Showing papers presented at "NATO ASI RTC in 1994"

Is Time a Real Time? An Overview of Time Ontology in Informatics

Abstract: Aim of this paper is to introduce several aspects of time in the heterogeneous world of Informatics and define ontologies for time in different domains of computers and their applications Some philosophical and physical backgrounds are given, to show how, from the richness of often contrasting ideas developed in the framework of these disciplines, many useful concepts have been derived also for computer science Architectural aspects of computer systems, Information Systems applications, and Real-time systems, are considered as temporally problematic domains

Asynchronous Decentralized Realtime Computer Systems

Abstract: Realtime computer systems which perform physically and logically decentralized mission management such as collaborative direction within a team of autonomous entities conducting manufacturing, maintenance, or combat must accommodate significant run-time uncertainties in the application environment and system resource state, by being dynamically adaptive. In particular, such systems have mission-critical time constraints which must be satisfied acceptably as specified by the application given the current circumstances.

Engineering Predictable Real-Time Systems

Abstract: The demands placed upon real-time systems (or portions thereof) in respect to predictability continue to increase as these systems become integrated into a wide variety of safety critical applications. It is essential to be able to guarantee that all critical processing is accomplished accurately and on time. In this contribution, a point of view is established in respect to achieving predictability in combination with another vital ability, namely understandability. The argument is put forward that these two goals can be jointly attained; further, that understandability is the most important common denominator for achieving a variety of other important real-time computing abilities; for example, reliability, testability, verifiability, maintainability, and so on.

Formal Methods for Real-Time Systems

Abstract: We believe that formal methods are both under-used and over-sold and consider here when and to what extent it is appropriate to use formal methods in the development of safety critical systems.

Combat System Prerequisites on Supercomputer Performance Analysis

Abstract: Modern naval combatants host many highly complex systems. Each system performs one or more tactical capabilities. The single large-scale system formed via integration of these complex systems is a Combat System. This system is required to exhibit behavior described by the following attributes: deterministic, responsive, fault tolerant, automated and controlled.

Design of Real-Time Fault-Tolerant Computing Stations

Abstract: The steady increase observed during the past decade in distributed computer system (DCS) use in safety-critical real-time applications is expected to continue through the 1990’s. For example, DCS’s have been increasingly adopted in applications such as space navigation, air-traffic control, hospital automation, national defense, etc. [11, 16, 21, 43, 47]. To attain the desired level of reliability, such DCS’s must be designed to possess effective fault tolerance capabilities.

Communications for Real-Time Industrial Control: The Design Issues

Abstract: In most real-time automation applications we are required to control real-world, physical processes — processes which have their own intrinsic time-related properties. A chemical reaction, once started, proceeds according to its own dynamics, and the chemistry takes little notice of the controlling computer. Likewise, once a piece of metal is removed by a milling-machine, it cannot be replaced! These processes operate according to their real-time dynamics, and as a result, any controlling system must recognise these inherent real-time characteristics. A real-time computer control system is thus one in which the correctness of the system depends not only on the logical results of computations, but also on the time at which the results are produced. “Real-time” does not necessarily imply “fast”, but does mean “fast enough for the chosen application”.

Safety Engineering and Assurance for Real-Time Systems

Abstract: Many real-time systems are also safety-critical, that is they are used in applications where their inappropriate behaviour, or failure, could lead to loss of life, or severe environmental damage. Examples include the flying control systems for aircraftl, reactor protection systems, and anti-lock braking systems for cars. The primary difference between the development of non-critical real-time systems and safety-critical real-time systems is that we are concerned with failures, both of the computer system and the wider system in which it is embedded. Our discussion will focus on the issue of identifying failure modes and showing that safety requirements are met, despite the fact that failures can occur.

The Reality of a Real-Time UNIX Operating System

Abstract: Industry standards will be the basis for the next generation of real-time systems. These industry standard systems, termed Open Systems, provide a number of advantages for users. Open systems reduce system cost and time to market, increase availability of software packages, increase ease-of-use, facilitate system integration, and can provide complete software portability.

An Overview of the HOOD Software Design Method

Abstract: HOOD or Hierarchical Object Oriented Design has been developed as the standard software design method of the European Space Agency (ESA) for Ada projects. It is used in all major space programs and in the European Fighter Aircraft program. The concepts and the method in HOOD are described in detail. Aspects of the application of HOOD are also discussed.

A Real-Time Image Processing Language?

Abstract: Real-time image processing is widely used in multi-media systems, virtual reality and simulation, and remote command and control. Unfortunately, there is no standard language to support image processing in a real-time framework. Moreover, it is unwise to assume that other real-time programming languages can support the specialized needs of image processing.

A Consensus-Based Framework for Responsive Computer System Design

Abstract: The concept of responsive computer systems is presented. The emerging discipline of responsive systems demands fault-tolerant and real-time performance in both parallel and distributed computing environments. The responsiveness measure is discussed and a new design framework for responsive systems is introduced. The new framework is based on the fundamental concept of consensus and on application specific responsiveness. It is shown that consensus is crucial in responsive synchronization, communication, diagnosis, and reconfiguration.

Real-Time Operating Systems: Can Theoretical Solutions Match with Practical Needs

Abstract: Constructing real-time systems, which are predictable in a very restrictive sense, is a challenging task for scientists. On the other hand today’s real-time operating systems do not meet these strong requirements. This paper gives an insight into sources of delays for application processes caused by the operating system. It follows from this analysis that most of the services of real-time operating systems insert unexpected delays to the application processes and worst case values are hardly to determine. Regarding the fact that real-time operating systems are used successfully certain deviations from the model of strict predictability seem to be tolerable for most applications. The paper presents some ideas to describe such tolerable deviations more precisely.

Formal Specification and Timing Analysis of High-Integrity Real-Time Systems

Abstract: We motivate the need for formal specification and verification of specifications in the case of high-integrity real-time systems After a review of sample approaches, we concentrate the attention on Petri nets augmented to support timing requirements We illustrate a very general formalism and an associated timing analysis procedure We then illustrate a restriction of the general model and a specialized analysis procedure Concepts are mainly illustrated via examples and informal descriptions The reader is directed to the published literature for formal details

Algorithmic Methods for Real-Time Scheduling

Abstract: Task scheduling is a wide research area whose results gained increasing interest during the last decades. A great variety of different algorithmic methods has been developed. Though these methods are very often adjusted to the specifics of the various scheduling problems, there are few principles along which scheduling algorithms work. The purpose of this contribution is to give an overview on the main algorithmic approaches applied in real-time scheduling. Different methods for solving selected problem classes are discussed.

Synchronization Techniques, Illustrated by the Concepts of the Dependable Distributed Operating System DEDOS

Abstract: This article gives an overview of the most important synchronization issues in real-time systems in general and in distributed systems in particular. Special attention is paid to static (deterministic) methods and the role of distributed algorithms. Many of the concepts discussed are illustrated with concepts of the Dependable Distributed Operation System (DEDOS), which is presently under development at the Department of Computing Science of the EUT. It is shown that synchronization is one of the most important issues in concurrent and distributed real-time systems.

Scheduling in Fieldbus Based Real-Time Systems

Abstract: In this paper we discuss the characteristics that hard real-time networks (in particular fieldbuses) should satisfy to meet real-time requirements. Fieldbuses are networks that interconnect sensors, actuators and machines at the lowest level of a Computer Integrated Manufacturing (CIM) architecture. We propose that they should provide the services of a distributed real-time database. We found scheduling as the essential support to ensure the maintenance of this distributed real-time database in a consistent state, and we propose the analysis of pre-runtime schedulability of both tasks and network traffic to ensure predictability.

Off-Line Scheduling of Hard Real-Time Distributed Systems Using Windows

Abstract: The last decade dependability of computer systems has become a more and more important topic in computing science. Dependability of a computer system includes besides correct functionality also timeliness, i.e., hard real-time constraints imposed on the tasks of the system must be satisfied. In our opinion, timeliness can only be guaranteed when the event rate of the environment is bound and all hard real-time tasks are scheduled off-line incorporating all resources. An approach for off-line scheduling of tasks in hard real-time distributed systems is presented.

Issues in Object-Oriented Real-Time Language Design

Abstract: To express real-time applications, most real-time languages introduce specific constructs to specify features like deadlines, periodic behavior and time intervals. These constructs, in general, can be seen as annotations to conventional language structures. Object-oriented programming languages have gained popularity in non real-time applications. These languages are highly modular and provide protection through strictly encapsulated abstract data types called objects. In addition, classes and inheritance mechanisms enable programmers to reuse existing software. Recently, there have been some attempts to define real-time object-oriented languages. One motivation for introducing these languages is to reduce the complexity of applications through modularization so that predictability and reliability of applications can be increased. Secondly, inheritance mechanisms allow reuse of software modules that have well-defined behavior. This may simplify analysis for a particular real-time application. Thirdly, since object-oriented languages are now more frequently applied to software implementations, it would not be practical to adopt different languages for real-time and general parts of an application. However, there are several issues to be addressed in order to fully utilize object-orientation in real-time applications. Firstly, real-time specifications must be reused separately from the ‘application code’. This allows the reuse of classes in applications with different real-time behavior. Otherwise, changes made to the application requirements or real-time specifications in the sub-classes may result in excessive redefinitions of super-classes although this seems to be intuitively unnecessary. This we refer to as the real-time specification anomaly. Secondly, since a subclass may extend, exclude or replace the real-time specifications of its super-classes, semantics of inheritance must be clearly defined. Thirdly, there must be language mechanisms to modularly specify and reuse alternative implementations. For example, inter-object interactions often result in blocking execution threads. Blocking time can be minimized using dedicated strategies. This requires abstraction of inter-object communications and large scale synchronization among objects as first class objects. In addition, an object may adopt different implementations for its public interface. Lastly, all these language mechanisms must be uniformly integrated in a single consistent framework. We believe that the conventional object-oriented model is far too restricted to fulfil these real-time requirements. Language annotations made to conventional languages may result in real-time specification anomalies and non-uniform language constructs. At the University of Twente, we have been working on new object-oriented language mechanisms using the composition-filters approach [1]. Composition filters affect the received and sent messages to or from an object. By proper configuration of filters, one can specify inheritance, delegation, inter-object communications and “real-time constraints” in a single framework.

Concurrent Engineering of Real-Time Systems Through Heterogeneous Prototypes

Abstract: A heterogeneous prototype is an executable system model whose different parts may present different levels of abstraction (maturity), and yet they can be executed together as a total system. Abstract models may be presented by means of a graphical specification language whereas more mature physical models may be presented by programming language code.

A Generic Systems Integration Framework for Large and Time-Critical Systems

Abstract: The concepts presented in this paper propose a pre-planned and integrated development process for large and complex systems which include time-critical components. A system is partitioned into subsystems which are developed relatively independently within an integration framework, but work in an integrated manner — as a “system of systems” — to provide the necessary functionality within an application domain. To do this, we propose a generic systems integration framework (GenSIF) which includes a meta-level of control and management, coordinating the independent development projects. Domain modeling, integration architecture design and infrastructures design are identified as the main elements of the proposed framework. Based on these basic concepts, a generic process model for system development and integration is presented. A possible way to handle time-critical requirements in this environment is discussed.

Real-Time Databases

Abstract: The questions posed by the panel moderator include: what are the correct transaction and data models, the correct database architecture, which applications require real-time databases, and what are the open research questions. I’d like to address these questions in an integrated fashion, twice: once for soft real-time databases (SRTDB) and once for hard real-time databases (HRTDB).

Graphical Design of Real Time Applications

Abstract: The major difficulties encountered by those who have to solve Industrial Real Time Applications, in a multitask computing environment, are: the expression of the dynamic aspect of this type of applications, the number and the variety of these Real Time Executives.

Fundamentals of Real-Time Scheduling (Extended Abstract)

Abstract: For a given set of jobs the general scheduling problem asks for an order according to which the jobs are to be executed such that various constraints are satisfied. Typically, a job is characterized by its execution time, ready time, deadline, and recourse requirements. The execution of a job might or might not be interrupted (pre-emptive) or (non-preemptive scheduling). Over the set of jobs, there is a precedence relation which constrains the order of execution. Specifically, the execution of a job cannot begin until the execution of all its predecessors (according to the precedence relation) is completed. The system on which the jobs are to be executed is characterized by the amounts of resources available.

Formal Methods — How to Make Them Understandable and Usable for Engineers?

Abstract: The discussion was motivated by a long term observation, that each time when theoretically oriented computer scientists deeply involved in formal methods talk with “engineers from practice”, they don’t really understand each other. However, on the background of increasing importance of formal methods in many technically oriented application domains with high safety and reliability requirements, a cooperation between both parties becomes more and more important. This requires a common understanding.

Next Generation of Real-Time Operating Systems: Industrial Prospective

Abstract: The presentation begins with discussing major driving forces which will impact the next generation of real-time operating systems. These are: 1. Approval of the ieee posix standards 1003.4 (Real-time standard), and 1003.4a (Threads). Once these standards will be approved (expected time frame is end of 1992 to middle of 1993), every real-time operating system in order to be successful, must comply with the standard. In other words, the domination of proprietary real-time operating systems will be over! 2. Big players, such as ibm, des, Hewlett-Packard, and Sun, have entered the real-time computer battle by introducing their real-time unix operating systems. Their rt unix systems contain a fully preemptive kernel, fixed priority schedulers, and run on powerful risc-based systems. These are ibm, aix, des osf/1, sun Solaris 2.0, and hp/rt.

The Design of Distributed, Dependable Real-Time Systems Using a Functional Paradigm

Abstract: Distributed real-time computer systems are replacing conventional control systems in many applications. Most of these systems are higly critical ones, e.g. design of a traffic flight control or an industrial process control system. In addition to the specified functional capabilities, these applications demand predictable timeliness and a high level of dependability which includes non functional attributes such as availability, reliability, safety and maintainability [4]. The systematic development of fault-tolerant real-time systems with guaranteed timeliness and dependability requires an appropriate system architecture and a rigorous design methodology.

The State-of-the-Art in Real-Time Computing

Abstract: The emerging generation of complex software systems presents significant challenges that must be addressed with new development technologies. Such systems are highly distributed and employ many heterogeneous processors, some of which may be parallel processors. Additionally, there are many non-functional requirements (related to timing, reliability, security and fault tolerance).

Usability of ATM Connections for Sensitive Real-Time Systems

Abstract: Expected real-time applications like video telephony, video conferences and document retrieval can either take advantage of human’s ability to cope with missing pictures or bad audio quality, or of — time consuming, therefore inconvenient — retransmission.

What Will We Get from Real-Time Research Within the Next 50 Years (1992–2042)

Abstract: It was in St. Maarten where a group of more than 100 computer scientists surprized the rest of the world by stating that they have been doing real-time research for years in order to generate automated solutions for problems and in spite of their efforts and in spite of tricky schedulers, excellent methods, fast CPUs and networks, not all problems concerning real-time have yet been solved.