Computing Architectures

Computing Architectures concentrates on understanding and exploiting the fundamental aspects of hardware and associated system software for emergent and future computing architectures for mobile, scientific, and data intensive applications. Computing systems include both mobile and fixed/virtual architectures optimized for faster communications, lower power consumption, larger hierarchical memory, novel and robust algorithms, resilience, and HPC networking.

Tactical High Performance Computing

Research novel ways to provide High Performance Computing (HPC) levels of computational capacity to Army's tactical operational realms and provide innovative HPC capabilities to increase the effectiveness of Army Soldiers around the world.  Tactical HPC will be possible only through a combined effort of advanced computing architectures research, mathematical models and techniques to network and manage deployed friendly computing devices.

Tactical computing is intrinsically distributed and heterogeneous and dominated by time sensitive, critical applications.  With the advent of cognitive devices, a computer system can do more than broadcast its presence on the network; it can broadcast its capability, status and intelligently pull/push applications during availability windows.  Cognitive devices allow the battlefield to move toward autonomous distribution of computational workload using push and pull of services, and applications with intelligent power-aware, location-aware, bandwidth-aware, and mobility-aware resources.

Algorithm design and software engineering approaches will be investigated to effectively partition and use binary processing cores to reduce time to solution for Army-relevant problems.  Factors such as performance, portability, and power will be considered in conjunction with developing new models to quantify computing capabilities in hybrid systems to facilitate algorithm signature mapping to available resources.

Additionally to reduce energy consumption for disparate computing technologies, more efficient software design and implementation approaches need to be investigated.  Areas of interest include code refactoring, variations in low-level instruction scheduling, dynamic gating utility, and run-time execution models.  Software run-time model parameters will allow tuning between power utilization, execution time, and software efficiency.

Principal Investigator:

Dr. Barry Secrest, barry.r.secrest.civ@mail.mil, (410) 306-1313

Programmable Networks

This effort is focused on the development of non-traditional network architectures and protocols to solve complex network related problems specific to the war fighter. Basic research is focused on quantum and classical network controls, algorithms, circuits, development of domain specific languages and high performance computing models and simulations. Current research includes:

Development of software defined quantum communication control plane abstractions and algorithms for quantum applications. Programmable HPC network fabrics for tactical smart edge functions. Design and development of autonomous tactical computing and communication platforms.

Developing quantum control protocols for, and exploring applications of, quantum networks, as well as developing efficient methods for modelling controlled open quantum systems. Topics of interest included entanglement distribution and metrology.

Design of communication protocols and prove properties about them using quantum logic. Theorem provers such as COQ and Agda are the main tools of investigation. Other areas of interest include study of open quantum walks within SLH framework.

Principal Investigator:

Dr. Kurt Jacobs, kurt.a.jacobs5.civ@mail.mil, 301-394-2306

Next Generation Architectures for Modeling and Simulation

The Next Generation Architectures for Modeling and Simulation (M&S) research effort refines and demonstrates advances in computer science that support the development of M&S architectures required to support the M&S Communities (Acquisition, Test & Evaluation, Intelligence, Experimentation, Analysis and Training) in the future. Research under this effort is a forward looking approach, working to identify demonstrable concepts for investigation that will be used to accomplish Army M&S 3-5 years in the future. Areas of potential research include methods for M&S execution supporting Software-as-a-Service and Platform-as-a-Service across a distributed environment; use of multi-processor environments to research and refine parallelized and distributed M&S architectures; technologies that support the concept of M&S delivered to the Point of Need; methods and means to leverage the continued growth in computing power and the increasing affordability of those capabilities while achieving efficiencies transcending current hardware, middleware, protocols and data structures; and, hardware and software technology solutions for current and future M&S challenges identified by the M&S Communities.

Principal Investigator:

Christopher McGroarty, christopher.j.mcgroarty.civ@mail.mil, (407) 208-3323