- Internet Protocol Security (IPsec)
ITD developed a solution to the problem of Internet Protocol (IP) packet encryption and authentication and codified its solution through engagement with the Internet Engineering Task Force (IETF). This led to Internet standards as documented in RFC 1825, Security Architecture for the Internet Protocol; RFC 1826, IP Authentication Header; and RFC 1827, IP Encapsulating Security Payload (ESP). The specified architecture applies to both IPv4 and IPv6 and is cryptographic algorithm agnostic, so both commercial and NSA type 1 algorithms are compatible with the standard.
- World’s First Implementation of IPv6
ITD developed the first implementation of the IPv6 next generation internet protocol. This allowed NRL to confirm that specifications in the RFCs were correct, and most importantly, demonstrate to industry that it was feasible to build a network protocol stack with integrated encryption and authentication. NRL made the implementation, including its source code, available to industry to facilitate broad adoption.
Onion routing is a general approach to securing both routing and traffic-flow information for network communications by separating identification of network location from routing. In basic form, it uses public-key cryptography to build a cryptographic circuit along an unpredictable path of onion routers. Symmetric-keys, established during path building, are then used to pass data back and forth between the path originator and a network destination, such as an Internet web server. The onion routing patent received the NRL Edison Invention Award in 2002 and was identified as one of NRL’s top 100 technology contributions to sea power and national security during its Centennial celebration in 2024. Numerous follow-on onion routing networks have been deployed, and now underpin a broad field of research on privacy-enhancing technologies. Onion routing’s most identifiable implementation is the Tor network that is maintained by the Tor Project, a U.S. 501(c)3 incorporated in 2006. As of 2023, the Tor network is accessed by 2 million daily users over 8,000 relay nodes, with public accounts of peak usage nearing 7.5 million users.
- Programmable Embeddable INFOSEC Product (PEIP)
The Programmable Embeddable INFOSEC Product (PEIP) is a family of programmable cryptographic devices that provide cryptographic algorithm execution for all COMSEC devices, up to 10 simultaneous COMSEC device emulations, and supports field-deployable cryptographic algorithm upgrades. The first phase of the PEIP family of cryptographic devices provided receive-only capabilities to replace the KG-3X family of cryptographic devices and was denoted as the KOV-17. The KOV-17s (and its radiation hardened variant) replaced all of the KG-3X cryptographic devices in U.S. Navy SSNs, SSBNs, and USAF Minuteman Silos.
The Network Pump is a high assurance cross domain solution (CDS) product that allows applications operating on lower security level networks to exchange information with applications on higher security level networks. The Pump provides a failsafe and redundant security architecture that maintains network and security domain separation, as well as high assurance filtering without leakage of information from the high network to the low network. The Pump is system and network agnostic and can serve several applications simultaneously.
NORM is a transport layer Internet protocol designed to provide reliable data transport in multicast groups. It is formally defined by the IETF in RFC 5740, which was published in November 2009. It includes reliability mechanisms, congestion control, flow control and forward error correction to support bulk transfer, messaging, and streaming content at optimized rates with delay and lost packet tolerance. NORM has transitioned to many materiel acquisition programs in every service branch, as well as other U.S. Government agencies and public applications.
Flying Squirrel is a software application suite that provides real-time discovery, analysis, and mapping of IEEE 802.11a/b/g/n wireless networks and Bluetooth devices. In an effort to standardize wireless security for the purpose of detecting, and thus deterring, unauthorized wireless activity, the DoD’s Enterprise-Wide Solutions Steering Group, in 2006, selected Flying Squirrel as the standard DoD tool for wireless discovery and mapping. In order to meet follow-on requirements for continuous monitoring, a variant of Flying Squirrel called Orb-weaver was developed to fulfill this requirement using multiple fixed sensors. Flying Squirrel and Orb-weaver technology helps combat the exfiltration of sensitive information over wireless signals. Flying Squirrel was identified as one of NRL’s top 100 technology contributions to sea power and national security during its Centennial celebration in 2024.
ITD developed ARCADIA, a computational model for creating intelligent agents that embodies a theory of human cognition. It emphasizes the concept of attention for how it performs perception, reasoning, and decision making. ARCADIA agents have performed well on a wide range of tasks (e.g., object recognition, tracking, and autonomous driving). Current work includes deploying an ARCADIA agent to control the actions of a turret-mounted camera that performs active perception to ensure interesting objects and events are sensed.
- Advances in Autonomy and Artificial Intelligence Goal Reasoning
ITD pioneered research on autonomous agents that can dynamically adapt their goals to their current environment. They have been demonstrated for controlling underwater autonomous vehicles (e.g., to avoid potential threats), simulated unmanned air vehicles in human/UAV air combat teams, and for satellite constellation decision making during ISR missions. NRL researchers extended these algorithms to reason with networks of goals, which supports collaborative autonomy objectives involving reasoning about the goals of human-robot teams.
ITD researchers have developed methods for reducing training time (by an order of magnitude), increasing the accuracy, and simplifying the deployment of deep (neural) networks. This has included automated techniques for tuning hyper-parameters (e.g., using cyclical methods for setting the learning rate) and for leveraging evidence of underfitting or overfitting the training data.
ITD has developed a variety of swarm intelligence algorithms for several unmanned air and ground platforms, and demonstrated their ability to learn emergent behaviors, follow gesture commands, and communicate in GPS-denied environments. Bio-inspired designs have been a strong focus; this has included using bat echolocation models to enable location and navigation of small, unmanned air systems, using slime mold behavior models to discover targets and establish communication networks among a swarm, and mosquito larvae models for locomotion of a heterogeneous swarm of air and underwater vehicles.