Operating System Research Research Articles

A survey of customizability in operating systems research

An important goal of an operating system is to make computing and communication resources available in a fair and efficient way to the applications that will run on top of it. To achieve this result, the operating system implements a number of policies for allocating resources to, and sharing resources among applications, and it implements safety mechanisms to guard against misbehaving applications. However, for most of these allocation and sharing tasks, no single optimal policy exists. Different applications may prefer different operating system policies to achieve their goals in the best possible way. A customizable or adaptable operating system is an operating system that allows for flexible modification of important system policies. Over the past decade, a wide range of approaches for achieving customizability has been explored in the operating systems research community. In this survey, an overview of these approaches, structured around a taxonomy, is presented.

Near-Optimal Parallel Prefetching and Caching

Recently there has been a great deal of interest in the operating systems research community in prefetching and caching data from parallel disks, as a technique for enabling serial applications to improve input--output (I/O) performance. In this paper, algorithms are considered for integrated prefetching and caching in a model with a fixed-size cache and any number of backing storage devices (disks). The integration of caching and prefetching with a single disk was previously considered by Cao, Felten, Karlin, and Li. Here, it is shown that the natural extension of their aggressive algorithm to the parallel disk case is suboptimal by a factor near the number of disks in the worst case. The main result is a new algorithm, reverse aggressive, with near-optimal performance for integrated prefetching and caching in the presence of multiple disks.

An Introduction to this Special Issue of AI Magazine

Web. Indeed, the authors of each article included in this special issue have promised to create and maintain a URL pointing to a working prototype. Now, almost a year later, we have the fruit of this labor. A series of high quality articles describing a wide range of AI systems that are accessible on the Web right now (or so I hope). The articles describe a broad and diverse set of systems. The AI technologies used span the gamut from machine learning to natural language processing, from case-based reasoning to knowledge representation, and In AAAI-96 I gave an invited talk exhorting AI researchers to deploy their prototype systems on the World Wide Web. Deploying AI systems on the Web provides tangible evidence of the power and utility of AI techniques. Next time you encounter AI bashing, wouldn’t it be satisfying to counter with a handful of well-chosen URLs? Well, see the sidebar. At the conference, Jude Shavlik asked me to edit a special issue of AI Magazine describing AI systems that have the Web as their domain. I suggested that the authors make their prototype systems accessible on the more. Applications include Web page filtering, a grant finder, a FAQ finder, a home page finder, a shopping assistant, and more. Is the challenge of deploying systems on the Web a distraction from our long-term goal of understanding intelligence? I believe that the field benefits from a mixture of long-term and short-term goals and from both empirical and theoretical work. Work toward the goal of deploying systems on the Web is a valuable addition to the current mix for two reasons. Internally, the Web suggests novel research challenges and new constraints on existing techniques. Externally, intelligent Web “apps” will help to enhance the reputation of AI and to counter the popular misconception that “if it works, it ain’t AI.” Database and operating system research has spawned software industries with annual revenues in the billions. I look forward to the day when we have a comparable AI industry. Hopefully, Intelligence on the Web will be a step in that direction. An Introduction to This Special Issue of AI Magazine

Real-time and embedded systems

Real-time systems are those systems in which the correctness of the system depends not only on the logical results of computation but also on the time at which the results are produced [Stankovic 1988]. They span a broad spectrum of complexity from very simple microcontrollers in embedded systems (a microprocessor controlling an automobile engine) to highly sophisticated, complex, and distributed systems (air traffic control for the continental United States). Other real-time systems include command and control systems, process control systems, and multimedia and high-speed communication systems. Some real-time systems are beginning to add expert systems and other AI technology, creating additional requirements and complexities. At least three major trends in the real-time and embedded systems field have had a major impact on its technology: the increased growth and sophistication of embedded systems, the development of more scientific and technological results for hard real-time systems, and the advent of distributed multimedia. Most embedded systems consist of a small microcontroller and limited software situated within some product such as a microwave oven or automobile. However, with the increasing sophistication of such systems, powerful microcontrollers and digital signal processing (DSP) chips are commonly used, as are off-the-shelf real-time operating systems and design and debugging tools. Many people working with embedded systems deal on a daily basis with sensors and data acquisition technology and systems; others construct architectures based on single-board computers (many are still 68000-based, but RISC processors are being used more and more) and buses such as the VME bus. Many people are involved with the programming and debugging of embedded systems, largely using the C programming language and cross development and debugging platforms. Embedded systems may or may not have real-time constraints. Hard real-time systems are those in which missing an important deadline can cause severe consequences, even death. In this area, many fundamental results have been developed. For example, in real-time scheduling, rate monotonic analysis has enabled careful evaluation of many practical systems; the concept and analysis of competitive algorithms has provided important scheduling bounds and limits; and on-line planning has added flexible and dynamic capabilities to real-time systems. Operating systems research has produced predictable primitives, time-constrained synchronization techniques, and reservation and admission-control paradigms. Many other results exist in real-time architecture, fault tolerance, communication protocols, specification and design tools, formal verification, databases and object-oriented systems. Emphasis on all these areas is expected to increase in the foreseeable future. Many hard real-time systems are embedded systems. Distributed multimedia have produced a new set of soft real-time requirements. Real-time principles lie at the heart of distributed multimedia, but without the concomitant high reliability

Distributed operating systems

In the past five years, distributed operating systems research has gone through a consolidation phase. On a large number of design issues there is now considerable consensus between different research groups. In this paper, an overview of recent research in distributed systems is given. In turn, the paper discusses overall system structure, protection issues, file system designs, problems and solutions for fault tolerance and a mechanism that is rapidly becoming very important for efficient distributed systems design: hints. An attempt was made to provide sufficient references to interesting research projects for the reader to find material for more detailed study.

Blocks and procedures

Mössenböck proposes the argument block as a language construct for using statement sequences as arguments for procedures [Mös92]. Such a construct enables users to define their own control structures. Mössenböck is inspired by S MALLTALK -80 [GR83] and claims that most languages lack support of argument blocks but does not take into account a quality which in practice is crucial to their utilization. Stolz [Sto93] therefore falsely concludes the equivalence of argument blocks and anonymous procedures and points out that anonymous procedures were already supported in ALGOL 68 [vWMP + 76]. We show the difference between blocks and procedures and describe implementation patterns which are taken under consideration for the programming language A RISTARCH /L. A RISTARCH /L is supposed to support the distributed operating system A RISTARCH /OS which is under development in S AMOS , a project undertaken by the Operating Systems Research Group at the University of Karlsruhe.

Operating Systems Research Related to the Galactica Net Architecture

Operating Systems Research Related to the Galactica Net Architecture

How Can We Make Operating Systems Research More Effective?

No abstract available.

Technical summary of the second IEEE workshop on workstation operating systems

The broad spectrum of universities, industrial research laboratories, and computer companies represented at the Second IEEE Workshop on Workstation Operating Systems provided a rich snapshot of current activities in operating systems. There were representatives of 19 operating system research projects among the participants and several from commercial offerings. The attendees came from seven countries on three continents: North America, Europe, and South America.Since the last Workshop in 1987, there have been more advances in hardware than in software functions. Software standards continue to emerge in the areas of operating system interfaces, page description languages, window management interfaces, and communication protocols. New software applications exist in the areas of multimedia and multi-node computing. Object-oriented technology is already present in running systems and gaining importance. The areas that the participants perceived needing most future work were operating system abstractions, workstation operation, system responsiveness, input output, network services, management of clusters of workstations, and failure handling.While processor speeds, main memory access speeds, memory density, and secondary storage capacity continue to increase fast, disk seek times have decreased only slightly, and the bandwidth of most local-area networks has not increased at all. FDDI networks are just beginning to be deployed. The software is adjusting to this hardware scenario by using caching at multiple levels of the systems.In the last two years large main memories at individual computing nodes and multi-node computer installations have become common. It is expected that most future computing nodes will have substantial local storage and that high-bandwidth networks will enable the support of continuous media like vocie and video. Input output, to disks, to networks, and to user-oriented devices, is expected to become the central problem in future systems.

Technical summary of the second IEEE workshop on workstation operating systems

The broad spectrum of universities, industrial research laboratories, and computer companies represented at the Second IEEE Workshop on Workstation Operating Systems provided a rich snapshot of current activities in operating systems. There were representatives of 19 operating system research projects among the participants and several from commercial offerings. The attendees came from seven countries on three continents: North America, Europe, and South America.Since the last Workshop in 1987, there have been more advances in hardware than in software functions. Software standards continue to emerge in the areas of operating system interfaces, page description languages, window management interfaces, and communication protocols. New software applications exist in the areas of multimedia and multi-node computing. Object-oriented technology is already present in running systems and gaining importance. The areas that the participants perceived needing most future work were operating system abstractions, workstation operation, system responsiveness, input output, network services, management of clusters of workstations, and failure handling.While processor speeds, main memory access speeds, memory density, and secondary storage capacity continue to increase fast, disk seek times have decreased only slightly, and the bandwidth of most local-area networks has not increased at all. FDDI networks are just beginning to be deployed. The software is adjusting to this hardware scenario by using caching at multiple levels of the systems.In the last two years large main memories at individual computing nodes and multi-node computer installations have become common. It is expected that most future computing nodes will have substantial local storage and that high-bandwidth networks will enable the support of continuous media like vocie and video. Input output, to disks, to networks, and to user-oriented devices, is expected to become the central problem in future systems.

Distributed operating systems

In the past five years, distributed operating systems research has gone through a consolidation phase. On a large number of design issues there is now considerable consensus between different research groups. In this paper, an overview of recent research in distributed systems is given. In turn, the paper discusses overall system structure, protection issues, file system designs, problems and solutions for fault tolerance and a mechanism that is rapidly becoming very important for efficient distributed systems design: hints. An attempt was made to provide sufficient references to interesting research projects for the reader to find material for more detailed study.

APL in operating systems research

This paper will discuss three extensions to APL which will improve its usefulness for research in operating systems and systems programming. They are the introduction of processor variables, encased and bound statement fragments, and pointer or referent variables. APL has demonstrated capabilities in free-wheeling research environments. In an experimental operating system written in APL, much superfluous detail would be suppressed, and the essentials of various algorithms used would be brought to the fore.

Survey of recent operating systems research, designs and implementations

Survey of recent operating systems research, designs and implementations

Databsse system approach the management decision support

Traditional intuitive methods of decision-making are no longer adequate to deal with the complex problems faced by the modern policymaker. Thus systems must be developed to provide the information and analysis necessary for the decisions which must be made. These systems are called decision support systems. Although database systems provide a key ingredient to decision support systems, the problems now facing the policymaker are different from those problems to which database systems have been applied in the past. The problems are usually not known in advance, they are constantly changing, and answers are needed quickly. Hence additional technologies, methodologies, and approaches must expand the traditional areas of database and operating systems research (as well as other software and hardware research) in order for them to become truly effective in supporting policymakers. This paper describes recent work in this area and indicates where future work is needed. Specifically the paper discusses: (1) why there exists a vital need for decision support systems; (2) examples from work in the field of energy which make explicit the characteristics which distinguish these decision support systems from traditional operational and managerial systems; (3) how an awareness of decision support systems has evolved, including a brief review of work done by others and a statement of the computational needs of decision support systems which are consistent with contemporary technology; (4) an approach which has been made to meet many of these computational needs through the development and implementation of a computational facility, the Generalized Management Information System (GMIS); and (5) the application of this computational facility to a complex and important energy problem facing New England in a typical study within the New England Energy Management Information System (NEEMIS) Project.

NIE Plan to Build Educational R&D Capacity

The NIE Task Force on Resources Planning, Office of Research and Development Resources has announced initial plans for carrying out its legislatively mandated responsibility to "build an effective educational research and development system." The report, "Building Capacity for Renewal and Reform," includes both a rationale and recommendations for NIE activity to forward this important aspect of the NIE mission. Copies of the report may be obtained from Judith Cherrington, Office of R&D Resources, NIE, Washington, D.C. 20208. The plans are organized into efforts designed to build (1) internal NIE capability to monitor the R&D system and operating system of education, (2) the R&D system, (3) the linkage and support system, (4) capacity in the operating system (5) and field-initiated research on the processes of educational change. The chapters on the second and third programs, which are most directly pertinent to the mandate underlying the plans, are reproduced here in full text.