"This paper presents a non-invasive method for measuring interrupt latency and process dispatch latency. The significance of these measurements is discussed in the context of operating systems, real-time systems and performance measurement. The target of this method is a real-time UNIX operating system running on an MVME147 / VMEbus hardware platform."--Author's abstract.

A growing concern of mine has been the unrealistic expectations for new computer-related technologies introduced into all kinds of organizations. Unrealistic expectations lead to disappointment, and a schizophrenic approach to the introduction of new technologies. The UNIX and real-time UNIX operating system technologies are major examples of emerging technologies with great potential benefits but unrealistic expectations. Users want to use UNIX as a common operating system throughout large segments of their organizations. A common operating system would decrease software costs by helping to provide portability and interoperability between computer systems in today's multivendor environments. Users would be able to more easily purchase new equipment and technologies and cost-effectively reuse their applications. And they could more easily connect heterogeneous equipment in different departments without having to constantly write and rewrite interfaces. On the other hand, many users in various organizations do not understand the ramifications of general-purpose versus real-time UNIX. Users tend to think of "real-time" as a way to handle exotic heart-monitoring or robotics systems. Then these users use UNIX for transaction processing and office applications and complain about its performance, robustness, and reliability. Unfortunately, the users don't realize that real-time capabilities added to UNIX can provide better performance, robustness and reliability for these non-real-time applications. Many other vendors and users do realize this, however. There are indications even now that general-purpose UNIX will go away as a separate entity. It will be replaced by a real-time UNIX. General-purpose UNIX will exist only as a subset of real-time UNIX.

From the basics to the most advanced quality of service (QoS) concepts, this all encompassing, first-of-its-kind book offers an in-depth understanding of the latest technical issues raised by the emergence of new types, classes and qualities of Internet services. . This books provides sufficient depth for major QoS concepts and architectures. The book provides end-to-end QoS guidance for real time multimedia communications over the Internet. It offers you a multiplicity of hands-on examples and simulation script support, and shows you where and when it is preferable to use these techniques for QoS support in networks and Internet traffic with widely varying characteristics and demand profiles.

The 47 papers in this volume provide a useful reference tool for the state-of-the-art research in real-time programming.