Pure Storage FlashArray: Enterprise-wide Benefits More Than Worth the Cost
Updated May 22, 2020

Pure Storage FlashArray: Enterprise-wide Benefits More Than Worth the Cost

James Kelly | TrustRadius Reviewer
Score 10 out of 10
Vetted Review
Verified User

Overall Satisfaction with Pure Storage FlashArray

Chapman University owns two M20 FlashArrays. One provides primary high-speed storage for our production Oracle ERP database and middleware environments, for our highest-performance-tier VMWare clusters, and for our MS SQLServer-based university data warehouse. The second M20 provides ultra-high-speed shared scratch storage for our primary High-Performance Computing (HPC) computational-research supercomputing cluster. Our entire campus community, faculty, staff, students, researchers, use the resources hosted on these arrays every day. The applications, databases, and research workloads running on these arrays are literally how the university does its work.
  • Our FlashArrays have been 100% reliable. In the four-years-plus we have been relying on them, we have not experienced a single moment of unplanned downtime due to these arrays. Their hardware design is solid-state and fault-tolerant, redundant everywhere. This allows zero-downtime upgrades for all of their OS, management, and driver software. We've been through numerous (10+) software upgrades on two arrays and have never experienced any interruption in service or data loss.
  • Pure Storage's FlashArray support, phone-home, monitoring, and analytics services through their centralized Pure1 web portal have been head-and-shoulders above every other storage vendor we've worked with. Their tools allow their staff to be expert, professional, confident, and, best of all, proactive. Their platform allows them to know far more about the array than we do, and in real time.
  • Our FlashArrays have been blazingly fast. When we copied our production Oracle databases over to the first array for the first time, the databases immediately benchmarked out 2.5 to 3 times faster, with no other changes. Every one of the workloads we migrated to our FlashArrays experienced similar speed increases.
  • Compared to legacy hybrid or spinning-disk storage-array administration, managing our FlashArrays is extremely simple. No more external spreadsheets tracking raid categories, LUN sizes, or storage pools built on different underlying types and arrangements of disks. Allocate a volume, give it a name, hook it up to a host or a cluster. Rescan the storage, and start using it.
  • Pure's implementation of data reduction/data deduplification has been amazing for us. Absolutely a game-changer. It has allowed our arrays to function like arrays five or even ten times their actual size. It works especially well (really, really, really well) with large database and database-development environments, where dev/test/qa databases are close copies of production databases. You basically get all of the copies for free, storage-wise. Something very similar happens with array storage for VMWare VM's, also. Copies of VM's are basically free.
  • In our environment, Pure could provide additional network interconnect technologies. HPC is not that common a market niche, though. Pure could at least test and validate limited compatibility with third-party hardware, even if they then needed to limit their support or warranty if customers chose to implement it.
  • Pure provides a lot of API support for building your own integration solutions for things like VMWare administration, container orchestration, and hybrid cloud environments. They do seem to be investing in these technologies, but full-featured, pre-packaged, tested and supported solutions would be a lot better. Most enterprises don't want to build their own automation and integration, they want things to just work, out of the box.
  • Our FlashArray implementations have had extremely positive impact on our business goals. Our main, heavy-lifting Oracle and MS Sql databases and apps run much faster and provide much faster response times. Our data warehouse ETL completes in much, much less time. (From >12 hours to
Working with Pure's support team is 180 degrees different than working with any other storage vendor we've ever dealt with. Their technology and their product design makes their job a lot easier, because of its straightforwardness and reliability. Their centralized phone-home/management/analytics support platform also means that they will always know everything about your array that they need to, in real time. You don't need to run tests or gather data or run some kind of utility that you then send the results of to the support team. They already know. This allows them, sometimes, to even be proactive, making you aware of some emerging issue, and then fixing it, before you would otherwise have even started to notice it.
The FlashArray beat the hybrid arrays in performance, the other all-flash arrays we evaluated in price, and all of the legacy spinning-disk arrays in ease of management and in reliability. Their two huge advantages were their "Evergreen" upgrade-everything-forever option and the visibly different Pure Storage organizational culture. Pure was just so much better to work with as an organization. The difference was night and day.
  • Pure's truly excellent support for automation and integration programming has allowed us to build a completely new database- and application-automation solution. This solution allows us to clone, manage, and deploy complex, integrated database-plus-application environments as dev/test copies in less than 10% of the time it previously took, using less than 10% of the previous DBA/SAN admin effort. This has allowed our development teams to complete their projects faster, delivering features and performance to our entire campus community, increasing customer satisfaction and other higher-ed key performance metrics.
All-flash storage is expensive, and full-featured all-flash, like the FlashArray, is even more expensive. You need to have a need for it. If you use virtual desktops in your environment, or if you have performance-sensitive databases or database-integrated applications, or if you are doing any sort of significant machine-learning, deep-learning, or data analytics, these are all workloads that will benefit from FlashArray storage. Anyone of those, much less all of them, would easily rationalize spending the extra money to get the benefits of a FlashArray.