SolidFire_logo_RGBI’m on my way home after a couple of days spent in Boulder (CO) with the SolidFire team for their annual meeting with analysts.

My opinion about their product and technology were already clear but, this week, I had the chance to talk with various customers and executives and now I have an even clearer view about this company and its potential.

Solid customers

I’m going to start with the customers because I think it was one of the most important moments of the meeting. In fact, SolidFire’s customers (mostly big ISPs, *aaS providers and large enterprises) have chosen to work with SolidFire for its basic characteristics (scalability, QoS, high availability, resiliency and automation capabilities) but then, what seemed to come up repeatedly from the end users was “because it [SolidFire] lets me sleep at night”. Considering that these customers are consolidating huge amounts of diverse workloads (and data) on the same clusters in a relatively short time this is quite remarkable.

As is common for most startups, SolidFire listens to its customers and reacts very quickly. This is another plus for these customers; They want from vendors the same agility that is asked to them!

Solid Design

I’m going to repeat myself by saying that SolidFire architecture is on top of the chart when it comes to scale-out flash storage. But I want to add a couple of interesting points that you should take into account if you are looking at similar systems.

A new type of node, the SF9605, was announced at the event. This makes 4 different available models now, and they are all mixable, even with older models, in the same cluster. It’s not only about freedom of choice for new deployments, but it also means longevity and flexibility. I think this is a big plus, especially when it is officially covered by the guarantee program (FlashForward) announced during the event.

New features that are added with a notable frequency without visible changes to the system architecture (the 8th software release was announced at the event). This, in conjunction with the fact that all the upgrades are carried out in a non-disruptive fashion, is quite remarkable, and I don’t think this will be changing in the future. In fact, SolidFire’s is now leveraging its own cloud-based analytics as welll (Active IQ), which is used to improve support and give users smart reporting tools, but also to collect useful information to develop new features and improve the product faster. This solution is not unique, but vendors working this way are still very few at the moment.

Solid strategy and vision

Infrastructure Consumption Continuum GraphicIn these two days, SolidFire has clearly demonstrated the potential of its solution, where it is positioned and where it is heading.

The target markets are still ISPs and large enterprises (which, in the last Q, have counted for 54% of the bookings!). In practice, organizations which are looking for high performance, highly scalable and automated storage for primary data. The set of capabilities available in the product largely concur to position this solution a step above other All-Flash arrays in some specific scenarios. For example, and on the contrary of what you usually see with other storage vendors, a large percentage of SolidFire’s customers are deploying OpenStack in production.

The announcement of ElementX, a fully-featured software-only version of the product for customers who want to buy only SolidFire’s OS and install it onto their hardware, totally confirms this strategy. In fact, this is much more than an attempt to separate the HW from the SW. ElementX is intended for hyper-scale end users who want to maximize the efficiency of their infrastructure without compromises. This means that SolidFire collaborates with these end users to validate and support specific hardware configurations, leaving them all the flexibility for deployment, HW support and management.

Closing the circle

Solidfire is one of the most interesting storage startups out there, they produce an All-Flash array like many others but, because of its specific characteristics and positioning, it competes at best in the high-end segment of the market and very high demanding environments. In fact, even if the starting price for the solution is around $100K (4nodes), it’s not unusual to see much larger installations.

My only concern, if I can be a bit picky, is about the potential of their architecture in covering smaller deployments. If not for serving the low-end, just because large organizations have remote/branch offices and sooner or later someone will be asking for a solution for those smaller environments.

Disclaimer: I was invited to this meeting by SolidFire and they paid for travel and accommodation, I have not been compensated for my time and am not obliged to blog. Furthermore, the content is not reviewed, approved or edited by any other person than the Juku team.