IONS 3.1 Scheme 48 Kali Scheme is implemented as an extension to Scheme 48 [Kelsey and Rees 1994], an implementation of Scheme [Clinger and Rees 1991]. Scheme is a lexically scoped dialect of Lisp. Scheme 48 is based on as byte-coded interpreter written in a highly optimized, restricted dialect of Scheme called Pre-Scheme, which compiles to C. Because of the way it is implemented, the system is very portable and is reasonably efficient for an interpreted system. 2 Unlike other Scheme implementations, 2 Scheme 48 is roughly 10-15 times slower slower than a highly optimized Scheme compiler generating native code [Kranz et al. 1986]. (define-record-type thread : : : continuation : : : ) (define current-thread : : : ) (define (spawn thunk) (let ((thread (make-thread))) (set-thread-continuation! thread (lambda (ignore) (thunk) (terminate-current-thread))) (context-switch thread))) (define (context-switch thread) (add-to-queue! runnable-threads current-thread) (switch-to-thread thre...

Rights to individual papers remain with the author or the author's employer. Permission is granted for noncommercial reproduction of the work for educational or research purposes. This copyright notice must be included in the reproduced paper. USENIX acknowledges all trademarks herein.

In group communication, in order to deliver multicast messages reliably in a group, it is common practice for each member to maintain copies of all messages it sends and receives in a bu er for potential local retransmission. The storage of these messages is costly and bu ers may grow out of bound. A form of garbage collection is needed to address this issue. Garbage collection occurs once a process learns that a message in its bu er has been received by every process in the group. The message is declared stable and is released from the bu er. An important part of garbage collection is message stability detection. This dissertation presents the result of an investigation into message stability detection protocols. A number of message stability detection protocols used in popular reliable multicast protocols are studied with a focus on their performance in large scale settings. This dissertation proposes anewgossip-style protocol with improved scalability and fault tolerance. This dissertation also shows that by adding a hierarchical structure to the set of basic protocols, their performance can be signi cantly improved when the number of participants is large.

. A critical challenge to creating effective open multi-agent systems is allowing them to operate effectively in the face of potential failures. In this paper we present an experimental evaluation of a set of domain-independent services designed to handle the failure modes ("exceptions") that can occur in such environments, applied to the wellknown "Contract Net" multi-agent system coordination protocol. We show that these services can produce substantially more effective fault handling behavior than standard existing techniques, while allowing simpler agent implementations. Keywords: contract net, failure detection, failure resolution, fault handling, multiagent 1 1. The Challenge A critical challenge to creating effective agent-based systems is making them robust in the face of potential failures. Most work to date on multi-agent systems has focused, however, on supporting such basic functionality such as matchmaking (Decker, Sycara et al. 1997) and inter-agent communication (Fin...

This paper discusses the cooperation exception handling model that comes along with a mechanism for multi-party interaction, in order to support the development of robust distributed applications running over a local area network. Lessons learnt from this work and its relation with today's common practice in the area of distributed computing are further considered.

Network programming is notoriously hard to understand: one has to deal with a variety of protocols (IP, ICMP, UDP, TCP etc), concurrency, packet loss, host failure, timeouts, the complex sockets interface to the protocols, and subtle portability issues. Moreover, the behavioural properties of operating systems and the network are not well documented.

. Increasingly, complex artifacts such as cars, planes and even software are designed using large-scale and often highly distributed collaborative processes. A key factor in the effectiveness of these processes concerns how well conflicts are managed. Better approaches need to be developed and adopted, but the lack of systematization and dissemination of the knowledge in this field has been a big barrier to the cumulativeness of research in this area as well as to incorporating these ideas into design practice. This paper describes a growing repository of conflict management expertise, built as an augmentation of the MIT Process Handbook, that is designed to address these challenges. 1. The Challenge Increasingly, complex artifacts such as cars, planes and even software are designed using large-scale and often highly distributed collaborative processes. Conflict (i.e. incompatibilities between design decisions and/or goals) is common in such highly interdependent activities. In on...

Monitoring the behaviour of computing systems is an important task. In active database systems, a detected system behaviour leads to the triggering of an ECA (event-condition-action) rule. ECA rules are employed for supporting database management system functions as well as external applications. Although distributed database systems are becoming more commonplace, active database research has to date focussed on centralised systems. In distributed debugging systems, a detected system behaviour is compared with the expected system behaviour. Di erences illustrate erroneous behaviour. In both application areas, system behaviours are speci ed in terms of events: primitive events represent elementary occurrences and composite events represent complex occurrence patterns. At system runtime, speci ed primitive and composite events are monitored and event occurrences are detected. However, in active database systems events are monitored in terms of physical time and in distributed debugging systems events are monitored in terms of logical time. The notion of physical time is di cult in distributed systems because

. In this paper, we present Bast, an extensible library of protocol classes. The latter is aimed at helping system programmers to build distributed programming environments. Protocol classes constitute the basic structuring components for higher-level programming models, such as the transactional model, and add flexibility to distributed environments. We focus on classes that implement a generic agreement protocol named DTM (Dynamic-Terminating-Multicast). To the programmer, the DTM generic protocol appears as a set of classes that can be specialized to solve agreement problems in distributed systems. In particular, we show how those classes can be derived to build atomic commitment and reliable total order protocols. An overview of the Smalltalk design and implementation of the Bast library is also presented. 1 Introduction This paper describes Bast, an extensible class library of distributed protocols. Bast is aimed at assisting system programmers in building distributed programm...

Abstract: Web services technology is all about distributed computing. There is no fundamentally new basic concept behind this and related technologies. What is really new is the reach of Web services and its ubiquitous support by literally all major vendors. Most likely, heterogeneity will at the end no longer be an obstruction for distributed applications. This will have impact on application architectures, middleware, as well as the way in which people will think about computing and businesses use computing resources. We sketch these impacts as well as some exemplary research work to be done to actually build the outline environment. 1