Citation
Edward A. Lee. "Time for High-Confidence Software Systems". Talk or presentation, 1, May, 2011.

Abstract
All widely used software abstractions lack temporal semantics. The notion of correct execution of a program written in every widely-used programming language today does not depend on the temporal behavior of the program. But temporal behavior matters in almost all systems. Even in systems with no particular real-time requirements, timing of programs is relevant to the value delivered by programs, and in the case of concurrent programs, also affects the functionality. In systems with real-time requirements, such as most cyber-physical systems, temporal behavior affects not just the value delivered by a system but also its correctness. The lack of temporal semantics in programs has many consequences that can undermine confidence in systems. For example, without temporal semantics, denial-of-service attacks are no threat. Typically, a denial-of-service attack only affects the timing of services, and if timing is irrelevant to correctness, then such an attacks is also irrelevant. But we know that such attacks are not irrelevant. Worse, without temporal semantics in programs, there is no semantic basis for distinguishing between normal behavior and behavior under attack. Hence, detection of denial-of-service attacks is at best a heuristic. In this talk, we will argue that time can and must become part of the semantics of programs for a large class of applications. To illustrate that this is both practical and useful, we will describe two recent efforts at Berkeley in the design and analysis of timing-centric software systems. On the design side, we will describe PTIDES, a programming model for distributed real-time systems. PTIDES rests on a rigorous semantics of discrete-event systems and reflects the realities in distributed real-time, where measuring the passage of time is imperfect. PTIDES enables deterministic time-sensitive distributed actions. It relies on certain assumptions about networks that are not trivial (time synchronization with bounded error and bounded latency), but which have been shown in some contexts to be achievable and economical. PTIDES is also robust to subsystem failures, and, perhaps most interestingly, provides a semantic basis for detecting such failures at the earliest possible time.

Electronic downloads

• HSCC_Time.pdf · application/pdf · 2580 kbytes

• HTML

  Edward A. Lee. <a
  href="http://chess.eecs.berkeley.edu/pubs/840.html"
  ><i>Time for High-Confidence Software
  Systems</i></a>, Talk or presentation,  1, May,
  2011.

• Plain text

  Edward A. Lee. "Time for High-Confidence Software
  Systems". Talk or presentation,  1, May, 2011.

• BibTeX

  @presentation{Lee11_TimeForHighConfidenceSoftwareSystems,
      author = {Edward A. Lee},
      title = {Time for High-Confidence Software Systems},
      day = {1},
      month = {May},
      year = {2011},
      abstract = {All widely used software abstractions lack
                temporal semantics. The notion of correct
                execution of a program written in every
                widely-used programming language today does not
                depend on the temporal behavior of the program.
                But temporal behavior matters in almost all
                systems. Even in systems with no particular
                real-time requirements, timing of programs is
                relevant to the value delivered by programs, and
                in the case of concurrent programs, also affects
                the functionality. In systems with real-time
                requirements, such as most cyber-physical systems,
                temporal behavior affects not just the value
                delivered by a system but also its correctness.
                The lack of temporal semantics in programs has
                many consequences that can undermine confidence in
                systems. For example, without temporal semantics,
                denial-of-service attacks are no threat.
                Typically, a denial-of-service attack only affects
                the timing of services, and if timing is
                irrelevant to correctness, then such an attacks is
                also irrelevant. But we know that such attacks are
                not irrelevant. Worse, without temporal semantics
                in programs, there is no semantic basis for
                distinguishing between normal behavior and
                behavior under attack. Hence, detection of
                denial-of-service attacks is at best a heuristic.
                In this talk, we will argue that time can and must
                become part of the semantics of programs for a
                large class of applications. To illustrate that
                this is both practical and useful, we will
                describe two recent efforts at Berkeley in the
                design and analysis of timing-centric software
                systems. On the design side, we will describe
                PTIDES, a programming model for distributed
                real-time systems. PTIDES rests on a rigorous
                semantics of discrete-event systems and reflects
                the realities in distributed real-time, where
                measuring the passage of time is imperfect. PTIDES
                enables deterministic time-sensitive distributed
                actions. It relies on certain assumptions about
                networks that are not trivial (time
                synchronization with bounded error and bounded
                latency), but which have been shown in some
                contexts to be achievable and economical. PTIDES
                is also robust to subsystem failures, and, perhaps
                most interestingly, provides a semantic basis for
                detecting such failures at the earliest possible
                time. },
      URL = {http://chess.eecs.berkeley.edu/pubs/840.html}
  }
  

Posted by Mary Stewart on 5 May 2011.
For additional information, see the Publications FAQ or contact webmaster at chess eecs berkeley edu.