SUMMARYIn the light of the increasing throughput of local area networks, Networks Of Workstations (NOWs) which provide a Distributed Shared Memory (DSM) have become a convenient and cheaper alternative to parallel architectures in the framework of parallel scientific applications. However, the probability that a failure occurs in such a system made up of a large number of components must not be neglected, especially for long-running applications. This paper presents the design, implementation and performance evaluation of ICARE, a page-based recoverable DSM implemented on top of an ATM-based NOW running the CHORUS microkernel. ICARE relies on a Backward Error Recovery (BER) mechanism, and provides a way to combine both efficiency and high-availability. The fact that checkpoints are stored in volatile memory provides a low-cost fault-tolerance mechanism, as well as the opportunity to exploit the symbiotic relationship between the data replication implemented in DSM systems and that needed for fault-tolerance. Furthermore, ICARE efficiently implements transparent process rollback recovery. Performance evaluations show the efficiency of the ICARE prototype that implements the proposed algorithms.

This paper focuses on the problem of fault tolerance in shared memory multiprocessors, and describes an architecture designed for transparently tolerating processor failures. The Recoverable Shared Memory (RSM) is the novel component of this architecture, providing a hardware supported backward error recovery mechanism which minimizes the propagation of recovery when a processor fails. The RSM permits a shared memory multiprocessor to be constructed using standard caches and cache coherence protocols, and does not require any changes to be made to applications software. The performance of the recovery scheme supported by the RSM is evaluated and compared with other schemes that have been proposed for fault tolerant shared memory multiprocessors. The performance study has been conducted by simulation using address traces collected from real parallel applications.

No abstract

Todays, we assist to the explosive development of mobile computing devices like PDAs and cell-phones, the integration of embedded intelligence (like Web server) in more and more common devices, and the proliferation of wireless communication technologies (IRdA, Bluetooth, IEEE 802.11, GPRS). All these trends contribute to move us closer to the ubiquitous computing world described by Mark Weiser.But while the technology is here, applications, and more important, models and tools for designing future ambient computing systems are still rare. One of the first innovative concept of ubiquitous computing, context-awareness, is still hard to use and understand from a programming perspective. We think that the problem resides in the lack of system support : in traditional computing, operating system offers simple to use and easy to understand abstractions of computational resources. Ubiquitous computing involves an integration of "computing" into the real-world, which is a radically different environment for applications. We think that this environment requires new operating system services and abstractions. Because the real world is made of physical entities, "living" in the physical space, ambient computing software should be able to use abstraction representing such objects, in a simple way.In this paper, we present a light framework to design ubiquitous computing software, called SPREAD. Unlike many approaches which hides too much of the real-world behind traditional computing abstraction, SPREAD defines programming abstraction based on the properties of the physical space. Hence, physical properties, like relative proximity, are used as implicitly in SPREAD as variable addressing in a computer memory. In SPREAD, application (or process) behavior can be "mechanically" driven, in the sense that actions flow can be directly dependent of physical mobility.To support this concept, we introduce a programming and execution model allowing to design computing and information systems driven directly by arranging and moving physical objects in the space. We demonstrate the use of the model to implement a few practical applications, highlighting its simplicity and expression power.