- NVMe Storage
- Enterprise NVMe
Our engineering team comprised of industry leaders sat down to solve the decades old storage bottleneck, not to find just another work-around for resource heavy, latency burdened protocols and software defined storage that require scale-out of not only the hardware and software, but IT budgets as well.
Our team started with fresh eyes on the challenge and asked a very logical question. How do we remove the problem altogether? It was simple, remove the controller, remove the archaic storage protocols, and remove the slow fabric networks that tied it all together, no small feat. But that’s exactly what we did. We must confess, we had a little help, the advent of both the NVMe protocol and the proliferation of storage aware applications and frameworks such as Splunk, Hadoop, Spark, Security Onion, Cloudera, Hortonworks and more, that mitigated the need for additional storage management utilities, created the perfect combination for our technology to thrive.
What makes companies like Intel, NEC, Splunk, Hadoop, and some of the word’s largest IT distributors take note? A paradigm shift in the way the future of storage will be managed. Yes, there are other NVMe solutions available, but all of those options pair NVMe drives with sluggish legacy technology, slow hardware controllers, cumbersome protocols, and lots of software code to manage the stack. It seemed like a logical first step, NVMe over Fabrics (NVMe-oF). Leverage the old technology to introduce a new one, a gentle easing in of the future. Although a great improvement over the status quo now existed, the bottleneck was still present. Sure there were speed improvements, but it still required racks and racks of server captive flash connected by slow networks, and bloated software, most of the technology advantage afforded by NVMe went untapped. That is until Aperion developed the world’s fastest and only networked storage platform utilizing native NVMe technology across a built in 40Gbe network. What’s more is we removed the storage controller completely by designing HBA’s and built in array network switches that take advantage of Intel FPGA’s to deliver the native data packets encapsulated within a 4K wrapper across Ethernet (NoE). The answer was a novel, hardened layer 2 Ethernet transport protocol to facilitate the delivery of data. This represents a paradigm shift in network storage, and one that is being validated by such industry giants such as Intel. Not since the advent of SAN and NAS technologies themselves, has there been such a dramatic shift in storage than with Aperion’s native NVMe technology. It’s unthinkable, but true.
If your application is storage aware and is craving more high performance scale-out storage of any capacity, you must consider using the ADS platform or suffer degradation in application response, and deal with increased hardware and software budgets, and losing ground to the competition.
Each 2U ADS 1000 provides a tile and budget busting 18.4M IOPS and 72GB\sec throughput that’s perfectly scalable to 1000’s of NVMe drives and 100’s of millions of IOPS, each and every drive and enclosure presents across our HBA’s as sever captive storage, often at faster than server captive speeds.
If you are looking at NVMe storage for your enterprise, Apeiron is the clear leader, and in a class all by itself. The ADS provides unparalleled scalability, untouchable blazing fast performance, and the greatest density footprint by a margin of any scale-out storage platform available today.