Today, every business is powered by software, and having a DevOps strategy is key to innovation and developer productivity. Pure can help. A data-centric architecture from Pure Storage ...
First off, I’ll apologize for a bit of a blogging hiatus, we’ve been pretty busy over here for the past few months getting customers up and running on the Pure Storage technology through our Early Adopter Program. The EAP allows customers to try the Pure Storage FlashArray in their environment at no cost, then transition directly into a production deployment upon successful validation in their workload. So far we have over three dozen customers deployed on the Pure Storage FlashArray, and we work with those customers weekly to understand their successes, challenges, and how they are evolving their application environments for flash. Through the process we’re improving our product, and we’re learning a lot about the real-world benefits of flash in the enterprise data center, and how to prepare customers for the improvements that flash makes in their application infrastructure.
Let’s start by looking at the program overall to begin with, and then we’ll share some specific quotes and results at the bottom of this post.
Not surprisingly, many of our early adopters are typical technology early adopters: tech companies, web companies, and finance companies (Wall Street, hedge funds, and insurance companies in this case). That said, these typical early adopters represent less than half of our adopters, the remainder are what you might call decidedly “mainstream” companies: manufacturers, agribusiness, educational institutions, pharmaceutical manufacturers, law firms, etc. It turns out that flash storage appeals to a lot of people, in a lot of different industries and use cases, it’s not just for Wall Street and un-mentionable government agencies. Remember, Pure Storage isn’t just about performance, we’re focused on delivering a 10x improvement in performance, power consumption, space usage, and administrative efficiency — all at the same cost or less than traditional storage.
OK, so what are they running on Pure Storage? One of the nice benefits of flash is that it serves multiple workloads well – no need to worry about spindle contention! So the majority of our customers are actually testing and deploying multiple applications on a consolidated Pure Storage environment (this is in stark contrast to most PCI flash or flash appliance deployments to date, which are generally geared towards accelerating a single problematic application). As such, the number of applications above is greater than the number of program participants to date (>50 different application environments). Here are the major use cases:
The really fun part is starting to see the success stories roll in…we know that putting 10x faster storage under an application can have profound impacts, but we’re never quite sure what real value the customer will actually realize, as it is totally tied to their unique application. Will they see a dramatic improvement in end-user experience, or will their application bottle-neck somewhere else? Here’s a few of the results we’ve seen thus far, in our customers words:
While the realized performance gains are awesome, it’s actually the last comment that I’m most proud of. One of the consistent themes from many of our customers is how blown-away they are by the simplicity of the Pure Storage FlashArray. All-flash storage isn’t just about performance, its also about a whole new level of operational simplicity, completely removing scores of tasks that plagued disk storage administration from the storage admins life (RAID management, performance troubleshooting, LUN “carving”, host/array block alignment, etc.). We hope people come give us a spin for the performance, but we believe many will stay for the simplicity.
Every Pure Storage array calls home nightly to report any challenges, as well as usage information around realized performance and data reduction….so we get a real-time view into what level of performance we are delivering for folks, and what level of data reduction they are getting on different workloads. But…we’ll save that data for a future post coming soon 🙂