PRISMA Workshop Noordwijk, The Netherlands, September 24–26, 1990 Proceedings
chapter
Many different paradigms for parallel programming exist, nearly each of which is employed in dozens of languages. Several researchers have tried to compare these languages and paradigms by examining the expressivity and flexibility of their constructs. Few attempts have been made, however, at practical studies based on actual programming experience with multiple languages. Such a study is the topic...
chapter
An important feature of database technology of the nineties is the use of distributed computation for speeding up the execution of complex queries. Today, the use of parallelism is tested in several experimental database architectures and a few commercial systems for conventional select-project-join queries. In particular, hash-based fragmentation is used to distribute data to disks under the control...
chapter
chapter
Conventional data models embodied by current database management systems (DBMS) do not provide satisfactory support for emerging applications. A major reason for this deficiency is the absense of concepts for complex object processing. In this paper, we explain the motivation and key properties of a new data model explicitly designed for the management of complex objects. Furthermore, the most important...
chapter
The paper discusses the parallel programming lessons learnt from the ESPRIT SuperNode project that developed the T800 Transputer. After a brief review of some purportedly portable parallel programming environments, the Genesis parallel benchmarking project is described. The next generation of Transputer components are being developed in the ESPRIT-2 PUMA project and the goals of this project are briefly...
chapter
Data structures such as sets, relations and graphs occur frequently in Artificial Intelligence applications. When these structures are large, as in deductive databases and knowledge bases, the performance of conventional computers is poor. The generic operations on these structures, such as pattern-directed search, set intersection and transitive closure, all involve associative matching of variable-length...
chapter
The European Declarative System (EDS) project (ESPRIT II EP2025), supported by the Commission of the Europen Communities (CEC DG XIII/A/4) is developing a system supporting a parallel relational database, along with parallel LISP and PROLOG languages implementations. This paper describes a performance model for a JOIN under the relational database on the multi-processor distributed store EDS prototype...
chapter
This paper describes the parallel Relational Database Management System (RDBMS) currently being designed and built by the ESPRIT II EDS project. The system components include a high performance distributed store parallel machine consisting of up to 256 processing elements connected by a delta network, an operating system designed for the parallel system, and a RDBMS which exploits both inter and intra...
chapter
Using a large multiprocessor, consisting of 100 processing nodes, in the area of data and knowledge processing poses challenging research and engineering questions. The PRISMA project presented here has addressed many of the problems encountered by the design and the construction of such a system. Among the results obtained are an implementation of a parallel object-oriented language, a hardware platform...
chapter
The design of a large database management system on top of an object-oriented platform which is implemented on a tightly coupled multiprocessor system requires a large number of design decisions. There exist no methodology to evaluate the consequences and compare the differences between them. Looking into experimental methodology in physics we find a lot of similarities and techniques we think that...
chapter
The paper discusses an implementation of POOL-X that aims to be all of available, reliable, portable, modifiable and efficient. This implementation consists of both compiler and run-time-support system. In order to attain the stated goals, the system is kept as simple as possible. Design trade-offs are be discussed, and the final system is evaluated.
chapter
The recovery methods for main-memory database systems are mostly based on maintaining a disk based log and checkpoint. The information written to the log usually consists of low level information, before-after images or single tuple updates. In this paper we examine the effect of writing higher level update information to the log. This results in less logging overhead during normal processing, but...
chapter
Integrity constraints form an important part of a data model. Therefore, a complete integrity constraint handling subsystem is considered an important part of any modern DBMS. In implementing an integrity constraint handling subsystem, there are two major problem areas: providing enough functionality and delivering good performance in constraint enforcement. In the PRISMA project, an integrity constraint...
chapter
Traditional query optimizers produce a fixed query evaluation plan based on assumptions about data distribution and processor workloads. However, these assumptions may not hold at query execution time. In this paper, we propose a dynamic query processing scheme and we present the performance results obtained by simulation of a queueing network model of the proposed software architecture.
chapter
The Oyster evaluation framework is developed to support high level simulation of computer architectures. A designer can make a simulation model of a proposed architecture, and gets feedback in the form of performance figures, and an analysis of the architecture. Oyster has a layered structure, allowing for both high level descriptions, low level specifications and the possibility to incorporate existing...
chapter
We describe a parallel implementation of a subset of the EQUIP expert system. In EQUIP, diagnostic reasoning is applied to the domain of silicon chip production control. In order to diagnose all probable process faults that occurred in the manufacturing of IC's, the necessary computations in the inference part of the program become very time-consuming. To reduce this computation time, a parallel inference...
chapter
This paper presents an overview of the problems encountered in implementing POOL, a parallel object-oriented language, along with solutions found for them. In order to get insight in the problem domain, also a short introduction to the language has been supplied.
chapter
This article describes the automatic recovery algorithm and its implementation for the POOSS stable file system. The recoverable faults by our scheme include non-consecutive disk loss, single duplication loss, single duplication corruption, and duplication inconsistency. The recovery algorithm is implemented on the 100-node parallel machine (POOMA). One important feature of the recovery algorithm...
