Atomicity in System Design and Execution
(Proceedings of Dagstuhl-Seminar 04181)
J.UCS Special Issue
Jon Burton (University of Newcastle upon Tyne, UK)
j.i.burton@ncl.ac.uk
Cliff B. Jones (University of Newcastle upon Tyne, UK)
cliff.jones@ncl.ac.uk
Abstract: The present volume contains a selection of papers
arising from a five-day workshop on "Atomicity in System Design
and Execution" that took place in Schloss Dagstuhl in Germany in
April 2004 and was attended by 32 people from various research
areas. This workshop was organised by Cliff Jones and Alexander
Romanovsky of Newcastle University, Dave Lomet of Microsoft Research
Redmond and Gerhard Weikum of Saarbruecken University.
The concept of atomicity is widely used as a means of abstracting
complex concurrent computations into simpler ones which exhibit less
concurrency. However, the interpretations and roles of the atomicity
concept(s) vary substantially across different research
communities. For example, the emphasis in database systems is on
algorithms and implementation techniques for atomic transactions,
whereas in dependable systems and formal methods atomicity is viewed
as an intentionally imposed (or sometimes postulated) property of
system components to simplify designs and increase dependability. On
the other hand, all communities agree on the importance of gaining a
deeper understanding of composite and relaxed notions of
atomicity. The Dagstuhl workshop brought together practitioners from a
number of different communities using atomicity — database and
transaction processing systems, fault tolerance and dependable
systems, formal methods for system design and correctness reasoning,
and hardware architecture and programming languages — in order
to explore the implications of their different concepts of atomicity,
to try to find common ground and to formulate a "manifesto"
for future research in this area. The hope of collaborations of this
type is that it will eventually be possible to unify the different
scientific view-points into more coherent foundations,
system-development principles, design methodologies and usage
guidelines.
During the week in Dagstuhl, a variety of talks and presentations were
given on recent results, and a number of very productive discussions
were held on areas in which further work is needed in order to advance
and improve the use of the atomicity concept.
The first paper in this volume — "The Atomic
Manifesto" — constitutes one of the main fruits of those
discussions and details the current state of research and challenges
which lie ahead. Of the remaining papers, two contain more detailed
blueprints for further research, while the others relate to results
presented at Dagstuhl. We hope that these papers give a faavour of the
breadth of interests represented at the workshop and of the great
wealth of research challenges which face us in the area of atomicity,
and conclude by looking forward to the second Dagstuhl workshop on
atomicity to be held in March 2006.
Jon Burton Cliff Jones
(Newcastle, April 22, 2005)
|