Architectures are greatly impacted by the explosion of data volumes to be stored in Datacenters. As a consequence, a new standard called « Scale-out-NAS » emerged, focusing essentially on scalability, and running an “Erasure Coding” mechanism. The aim of this latter is to halve the volume of data to be stored when High Availability is required, though at the expense of a well-known drawback: a significant reduction of the Input / Output performances.
But on the other hand, most applications still require the exact same essential capabilities as before : standard File System interface, real-time access performances to any file whatever its size or usage -who would seriously claim he doesn’t want performance?-, Hardware fault-tolerance and data integrity protection, all at a controlled cost and volume. Our vision is that a really mature Scale-out-NAS architecture should focus essentially on such capabilities, while offering of course the necessary scalability level and volume savings to cope with today’s world.
That is why we choose a different approach when designing our Scale-out-NAS solution: instead of just adding a standard Erasure Coding mechanism to our system, we completely built our solution around our original Erasure Coding mechanism, in such a way that it would accelerate the performances instead of reducing it, thus preserving the four essential features required by most applications. Namely a real File System delivering real-time performances along with a high level of data protection, all at a controlled cost.
We are Rozo Systems.