The grant enables NGD Systems to continue the work originally funded in the 2016 Phase I SBIR grant on developing applications for computational storage through in-situ processing for growing petabyte-scale analytics workloads. The award represents continued commitment and further confidence in the deployment of computational storage and in the ability of NGD Systems to deliver market-changing products that embody in-situ processing.
“We are excited to be the recipient of this grant from the National Science Foundation,” said Vladimir Alves, chief technical officer, NGD. “The SBIR Phase IIB program is granted to companies whose achievements in the SBIR Phase I and Phase II programs demonstrated both technical merit and commercial potential. This award confirms the confidence of the NSF in NGD Systems and in the commercialization of computational storage.“
While the concepts of computational storage and in-situ processing exist, a recent webinar survey showed that concerns about the maturity and stability of computational storage vendors is one of the concerns of end-users. This grant by the NSF demonstrates the US Government’s confidence in both in-situ processing and in NGD ability to deliver enterprise products on the application of computational storage to petabyte-scale analytics workloads.
“Flash has transformed storage performance, and the NVMe protocols are now helping to extract even more speed from solid-state storage,” said Tim Stammers, senior analyst at 451 Research, LLC. “But while these changes have goal posts are being moved by emerging applications such as AI, which need yet faster access to data. That is driving a requirement for new architectures that will boost storage performance, while containing infrastructure costs, and taking advantage of the promised benefits of storage-class memories. Because even small amounts of data have gravity, conventional architectures need time to move data from storage devices into the processors that need to access it. NGD’s approach is instead to move the processing to the data, effectively making the storage smart.“
Regardless of use in edge computing, IoT applications, or real-time analytics, computational storage offers a means to nearly eliminate data movement through a system for analysis, which reduces the overall latency of the solution.
At the same time, NGD patented storage architecture can reduce the server footprint of these solutions through the largest capacity, intelligent storage devices. This is important for embedded applications encompassed by content delivery network (CDN) points of presence (PoPs), IoT, and edge computing. The NGD Newport platform, the latest in computational storage, builds upon the family of computational storage devices in use in the market.