Embedded and Secure Networked Embedded Systems

Document Sample
Embedded and Secure Networked Embedded Systems Powered By Docstoc
					              Embedded and Secure Networked Embedded Systems:
                     The Invisible Cyberinfrastructure
                                      Janos Sztipanovits
                                    Vanderbilt University
                             Janos.Sztipanovits@Vanderbilt.edu
                                        June 12, 2003


    Ongoing discussions on Cyberinfrastructure essentially miss the most pervasive area
of computing, embedded and networked embedded systems. Pervasiveness of embedded
systems in science and industry is rapidly creating an invisible cyberinfrastructure, which
has tremendous impact on scientific progress, on industrial competitiveness and on the
security of our critical infrastructure. In light of the rapidly emerging role of embedded
systems in science and engineering and that of its fundamental role in critical
infrastructure and industrial competitiveness, I submit the following recommendations
and supporting arguments.

Recommendations

   1. The scope of Cyberinfrastructure should include research infrastructure for
      embedded and secure networked embedded systems.
   2. This area offers excellent opportunity for cooperation between CISE and
      Engineering based on the facts that CISE has already started significant research
      efforts on establishing new foundations for embedded and networked embedded
      systems, and Engineering is one of the major stakeholders in the progress.
   3. Based on the overall significance of this emerging field, I recommend to organize
      a joint CISE/ENG Workshop focusing on Cyberinfrastructure for Embedded and
      Secure Networked Embedded Systems.


Supporting Arguments
    Embedded computing is an increasingly dominant part of future physical and
engineered systems. Systems are formed by interacting components. The new trend is
that a rapidly growing number of components and interactions in real-life systems are
computational. From electric shavers to airplanes and from cars to factory robots,
computers monitor and control our physical environment. Distributed control and process
automation systems integrate manufacturing production lines. Flight control and avionics
systems keep airplanes flying. This trend is based on a fundamental technical reason:
computing is uniquely suitable for implementing and controlling complex interactions
among physical system components.
    The steady diffusion of embedded software throughout our critical system
infrastructures motivates a renewed national-level attention to, and investment in,
research and engineering of embedded software. The Hybrid and Embedded Systems
Core Program at CISE, the large number of embedded and network embedded systems
related proposal submissions and projects in the ITR program, and the embedded systems
program suite at DARPA well illustrate this increasing national attention. Some of the
main drivers of this interest are the following:
    1. Embedded computing is becoming the universal system integrator for physical
       systems. The pervasiveness of this technology is well illustrated by the following
       facts: (a) the total shipment of microprocessor units (MPU) and micro control
       units (MCU) in 2000 was over 8.2 billion units, of this about 98% related to
       embedded applications1 and (b) between 1994 and 2004 the need for embedded
       software developers is expected to increase 10-fold.2
    2. Embedded and networked embedded systems are becoming important enablers of
       progress in science. (See e.g. the NEES and CLEANER presentations.) Cutting
       edge research in physics, environmental science, astronomy, structural
       engineering, biology, zoology, geo-sciences all use increasingly sophisticated
       networked embedded systems as a new way to instrument the physical and
       biological environment, to collect and process information in close interaction
       with the physical world. Closed-loop systems including thousands or hundreds of
       thousands of networked sensors and actuators integrated with distributing
       monitoring and control algorithms are the foundation for achieving totally new
       behaviors in smart structures, smart materials, adaptive optics, gossamer space
       antennas and many others.
    3. Embedded computing presents a vitally important opportunity and challenge for
       DoD. From avionics systems to smart weapons, embedded information processing
       is the primary source for superiority in weapon systems. The new wave of
       inexpensive MEMS-based sensors and actuators and the continued progress in
       computing and communication technology will further accelerate this trend.
       Weapon systems will become increasingly “information rich”, where embedded
       monitoring, control and diagnostic functions penetrate deeper and with smaller
       granularity in physical component structures.
    4. Embedded computing represents a disruptive technology for established
       industries. For example, 80-90% of all innovations in the automotive industry are
       now based on embedded computing.3 In particular, the expected impact of drive-
       by-wire technologies is revolutionary.4


   Embedded systems and software have been neglected as a topic in software,
networking, and security research until recently. Viewed as an ad hoc issue and

1
  Sastry, S., Sztipanovits, J., Bajcsy, R., Gill, H [Eds.]: Model-Based Design of Embedded Systems, Special
Issue of Proceedings of the IEEE, Vol. 91, No.1., January, 2003.
2
  R. H. Bourgonjon, “Embedded Systems in Consumer Products,” in Lecture Notes on Embedded Systems,
LNCS Vol. 1494, 1996, pp. 395-403.
3
  Gunter Heiner, DaimlerChrysler, Proc. of the Joint Workshop on Advanced Real-Time Systems, Vienna,
26 March 2001.
4
  Tom Fuhrman, GM, The Role of Embedded Software in the Automotive Industry, ARIES Workshop,
CMU, April 19, 2002
subordinated to design methodologies for engineered systems, embedded software
practice has largely focused on developing small, one-of-a-kind controllers that execute
in dedicated computing environments. Proprietary, monolithic, closed real-time
operating systems and incomplete tool chains have traditionally dominated embedded
systems infrastructure. Research opportunities have been damped in this environment,
while the challenges for modern complex engineered systems have increased rapidly in
military and industrial applications. The implications of this disarray are staggering:
      A university and industrial research infrastructure that is incapable of supporting
       next-generation IT-enabled physical systems, particularly large-scale mission-
       critical systems
      A lack of comprehensive advanced standards for embedded software, resulting in
       competitive disadvantage for the US in the software control technology and
       engineered product markets and
      Exceedingly fragile and vulnerable critical infrastructures, such as air traffic
       management, supervisory control and data acquisition systems for power grids,
       and national telecommunications systems.

The bold national initiative on cyberinfrastructure needs to place increased emphasis in
this area and address these problems without delay.