The challenge is to identify efforts that preserve, extend, and leverage the Army’s past, present, and future investments in software. The Army views integrated battlefield information systems and intelligent weapon systems as two of its most important sources of combat advantage into the next century. Yet, the software to support such integrated systems represents a challenge to conventional engineering, procurement, sustainment, and technology insertion practices.

Software technology encompasses a wide spectrum of highly technical specialties, activities, and processes, including, but not limited to, the following:

The Army relies on technologically superior systems to counter numerically larger forces, to reduce casualties and damage to urban infrastructure, and to enhance rapid, decisive action. Coupled with sophisticated applications software, high–performance computing (HPC) systems and advanced communication technology enable:

Research in this technology area encompasses computer and software engineering, operational simulation, battlefield environments, and science application tools.

Many Army S&T problems require computational performance rates measured in trillions of floating point operations per second (teraflops). These include problems in chemistry and materials science, computational fluid dynamics, parametric weight/vulnerability reduction, automatic target recognition, high–performance weapon design, and dispersion of hazardous materials. Since no single HPC architecture will effectively handle this spectrum of problems, Army S&T researchers require a variety of computer systems that, in aggregate, support the highest fidelity and greatest speed in analyzing problems of ever increasing size and complexity. These diverse S&T applications also require massive, hierarchical data storage and scientific visualization capabilities to provide meaningful results. HPC utility will fundamentally drive or limit solutions to these critical problems.

The profound impact of modern, computer driven technology has been amply demonstrated in recent hostile operations like Desert Storm and Joint Endeavor. Software is, and will continue to be, a force multiplier.

The Army is faced with a paradox. Systems are being extended in life and expected to achieve land force dominance with diminished resources, in a changing world, with a reduced defense industrial base. Yet, the Army is expected to field lethal, versatile, and rapidly deployable systems in response to the requirement to win decisively and quickly on any battlefield and to do so with minimum casualties.

Computer resources in general and software resources in particular offer a solution to this paradox. The U.S. defense strategy continues to be dominance based on superior technology. But changes in the world’s geopolitics combined with current economic constraints has broadened the focus of attention on technology to include issues of flexibility and adaptability. In today’s weapon system technology, software serves the role of providing these characteristics. Therefore, weapon systems will become more dependent on software to achieve these requirements. According to the Chief of Staff, Army, one of the most important lessons apparent from the Army’s performance in Operation Desert Storm was the profound impact of modern, computer driven technology on the outcome of battle. Desert Storm demonstrated the need to adapt and deploy the technology when and where it is needed.

The Army’s challenge is that existing hardware/software systems are being extended to achieve dominance through increased capability, while resources for that capability continue to shrink. Much of the evolving capability is provided by software. A change in hardware through product improvement has all the appearance of a new item while a change in the software supporting that hardware is not viewed as a new item. This visibility mismatch furthers the gap between the perceived and actual costs of hardware and software sustainment. The goal of the Army software S&T effort is to reduce software development and sustainment cost and schedules by an order of magnitude in the next 10 years, while increasing the capabilities of the software industrial base to allow more to be done with less.

Software allows for short lead times and can be deployed over satellite communications links with essentially no logistics volume, weight, or fuel cost. State–of–the–art training technology can provide expert systems that can train soldiers to use the new software on the battlefield. Changes to deployed systems can feasibly be made in theater through software modifications that have been previously tested in the Army’s stateside life–cycle software engineering centers (LCSECs) where synthetic environments, interacting with real materiel, are used to demonstrate successful performance of the changed system.

With technology progressing at a rapid pace, the dilemma is that systems that are state of the art today become enormous cost burdens in the near future. Some systems deployed today and still in production require dated software maintenance and change techniques that are frozen in time and appear to be enormously expensive to sustain (e.g., interoperate, respond to threats). Yet, the cost to make these changes in hardware, produce new hardware, refurbish materiel, and redeploy would be even more unacceptable.

The Army recognizes that research and development (R&D) in software engineering and life–cycle management and environments are to a large extent commercially driven. Systems currently under development and the employment of advanced concepts and operational scenarios that have a greater reliance on synthetic environments will exacerbate the current dilemma faced in supporting deployed software. A paradigm shift is required in the way that software is viewed, supported, and developed. Decreased budgets will increase reliance on commercial products, and possibly increase costs. It is imperative that we learn to leverage commercial advancements, while continuing to provide some level of support to maintain an industrial base in the software development market.

The Army software technology investment strategy represents the distillation of extensive work performed by technical experts from industry, academia, and government to create such a scenario. The work plan is focused on the needs of the Army, windows of opportunity, and a realizable implementation, given limited resources.

This subarea is concerned with development, exploitation, and deployment of high–performance computers offering scalable performance for a broad range of Army and DoD applications. Scalable parallel systems technology includes parallel architectures, compilers, and programming methodologies and tools essential to facilitate their effective use, systems software, mass storage, input/output (I/O), and visualization technologies. Application requirements drive the design of these systems.

Early access to new systems by DoD and Army users accelerates development of specific applications as well as knowledge, algorithms, and programming tools for solving problems. Current performance levels of 100 billion of floating point operations per second (gigaflops) will sustain a 10–fold increase by FY98 to reach the goal of 1 teraflop.

The Army relies on the DoD HPC modernization program to provide computing capabilities essential for the conduct of RDA and in support of the operational forces. The Army manages and operates two DoD HPC major shared resource centers (MSRCs) and five distributed centers (DCs) within the DoD modernization program. The Army MSRCs are located at the ARL Aberdeen Proving Ground (APG) and the Army Corps of Engineers Waterways Experiment Station (WES), which combine to offer full service HPC capability and high speed network access to both the DoD S&T and test and evaluation communities and the national HPC infrastructure.

The capabilities provided at the Army MSRCs are directly aligned to the DoD following objectives:

Deployment of state–of–the–art HPCs and exploitation of evolving computational algorithms provide an environment that allows the Army to solve critical mission problems and to tackle problems that were previously intractable. Improved HPC capability shortens design cycles and design costs by reducing the reliance on handcrafted prototypes and destructive testing. Robust high–speed network connectivity is essential for desktop access to remote resources and daily, interactive collaboration with remote users.

Issues include:

The high–performance specialized systems subarea includes the development of innovative technologies such as optical processing, embedded systems, neural networks, and systolic processing, that meet military requirements but have limited commercial potential. Target goals for these systems include a 200–fold increase in data processing reliability, a 10–fold system weight reduction, and a 5–time increase in digital data processing speed. The Army relies on DARPA and the other services to provide technology for its systems applications.

The diverse deployment criteria for specialized Army systems makes hardening and repackaging essential. In addition, image and speech recognition dictates that DoD and the services examine optical processing and neural computing. Incorporating fuzzy logic into neural computing for Army problems requires further research into expressing expert knowledge and combinatorial complexity in simple linguistic rules while reducing demands on computing resources.

Real–time access to information and data is required to realize one of the Army’s key modernization strategies of "winning the information war."

Integral to this capability are the computing and networking capabilities required to provide a secure and seamless battlefield computing environment. These capabilities include instant access to data, data extraction of the desired information in near–real time, and retrieval and presentation of the information in a form that the soldier can readily use to make educated decisions and better control the available resources. These capabilities require integrated networking of battlefield and research–based computing systems. High–speed and high–capacity networks enable interaction with research–based computing assets.

Networking has long been a mechanism to foster scientific collaboration, and the services were launched into this realm by the ARPANET initiative of the 1970s. This DARPA program has grown to be integrally responsible for the Internet explosion that serves as the catalyst and foundation for the National Information Infrastructure project. Ten gigabit (GB) per second to 100–GB per second networking will be available by the year 2000.

As part of the DoD HPC modernization program, the Defense Research and Engineering Network (DREN) is being designed to maintain intersite communication performance levels commensurate with I/O bandwidths of the HPC systems to which DREN will provide access (Figure IV–6). Bandwidth requirements are projected to approach 622 megabits per second (Mbps) within 2 to 3 years, and over 1 gigabits per second (Gbps) within 5 years to support and enable distributed computing performance in the TFLOPS range. These requirements represent an order of magnitude (x10) increase over currently available bandwidth within 1 year and more than two orders of magnitude (x100) increase over current bandwidths within 5 years.

Figure IV-6. IDREN Configuration
Click on the image to view enlarged version

The Army provides the technical lead in maintaining the interim DREN (IDREN) connectivity through transition to the DREN component of the DoD HPC modernization program. Current Army mission projects in networking include, but are not limited to:

The challenges include recognizing and identifying the most promising commercially available technologies and products and adapting these to Army needs. Since the environment and the conditions used in the commercial and military sectors are not the same, some adaptation may be required, especially in four areas: sensing, analysis, distribution, and assimilation. These factors turn combat information into knowledge, described by mathematical algorithms, and distribute the information in a hostile battlefield environment. The objective is to provide real–time, knowledge–based operations and seamless battlefield communications and computer processed C³I electronic warfare (EW) throughout the operational hierarchy.

Technical issues being addressed include protocols for reliable, seamless connectivity as remote hosts increase in number and explore high–bandwidth data channels to offset the need for large–scale localized data storage. Security and data integrity issues are also of interest as well as the configuration optimization, mobility and robustness of the computing systems.

Wearable computers and their applications are starting to become feasible. They can act as intelligent assistants and may take many forms, from small wrist devices to head–mounted displays. They have the potential to provide anywhere, anytime information and communications. Applications such as telemedicine (augmented reality), memory aids, maintenance assistance, distributed mobile computers in wireless networks (individual communication with soldiers on the battlefield), and desktop applications such as word processing, scheduling, and database applications.

The Army software technology investment strategy (ASTIS) is a targeted strategy based on a principle that capitalizes on conditions of imperfect competition with our adversaries and rapid technological change. Stated in warfighter terms, hit them where we are strong and they are weak, with the technology transfer equivalent of overwhelming force. The ASTIS vision includes:

This vision is realized through the establishment of a virtual advanced software technology consortium (VASTC). Assets of a VASTC will be a distributed matrix of an integrated government, academic, and defense industrial software and computer resource asset base.

The word "virtual" in VASTC implies:

A roadmap establishing, prototyping, demonstrating, and scaling up incremental capabilities hinging on this principle will yield an emphasis and a paradigm shift. Each effort in the roadmap has building blocks of integration, process, product teams, and a paradigm shift built in. The result will create a distinct techno–economic paradigm built around flexibility rather than simple volume production.

The ASTIS strategy consists of:

The ASTIS guides the industrial base toward key critical technology sectors. These sectors include computers and software support for the development of capital goods such as aircraft, ground transportation vehicles and systems, flexible manufacturing facilities, as well as telecommunication, decision support, visualization, and battlefield information systems. These are the sectors having the greatest growth and technological potential.

The VASTC offers industry and academia distributed yet integrated advanced technology transfer incubation facilities in which the emerging technologies come together to enable risk reducing proof–of–principle demonstrations conducted with access to materiel in an operational environment. This enivronment enables evolving synthetic environments, a distributed high–performance computing infrastructure, and advanced large–scale program management techniques. The VASTC establishes a rapid software technology transition channel for the Army and the nation.

Figure IV–7 depicts a single software technology incubation cell. The VASTC incubators scale up immature, emerging, and mature technologies, and integrate these technologies into existing environments. Real systems are the test articles and have the beneficial side effect of reducing risk on the actual programs. Deployed (in–service engineering), new developments, and advanced concept systems provide scale–up opportunities and real–world challenge problems. Yet, the artifacts from the incubators are reusable components that are targeted to domain–specific software architectures.

Figure IV-7. Software Technology Incubator Concept
Click on the image to view enlarged version

The VASTC offers the government an engine to continuously reduce risk and insert technology into existing weapon system software. The VASTC is also a software technology training factory. People are educated and trained in the use of the new technologies while they are analyzing and modernizing existing systems. The software training factory operates on existing systems with new technologies. The VASTC training factory will optimize resources and reduce risk by acting as a booster to future builds of existing systems.

Regardless of a VASTC participant’s role (e.g., academic, principal investigator, independent R&D (IR&D) explorer, governmental staff developer), the technology will flow with the participants. The VASTC will be a national asset and an engine of technology transfer influencing commercial practice that will be reflected in government products.

Key to realizing the vision of the VASTC will be the capability to provide integrated automation capabilities throughout the software life cycle. Process automation is a relatively new area of research with many technical challenges. A common underlying infrastructure that allows ease of integration and supports evolutionary development for each individual technology being automated will be necessary. Early efforts will be directed at developing this underlying infrastructure and providing an open interface that encourages tool vendors to build tools that support VASTC. Long–term efforts will be directed at finding technological advances that will make a seamless automated software development paradigm a reality.

The amount of Army software (old, modified, new) requiring life–cycle software engineering services is increasing exponentially along with life–cycle costs. To address this issue and bring costs under control, the Army has initiated a conceptual shift in how future life–cycle engineering services will be accomplished. At the core of this initiative is the next–generation life–cycle software engineering center (NGLCSEC) prototype. The goal of this new center is to reduce weapon system software development and support costs by at least an order of magnitude. The goal will be achieved by creating a seamless software engineering directorate within the Army Materiel Command (AMC) that shares resources, knowledge, and best practices among its members, with a focus on the customer. The concept is being prototyped at the Tank–Automotive and Armaments Command (TACOM) and scaled to an AMC–wide infrastructure capable of supporting Force XXI and the Army After Next.

Networking systems that can support greatly increased throughput, a supportable infrastructure, and mature domain–specific architectures must be sought out to fully achieve interoperability between geographically dispersed member organizations. Also, new management processes will be needed that can adapt to the many systems supported by member organizations and their organizational cultures.

All software systems are requirements driven. Software users have specific and general needs that must be fulfilled by the software they procure. In order for these software systems to satisfy those needs, the systems must satisfy the formal requirements outlined by users and engineered by designers. Automated systems that can analyze a software system’s formal design to validate the requirements are needed.

Embedded software packages, like software for aircraft control, are critical in the sense that if they fail, soldiers die. Battlefield information systems are critical because they provide critical information to the commander on the scene that facilitates sound decision making.

Some software requirements are difficult to specify. Methods for formal specification of these requirements are needed to enable automated validation.

Computer–aided prototyping is an evolutionary software development paradigm that involves the end user of the software in the requirements development process. This paradigm makes use of prototype demonstrations and user feedback to iteratively develop a functional prototype. Prototypes are executable specifications of software systems partially generated and partially built from atomic components retrieved from a reuse repository. Current efforts are directed at maturing and commercializing this technology to enable practical use by the life–cycle software engineering centers in the research, development, and engineering centers (RDECs). Our goal in FY98 is to continue the maturation of this technology and support its commercialization and incorporation into the NGLCSEC.

Computer–aided software engineering tools are difficult to commercialize. The long–term investment required to keep these tools viable in the software market is tremendous. Tools like computer–aided prototyping tools are important for the realization of the ASTIS vision, but are not attractive for the software industrial base. Efforts need to be concentrated on supporting their commercialization and influencing the industrial base to champion this technology.

Future system development will require vast amounts of data to be collected and made available throughout a system’s life cycle. A system evolution record (SER) is needed to serve as a cradle to grave repository for all artifacts and decisions made during the evolution of a software system. An initial model of a SER is being implemented. Our goal for the next and subsequent years is to model different pieces of the software development process to integrate with the SER.

New techniques for capturing design decisions must be developed to allow for the linking of these design decisions into the SER. Hypergraphs (nonlinear representations of information) must also be developed that will store not only the artifacts to be contained in the SER and the decisions already mentioned, but also dependencies between them. Additionally, new technologies for sharing information like the World Wide Web must be exploited to enable sharing of critical life–cycle information over extended distances.

Exploiting emerging high–performance computing, storage and retrieval, and communications systems for the Army’s electronic battlefield (EBF) requires advanced software capabilities incorporating AI. After 2000, DIS software capabilities are expected to include cooperating intelligent systems, coupling of symbolic and neural processing, and autonomous synthetic agents and robots. This will provide a large synthetic computing environment in which networking and process management are handled automatically and are transparent to the users. This includes multi–level secure data routing, loci of computation, workload partitioning, and interconnection of government and industry/academia expert and information centers with built–in ownership protection. By 2010, planning systems capable of complete support of military operations and deployment with less than 24 hours notice will become available.

The Army federated laboratory is focusing basic research in five areas, each of which will need AI technologies. These areas are advanced sensors, advanced and interactive displays, software and intelligent systems, telecommunications and data distribution, and distributed interactive simulations. Three approved consortia will work on Army–specific basic research over the next 5 to 8 years. The Army Artificial Intelligence Center manages the Army Artificial Intelligence Program, which is focused on applied research and prototyping to deliver artificial intelligence solutions in support of Force XXI and AAN. A number of expert systems have been delivered, and emerging technologies such as fuzzy logic, neural networks, and generic algorithms are being used to build advanced technologies.

The study of AI has produced advanced technologies in three categories: mature, emerging, and immature. Expert and rule–based systems are examples of mature technologies that are being widely used in commercial applications. The major challenge is to develop prototypes for Force XXI and identify appropriate technology insertion in existing systems and systems under development. Fuzzy logic, genetic algorithms, and neural networks are examples of emerging technologies. The development of prototypes for exploratory development and risk mitigation will clarify the technical issues. Finally, intelligent agents and machine learning are examples of immature technologies. These are the focus of the basic research efforts in the Army federated laboratory.

Human–computer interactions deal with the systematic application of scientific knowledge about humans to design the simulated human and its behavior as well as the interface software through which real humans interact with the synthetic environment. The Army programs addressing the physical human–machine interface and the human engineering aspects are described in Section III–N, "Human Systems Interface." Information display and human computer communications technologies are steadily advancing. COTS user interface management tools, standards–based approaches for product development, style guides, and graphical information visualization are now available for commercial and military applications. The Army programs addressing human computer interactions rely on these general tools to make computers and associated networks easier to use as well as to build. This is a continuous process.

An important aspect is the adaptation and interface of the large number of previously developed application–specific closed architecture codes with the COTS human–computer interaction tools. Connected speech systems with increasing natural language interpretation and voice recognition that can be trained quickly for different voices are appearing, but they lack robustness for military applications. Group system capabilities are needed to provide for multi–user interfaces in to software systems, and for group decision making capabilities in battlefield planning systems.

Safeguarding of information, loss–of–service protection, and damage prevention to programs and data through errors or malicious actions requires multilevel security, defense against malicious software, and credible procedures for technical evaluation, certification, and accreditation of software. The Army relies on the National Security Agency (NSA) to provide the required assured computing technologies.

Also relevant to this category is the short–term year 2000 problem. Essential management information systems must continue operation through January 1, 2000.

The biggest challenge facing the assured computing field is the year 2000 problem. Time has nearly run out for developing automated tools to find a solution to this problem, or to develop new systems to replace all legacy systems that display the problem. Manual editing methods will be necessary to solve the problem, and that means manpower. Effective means of keeping critically short software professionals in the Army to solve this problem must be developed.

Instant access to information on computer systems throughout the world is now a reality. Surfing the Web has become a national pastime for Internet users in and out of the government. The Web provides the capability for anyone with access to the Internet to access information on every imaginable subject at any time of the day or night, and on any machine that contains a Web server. This technology is being exploited in many ways to increase information sharing between agencies and to further our movement toward a paperless Army. Web servers have been established at virtually every organization that provides information or services to the Army. Publications and forms have been made available electronically and policies should encourage the use of electronic forms and publications.

This is a relatively new area of investigation and definitive near–, mid–, and far–term goals are still in the early stages of formulation. The tremendous rate of growth in Web technologies offers the promise of many significant advances within a very short time. Army planning will, in part, be driven by the rapid changes in available marketplace technologies.

The most critical challenge in this area is the ability to provide secure access to sensitive information, allowing easy access to authorized users while preventing unauthorized access. This technology is moving faster than even industry can keep up with. Most of the development of Web applications is being done by hackers working nights and weekends with no wish for compensation. Capabilities for increased information availability and increased interactivity have resulted in our inability to control what information flows and where. Future research must design ways to protect critical information while providing access to necessary information and capability.

The roadmap of technology objectives for Computing and Software is shown in Table IV–16. The Army software program is structured to take advantage of emerging commercial software technologies and relies on the DoD software program for most of the generic software technology, including tools and techniques for software engineering, reuse, and life–cycle management. This program is integrated with the tri–service Reliance program and addresses only those technology areas where DoD program investment will not satisfy Army–specific application needs.

The influence of this technology area on TRADOC FOCs is summarized in Table IV–17.