Commercial off-the-shelf (COTS) middleware is now widely used to develop distributed real-time and embedded (DRE) systems. DRE systems are themselves increasingly combined to form "systems of systems" that have diverse quality of service (QoS) requirements. Earlier generations of COTS middleware, such as Object Request Brokers (ORBs) based on the CORBA 2.x standard, do not facilitate the separation of QoS policies from application functionality, which makes it hard to configure and validate complex DRE applications. The new generation of component middleware, such as the CORBA Component Model (CCM) based on the CORBA 3.0 standard, addresses the limitations of earlier generation middleware by establishing standards for implementing, packaging, assembling, and deploying component implementations. There has been little systematic empirical study of the performance characteristics of component middleware implementations in the context of DRE systems. This paper therefore provides three contributions to the study of CCM for DRE systems. First, we describe the challenges involved in benchmarking different CORBA Component Model (CCM) implementations. Second, we describe key criteria for comparing different CCM implementations using key black-box and white-box metrics. Third, we describe the design of our CCMPerf benchmarking suite to illustrate test categories that evaluate aspects of CCM implementation to determine their suitability for the DRE domain. We demonstrate CCMPerf by using it to collect metrics from a CCM implementation designed for DRE applications.

Commercial of-the-shelf (COTS) distribution middleware is gaining acceptance in the distributed real-time and embedded (DRE) community as (1) the cost and time required to develop and verify DRE applications precludes developers from implementing DRE applications from scratch and (2) implementations of standard COTS middleware specifications, such as CORBA, mature. Although standard COTS specifications define the interfaces and policies to provision DRE application resources end-to-end, they do not yet provide sufficient abstractions to separate quality of service (QoS) policy configurations and adaptations from application functionality. DRE application developers must therefore configure QoS policies and program adaptation mechanisms in an ad hoc way. This tight-coupling tends to scatter the code that ensures end-toend QoS throughout many parts of DRE applications, making it hard to configure, validate, modify, and evolve complex DRE applications consistently. This paper provides three contributions to the study of the development of QoS-enabled DRE applications. First, we illustrate how standard component-based middleware can be enhanced to flexibly compose static QoS provisioning policies with application logic. Second, we describe how adaptive middleware capabilities enable developers to abstract and encapsulate reusable dynamic QoS provisioning and adaptive behaviors. Third, we illustrate how component-based middleware and adaptive middleware capabilities can be integrated to provide a total QoS provisioning solution for DRE applications.

Networked control systems could possibly constitute the next logical step in the evolution of control, leading to the convergence of control with communication and computing. A central challenge is that traditional digital control methods cannot be directly applied to such systems. However, if appropriate system abstractions can be engineered, then such methods and theory can still be utilized. Our thesis is that a well designed middleware framework can indeed manufacture such an abstraction of virtual collocation, and thereby, propel the further proliferation of networked control systems. In this thesis, we present such a middleware framework for networked control systems. Central to this framework is Etherware, a message oriented component middleware for such systems. We begin with a detailed description of the design and architecture of Etherware, and illustrate Etherware based development of networked control systems through a fairly complex traffic control testbed application. Building on the middleware, we address

Krishnakumar Balasubramanian, Nanbor Wang & Christopher Gill Douglas C. Schmidt fkitty,nanbor,cdgillg@cs.wustl.edu schmidt@uci.edu Department of Computer Science Electrical & Computer Engineering Washington University, St.Louis University of California, Irvine Abstract The growing complexity of building and validating software is a challenge for developers of distributed real-time and embedded (DRE) applications. While DRE applications are increasingly based on commercial off-the-shelf (COTS) hardware and software components, substantial time and effort are spent integrating components into applications. Integration challenges stem largely from the lack of higher level abstractions for composing complex applications. As a result, considerable application-specific "glue code" must be written, only to be rewritten from scratch when building subsequent DRE applications.

Abstract. Rapid advances in networking, hardware, and middleware technologies are facilitating the development and deployment of complex grid applications, such as large-scale distributed collaborative scientific simulation, analysis of experiments in elementary particle physics, distributed mission training and virtual surgery for medical instruction. These predominantly collaborative applications are characterized by their very high demand for computing, storage and network bandwidth requirements. Grid applications require secure, controlled, reliable, and guaranteed access to different types of resources, such as network bandwidth, computing power, and storage capabilities, available from multiple service providers. Moreover, they demand multiple, simultaneous end-to-end quality of service (QoS) properties, such as delay guarantees, jitter guarantees, security, scalability, reliability and availability guarantees, and bandwidth and throughput guarantees, for their effective operation. Existing grid infrastructure middleware, such as Globus, ICENI, and Legion,

Although component-based software development has widespread acceptance in the enterprise business and desktop application domains, developers of distributed real-time and embedded (DRE) systems have encountered limitations with the available component middleware platforms, such as the CORBA Component Model (CCM) and the Java 2 Enterprise Edition (J2EE). These limitations often preclude developers of DRE systems from fully exploiting the benefits of component software. In particular, component middleware platforms lack standards-based publisher/subscriber communication mechanisms that support key quality-of-service (QoS) requirements, such as low latency, bounded jitter, and end-to-end operation priority propagation. QoS-enabled publisher/subscriber services are available in object middleware platforms, such as Realtime CORBA, but such services have not been integrated into component middleware due to a number of development and configuration challenges.

Commercial off-the-shelf (COTS) middleware is increasingly used to develop distributed real-time and embedded (DRE) systems. DRE systems are themselves increasingly combined using wireless and wireline networks to form "systems of systems" that have diverse quality of service (QoS) requirements. Conventional COTS middleware does not facilitate the separation of QoS policies from application functionality, which makes it hard to configure and validate complex DRE applications. Component-based middleware addresses limitations of COTS middleware by establishing standards for implementing, packaging, assembling, and deploying component implementations. There has been little systematic empirical study of the performance characteristics of component middleware implementations in the context of DRE systems. This paper therefore provides three contributions to the study of component-based middleware. First, we describe the challenges involved in benchmarking different CORBA Component Model (CCM) implementations. Second, we describe key criteria for comparing different CCM implementations using key black-box and white-box metrics. Third, we describe the design of our CCMPerf benchmarking suite to illustrate test categories that evaluate aspects of CCM implementation to determine their suitability for the DRE domain. Our preliminary results underscore the importance of applying a range of metrics to quantify CCM implementations effectively.

SCM products are large and monolithic, difficult to adapt and evolve, with high entry cost. This paper describes a new approach to SCM in which the system is built from, potentially heterogeneous, existing pieces, with assembly mechanisms that enforce high-level properties. The approach does not provide a simple SCM tool, but a family of tools that is easily customized, fits both lowend users (only the required functionalities are present at a very low cost), as well as high-end users (for which very advanced features and/or specific features can be easily added). The paper describes the concepts and mechanisms of federations, and shows how our federation technology was used to develop a family of SCM systems.

The growing complexity of building and validating software is a challenge for developers of distributed real-time and embedded (DRE) systems. While building blocks of DRE systems are increasingly based on commercial off-the-shelf (COTS) components, substantial time and effort are spent integrating components into systems due to the lack of higher level abstractions for composing complex systems. As a result, considerable system-specific "glue code" must be written, only to be rewritten from scratch when building subsequent systems.

Component-based applications require runtime support to be able to guarantee non-functional properties.