Resumo em Inglês:

This paper presents a comprehensive approach to describe, deploy and adapt component-based applications having dynamic non-functional requirements. The approach is centered on high-level contracts associated to architectural descriptions, which allow the non-functional requirements to be handled separately during the system design process. This helps to achieve separation of concerns facilitating the reuse of modules that implement the application in other systems. Besides specifying non-functional requirements, contracts are used at runtime to guide configuration adaptations required to enforce these requirements. The infrastructure required to manage the contracts follows an architectural pattern, which can be directly mapped to specific components included in a supporting reflective middleware. This allows designers to write a contract and to follow standard recipes to insert the extra code required to its enforcement in the supporting middleware.

Resumo em Inglês:

This paper makes three contributions to research on QoS-enabled middleware for open distributed real-time embedded (DRE) systems. First, it describes the design and implementation of a dynamic scheduling framework based on the OMG Real-Time CORBA 1.2 specification (RTC1.2) that provides capabilities for (1) propagating QoS parameters and a locus of execution across endsystems via a distributable thread abstraction and (2) enforcing the scheduling of multiple distributable threads dynamically using standard CORBA middleware. Second, it examines the results of empirical studies that show how adaptive dynamic scheduling and management of distributable threads can be enforced efficiently in standard middleware for open DRE systems. Third, it presents results from case studies of multiple adaptive middleware QoS management technologies to monitor and control the quality, timeliness, and criticality of key operations adaptively in a representative DRE avionics system.

Resumo em Inglês:

Nowadays computing environments are made of heterogeneous networked resources, but unlike environments used a decade ago, the current environments are highly dynamic. During a computing session, new resources are likely to appear and some are likely to go offline or to move to some other place. The operating system is supposed to hide most of the complexity of such environments and make it easy to write applications using them. However, that is not the case with our current operating systems. Plan B is a new operating system that attempts to allow the applications and their programmers select and use whatever resources are available without forcing them to deal with the problems created by their dynamic distributed and heterogeneous environments. It does so by using constraints along with a new abstraction used to replace the traditional file abstraction.

Resumo em Inglês:

Reflection is now an established technique for achieving dynamic adaptability of middleware platforms. It provides a clean and comprehensive way to access the internals ofa platform implementation, allowing its customisation in order to achieve the best performance and adequacy under given operation environments and user requirements. In addition, the use ofa runtime component model for the design ofthe internal platform structure facilitates the identification of the elements to be adapted, as all platform aspects are built in terms ofcomponents. The major limitation ofthis approach, however, is related to the multitude ofaspects that make up a middleware platform, together with the requirement ofkeeping platform consis-tencyafter adaptations take place. This paper presents the results of ongoing research contributing to reduce this limitation. The approach is based on the use ofa common meta-model, together with meta-information techniques to provide a uniform way to specify and manipulate platform configurations. Both platform configuration and runtime adaptation are always specified using a small number of building blocks defined in the meta-model. The paper also describes the overall architecture ofthe Meta-ORB platform, which demonstrates this approach, and presents its two implementations: a proof-of-concept prototype written in Python, and a Java-based implementation aimed at supporting mobile devices. The results are also evaluated from a quantitative perspective, according to the requirements ofmultimedia applications, one ofthe major areas of application of reflective middleware.

Resumo em Inglês:

Non-functional requirements (NFRs) are rarely taken in account in most software development processes. There are some reasons that can help us to understand why these requirements are not explicitly dealt with: their complexity, NFRs are usually stated only informally, their high abstraction level and the rare support of languages, methodologies and tools. In this paper, we concentrate on defining how to reason and how to refine NFRs during the software development. Our approach is based on software architecture principles that guide the definition of the proposed refinement rules. In order to illustrate our approach, we adopt it to an appointmernt system.