companies who created solutions of this same ilk. To be fair, it was a very acceptable practice in the late 80s, 90s, and early 00s during the “appliance-era” to acquire packaged storage for your data center. You can hardly blame them for this model because it was well before we even thought about software-defined storage, but that is where those companies and even Pure Storage have fallen into a trap they cannot easily escape.
A look at some trends
As we look ahead or into a crystal ball, some industry analysts have said the HDD market will be killed off by Flash, be it 3D Xpoint, QLC, etc. From a financial perspective, it means you would first have to reach some kind of price parity for that to occur. Secondly, from an endurance perspective, you’d need to develop a flash technology that would come close to that of an HDD. QLC may be a low-cost leader, but its endurance is around 500 write cycles; this is not enterprise-class NAND. However, enterprise-class NAND is approximately 50,000 write cycles with a cost nowhere near parity than the HDD. It would make sense to me to use a portable software-defined storage solution to leverage both HDD and Flash in a hybrid model. Now that seems to be a sound financial model to which I believe most customers would agree.
Making broad predictions like this is very difficult. For example, many thought tape was dead 13 years ago, and we all know the “tale of the tape” on that one.
What does it mean?
Now, let’s get back to the focus of this article. What does it mean when a hardware company overnight decides it is now identifying itself as a software company because they are exiting the hardware business. Let me share my perspective. When a company suddenly decides they aren’t selling hardware but only software, it typically is based on how they recognize revenue. As you can imagine, there is a much higher margin in software versus hardware, and if the drag of hardware can be removed from the books, it makes a company’s overall run-rate look better than it is. Suppose the solution is still very much tied to particular hardware architecture. In that case, you are still forcing the customer to buy unique hardware irrespective of whether it comes from the vendor or the distributor. This is just kicking the can down the road and not truly disaggregating the software from the hardware. In the end, the customer is the one who is getting the short end of the stick.
Creating a genuinely software-defined storage solution is not easy. Creating a solution dependent on hardware is a lot easier, which is why legacy solutions like EMC Isilon and NetApp are still very reliant on their hardware and are finding it difficult to disaggregate the two, as I mentioned earlier in this article.
When you are genuinely software-defined, you have to consider “how” it will be done in software using reasonably available hardware the customer chooses, not what the vendor decides. You also look at all the new “wiz-bang” technologies out there and consider how they will work with your software stack without becoming 100% beholden to that technology platform. Qumulo knows a thing or two about disaggregating hardware from software. We had that vision years ago. I will say that Qumulo is a leader in software-defined storage architectures. It is likely the primary reason our customers have a consistent experience when using the Qumulo software platform anywhere they choose, whether on private cloud hardware or public cloud.We are genuinely a software-defined storage company, not a software-only company (the latter being a business model versus an architectural decision). We believe the customer should have the freedom to choose.
As I wrap up this article, I’m reminded of yet another Henry Ford quote that I believe fits very well here in closing.
“Vision without execution is a hallucination.” -Henry Ford
Sometimes start-up companies have a great idea and a terrific pitch deck but lack the execution against their intended vision. From my perspective, again, having been an industry analyst for several years, when hardware companies identify themselves as software companies, it is typically in response to a trend, competition, or investor pressure. When you observe shifts such as this, I believe it shows some short-sightedness for the young start-ups and a sign of desperation for the oligarchs. Only time will tell, but the good news is that if you are looking for a truly software-defined storage solution, Qumulo has you covered whether you are on private-cloud hardware or on the public cloud. We are executing toward our long-term vision to the delight of our customers.
-Chapa, signing off
Why We Don’t Use SCM for Performance (and Why You Don’t Need It Either)
Software Defined, Hardware Independent, What Does That Really Mean?
Technical Overview of Qumulo File Data Platform
Take a test drive. Demo Qumulo in our new, interactive Hands-On Labs, or request a demo or free trial.
Subscribe to the Qumulo blog for customer stories, technical insights, industry trends and product news.
David A. Chapa is a technical evangelist and strategist with over 30 years of industry executive and analyst experience helping organizations with data protection, recovery, and storage–and how to leverage cloud infrastructure, from hybrid to hybrid-multi-cloud. David is the head of competitive intelligence at Qumulo.