A national technology resource, MITRE is a not-for-profit corporation working in the public interest. It addresses issues of critical national importance, combining systems engineering and information technology to develop innovative solutions that make a difference.

MITRE specializes in system engineering of large, complex information systems, and is an internationally recognized center of excellence for military command and control, intelligence, and air traffic control. MITRE also works in other fields, including transportation, medicine, law enforcement, space exploration, and environmental remediation.

1999 was the 40th anniversary of MITRE's founding as an MIT Lincoln Lab spin-off, to provide the system engineering and ongoing support for the first major real-time command and control system - the massive, multi-billion dollar, continental air defense system called SAGE (more).

MITRE owns three Federally Funded Research and Development Centers (FFRDC's:¹ Command, Control, Communication and Intelligence; Aviation System Development; and Enterprise Modernization. MITRE's Defense and Intelligence FFRDC includes the Center for Air Force Command and Control Systems (CAFC2S), the Center for Integrated Intelligence Systems (CIIS), and the Washington C3 Center (WC3C). (It is slightly confusing that the Defense FFRDC components are also called centers.) I am part of the CAFC2S, which is the largest of MITRE's centers in any of its FFRDC's, and is located at the headquarters site in Bedford, MA. The Air Force Center has over 1100 staff members, and is matrix structured, being composed of eight "horizontal" technology Directorates and four "vertical" program (application domain) Directorates. I am the Chief Scientist in the Information Technologies Directorate, which has about 220 staff members. This Directorate has eight departments, including a Real-Time department with about 25 members. Unsurprisingly, I have an especially close relationship with the Real-Time department (and the real-time groups in other Centers within all three FFRDC's).

Below is the current 1-page summary of the real-time computing activities at MITRE that I am responsible for.

Real-Time Computing at MITRE

Conventional real-time computing concepts and techniques are focused on relatively small, simple, static, device level, sampled data monitoring and control. These concepts and techniques consider timeliness to mean that the latencies of non-schedulable activities, such as interrupt response times and operating system service calls, have known tight upper bounds. In addition, "real-time" is often equated to "real fast," in that lower magnitudes of these latencies are thought to imply "harder" real-time. To the limited extent that the conventional real-time computing perspective considers schedulable activities, such as tasks or threads having completion deadlines, it is confined to "hard" real-time in the sense that all deadlines must always be met. This conventional style of real-time computing, although limited in applicability, is a core competency in MITRE.

However, conventional real-time concepts and techniques do not scale up to larger, more complex, more dynamic, more distributed, higher level real-time systems – such as are common in the defense application domain (e.g., surveillance and combat platform management, mission management, battle management, C²) – especially for network-centric warfare. There, timeliness is about achieving application- and situation-specific acceptable predictability that a collection of activities (e.g., tasks for a computer or a radar or a weapon), which are competing for shared resources, will satisfy their time constraints according to an application-specific scheduling optimization criterion (e.g., minimize tardiness proportionally to relative task importances). In the case of distributed – meaning multi-node – applications and programs, predictable timeliness must be end-to-end. These application environments and computing systems are inherently dynamic and non-deterministic, which necessitates adaptive resource management.

MITRE's principal real-time computing value proposition to its clients (and, for our unclassified results, to the real-time research and product development communities) is based on the development and transition of innovative real-time computing concepts and techniques that are able to cost-effectively scale up to every level in an enterprise. For example, we have formulated a general timeliness framework that expresses the timeliness of each activity in terms of the utility provided to the application as a function of when the activity completes. To maximize the collective utility of these activities, we continue to devise application-specific, adaptive, scheduling optimization criteria and disciplines. An aspect that distinguishes our work from that of the traditional real-time computing community is that we emphasize the management and scheduling of application resources (e.g., sensors and weapons) as much as conventional computing hardware and software resources. For distributed real-time activities, we are continuing to develop a trans-node computational abstraction called "distributed threads" that propagates the context needed to perform distributed resource management for achieving end-to-end timeliness – both within an element (computer, radar, etc.), and among elements in a platform (aircraft, etc.) or combat theater (network-centric warfare). This abstraction was originally created by Jensen and his Ph.D. students at Carnegie Mellon University.

We compliment our own research and advanced technology development by actively engaging in important real-time standardization projects – notably real-time CORBA and real-time Java. For example, our distributed thread abstraction forms the basis of the programming model in OMG's Real-Time CORBA 2 and in Sun's Distributed Real-Time Specification for Java. We also engage in collaborations with real-time COTS operating system, middleware, and tool product developers.

MITRE strongly focuses on transitioning real-time and distributed real-time technologies, and applying the real-time computing expertise of its staff, by working intimately with DoD agencies, contractors, and (mostly classified) programs.

Regarding Visits to MITRE

Because so much of what we do at MITRE is classified, we regret that visitors must either be U.S. citizens or have a Green Card (unfortunately this excludes international students on F1 visas).

1. Many people are unfamiliar with the notion of FFRDC's, although the names are familiar. Here is a list of the DoD-sponsored FFRDC's (other federal government agencies also sponsor FFRDC's, including the ones listed above that MITRE owns):

Add to Favorites

Print Page