While everyday computer users are getting excited about the release of the newest two terabyte (TB) flash thumb drive, those in the IT and DevOps (development and operations) sphere are excited about the release of a new single all-flash SDD with a capacity for 60TB. But large businesses need more than terabytes for storing data, proprietary information, software, applications and room for developmental environments. Flash storage arrays with SSDs take up only 25 percent of the space HDDs do in the data center. In fact, Tintri’s all-flash storage system with the ability to support between 19TB–40PB uses just two rack units. And that’s just one upside to flash storage.
Flash storage that’s architected on applications is hands-down the fastest method for scaling up and tearing down applications, accessing data or erasing obsolete information. It delivers around 10 times the speed of legacy storage types, including HDDs—and it does it using one-tenth of the energy used by HDDs and physical servers. Where admins once waited minutes—or hours—for operations to complete, they now wait a few seconds. Or, with extremely large workloads, they may wait a few minutes. This remarkable rate of input/output (I/O) difference saves time for companies and for the consumers or clients they serve.
In a legacy HDD-driven data center created for storage, networking and computing, each of these disparate areas are siloed, which makes them inefficient for the company coffers and otherwise. Siloed storage is also difficult to build, and even harder to scale up from when HDDs are in use. What's more, the separation of all these environments means a lot of hands-on work for IT teams: manual configuration and processing are everyday events, and even then, siloed storage systems remain totally inflexible.
With flash memory and storage in application-level AFAs on SDDs, siloes become history, because flash erase the need for silos and removes the need for complex manual processes that still leave users with an inefficient system. Application-level flash allows the enterprise-level user to handle multiple workloads and share resources across their entire infrastructure simultaneously and without latency. With application-level flash storage, system updates, deployments and managing storage and applications is done at lightning speed when compared to its sluggish HDD counterpart.
In traditional data centers, the backup and management needs are endless, and require the effort of an IT army. When replaced with all-flash arrays on SSDs built for application-level operations, all these problems simply disappear. Business networks using application-level AFAs in the data center won’t only move faster, workers at the CPU-level get more done in less time. And CFOs should love flash, too: it slashes company cost of repairs while employees enjoy newer, faster applications, infrastructure and other features to speed up turnaround times. Human resources will also love having application-level AFAs in the data center. With automation and autonomy provided by application-level AFAs, IT staff has more free time on their hands. They can step away from data centers and focus on DevOps, where businesses gain more insight from their acumen and up-to-the-minute knowledge of DevOps, making better use of their time and the company payroll.
At the core of the legacy SSD or HDD are logical unit numbers (LUNs) and volumes: these are slower than flash memory by a huge factor, but more importantly, they don’t speak the language of the cloud enterprise—or any cloud, for that matter. Conversely, at the core of flash architected for applications are virtual machines and containers: these produce the high-end performance and agility businesses need for quick turnarounds, speedy deployments and more I/O per second (IOPS). The right architecture for enterprise cloud are VMs and containers. Evidence has shown that retrofitting LUNs and volumes within legacy storage just doesn’t create the same level of agility—and certainly doesn’t offer nifty features like automation, autonomy or real-time and predictive analytics.
Despite common rumors about how costly AFAs are, these anecdotal commentaries leave out the cost it takes to run HDDs within the data center. At the outset, flash storage may cost more per gigabyte, but what won’t you have to pay for when you make the transition to flash? To begin, you’ll lower your carbon footprint by reducing your need for all that cooling, which will immediately show up on your next monthly bill. You’ll be taking up a fraction of the space in the data center, so there’s another penny-saver. Because flash runs without a need for much maintenance or management, you won’t need that army of IT professionals anymore, so they become a more productive part of your data science or development teams, which could actually increase your earnings. And, of course, you’ll be getting a lot done in less time, thanks to speedier scaling up and tearing down, execution, deployment and computing with application-level AFAs.
As the old adage goes, time is money. The Tintri All-Flash Array offers all the benefits discussed above and more. Our AFAs bring public cloud-like agility to your enterprise cloud, offering the ability to isolate applications (goodbye, noisy neighbors), automate tasks that you once had to do manually and get real-time and predictive analytics.