Flash storage seems to be the popular kid on the block right now, with companies like Fusion-io sparking debate over the technology’s use at the enterprise level, while also incurring interest from the startup sector regarding innovative ways to make it a scalable and affordable storage solution.  FlashSoft is one such company, launching its shipping product, which is designed to eliminate IO bottleneck and improve enterprise application performance.

FlashSoft, as the name implies, is an all-software solution for freeing up server computing resources, leveraging flash as cache. This resides at the SSD level within the software, operating on the concepts of “tier minus one” to modify the inherent hierarchy of existing server stacks.  Placing the storage as close to the server as possible reduces processing and data transfer time, and to do so adequately, FlashSoft concentrates on keeping only the most important data at the SSD site.

The company has a grand industry welcome, with $3 million in funding led by Thomvest Ventures.  The introduction of its first commercial offering delivers FlashSoft SE, a product that’s already being used in business-critical production environments.  The first product supports Windows server platforms, with plans for Linux support rolling out in beta this summer, and virtual environments including VMware later on.  Zenprise is a launch partner, acting as an early supporter of a new system for managing data at the server level.

“Flash virtualization is a way for us to utilize software; to make the OS and appliances believe that there’s actually more flash than is actually present,” says Ted Sanford, CEO of FlashSoft.  “By representing the hottest data in cache, it appears there’s more data than there is.  It’s the 80/20 rule.  Instead of buying enough flash to handle it all, buy enough to handle the hottest 20%.”

Of course, determining what that 20% is becomes the real issue for FlashSoft to address.  It’s developed Active Data Management to handle this, a patent-pending piece of software FlashSoft uses to analyze the very behavior of data.  FlashSoft is cluster-aware, meaning it’s able to recognize other hot data caches in machines across the entire cluster.  What this does is provide a wider pool of data sets to understand data relationships, in turn allowing FlashSoft to influence that data in a uniform way.  It’s the persistent data caching in the server environment that signifies FlashSoft’s advancements around flash storage, making it an effective alternative even to some hardware solutions.

The type of data saved on your SSD is of no consequence; FlashSoft has read and write capabilities, attending to varied enterprise needs for applications.  Some may be sensitive to writes, getting the data written as quickly as possible to open bandwidth and reach.  FlashSoft has spent some time learning the inefficiencies data managers were dealing with, building solutions into their code.  With the vast majority of its code self-contained, FlashSoft is also pushing access through an API for quick porting to Linux and beyond.  This will become an important part of FlashSoft’s distribution and growth strategies, recognizing early on that usability and adoption on its technology relies heavily on easy integration with existing environments.

The way FlashSoft manages data, particularly through its Active Data Management system, is what sets its product apart.  The concept, however, isn’t entirely new.  Companies like Xiotech are also looking at algorithm-based elixirs for improving data management within the flash storage industry.  Both FlashSoft and Xiotech are certainly validating ongoing developments in this space, even as Fusion-io’s successful IPO launch demonstrates that these companies just may be on the right track.

“This is the closest storage has ever gotten to the CPU, if you consider this storage” Sanford says.  “You’re going to see people look at this as storage hierarchy, and it’s blurring the line.  Before, the difference was pretty clear.  By moving potentially terabytes out of the domain and into the storage, [the CPU] is its own storage.”

For more on virtualization’s affect on the storage industry, read Wikibon analyst Stu Miniman’s insight here.