DON26BZ01-NV022 — Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development

Award Maximum: $140,000 (Base) / $100,000 (Option) Period of Performance: 6 months (Base) + 6 months (Option) Phase Type: Phase I

OBJECTIVE: Develop a small and dense data recorder that can store >= 8 Petabytes of information in <= 4 u of 19-inch rack space, will be scalable and flexible in nature, and will demonstrate the interfacing to >= two different interface protocols each supporting > 400 GB/sec data transfer rates for >= 30 seconds.

DESCRIPTION: In today's environment, emphasis is put on how Artificial Intelligence/Machine Learning (AI/ML) can solve most of the Department of War's (DOW) problems as long as the AI/ML algorithms are trained correctly. This training requires vast amounts of relevant data. Unlike commercial websites where the algorithm developers can have the public train them based on security selection images, the DOW does not have vast stores of relevant data sets much less a global community to train the algorithms. Unfortunately, very few to none of the fielded program of record (POR) systems have the ability to record (at the tactical edge) relevant data products in sufficient quantity to help algorithm developers.

This SBIR topic is intended to develop extremely deep sensor data recorders for implementation/fielding on tactical platforms for tactical sensors at the tactical edge. These recording devices must be able to be integrated easily into the platform's sensor suite and be able to record the relevant data products for use in future algorithm development and training.

These recorders must easily adapt to various networking infrastructures (e.g., InfiniBand, NVLINK, PCIe, and or Ethernet, etc.) and support the extreme streaming bandwidths for wideband (500Mhz and greater I/Q data) Radio Frequency (RF) digital data and high definition (4K or greater) streaming video. These recording devices must be scalable in nature, at a minimum take up less than or equal to 4u of face plate volume in a 19-inch rack, and record greater than 8 petabytes of storage.

These devices must meet all NSA data at rest encryption requirements and be developed in a manner to easily acquire a volatility certification letter. These prototype devices will be installed on manned and unmanned platforms and must be developed with remote and/or autonomous operations in mind.

Key requirements: Less than or equal to 4u of 19-inch rack volume; Must meet class B shipboard installation Environmental Qualification Testing (EQT); Greater than 8 petabytes of data storage; Must meet data at rest security requirements; Must meet non-volatility certification requirements; Have networking architecture demonstrating ability to configure to multiple types of networks; Have a minimum of two different networking options where each networking option can sustain > 400 GB/s data rate; Compliance with shipboard installation environmental qualification requirements; Ability to perform data at rest encryption and the ability to meet volatility requirements for system posture changes; Ability to consume data from a defined sensor and parse/tag this data; Ability to record and playback from both local and remote users.

Work produced in Phase II may become classified.

PHASE I: Develop and provide a detailed schedule out through Phase II options, as well as a detailed technical description as to how they will achieve success. The initial deliverable of the Phase I award will be a kickoff meeting detailing how they will get to the final briefing. The final briefing will show specifically how to meet all key requirements listed in the Description. Phase I option will be showcasing software modules and fundamental breadboard designs and present the detailed plans for Phase II and Phase II option.

PHASE II: Hold a kickoff meeting with a detailed development plan including costing (recurring and non-recurring separated) development; and detailed security and testing plans. These plans will include detailed: technical plans; security plans; EQT plans; lab testing plans (both at developers facility and at government labs) utilizing different types of networks; ship installation and at sea testing. There will be at a minimum two lab demonstrations at the developers facility and one integration and demonstration at a government lab. It is probable that the work under this effort will be classified under Phase II.

PHASE III DUAL USE APPLICATIONS: The awardee will clearly and in detail describe how this capability will transition to a Navy program of record (POR). Any commercial industry looking for low cost optical processing on extremely large data sets will benefit from this technology.

KEYWORDS: Digital Recorders; Radar Interfaces; Combat Control Interfaces; Electro Optic data; Continuous Digital Intermediate Frequency; CDIF; Burst Digital Intermediate Frequency; BDIF