Skip to main content

Definition and Design of an Open Dependable Distributed System Architecture

Objective

The aim of DELTA-4 is to formulate, develop and demonstrate an open, fault-tolerant distributed system architecture. The proposed project has direct applicability to wide application areas which include computer-integrated manufacturing, office systems, integrated information, processing systems and process control systems.
The Delta-4 architecture features original fault concepts for the development and design of dependable open distributed computing systems. To tackle the complex dependability validation task, a comprehensive strategy incorporating both objectives of the validation process (fault removal and fault forecasting) has been followed that is based on the use of complementary validation methods. The methods considered include formal verification of the design of twospecific reliable communication protocols, experimental validation of the implementation of the fault tolerance mechanisms by means of physical fault injection and analytical modeling and evaluation of hardware and software. The paper summarixes the main characteristics (objectives, implementation and results) for each of the methods applied in this validation strategy.

The aim of open systems interconnection was to formulate, develop and demonstrate an open, fault tolerant distributed system architecture.
As an open system, Delta-4 has 3 important properties:
implementations based on off the shelf heterogeneous computer systems are possible (the fault tolerant properties of the architecture are in no way dependent on the use of proprietary, fail safe or self checking processors);
the architecture conforms to the open system interconnection (OSI) model and implementation are able to coexist with and interwork with systems communicating by the use of current standard International Standardization Organization (ISO) OSI protocols;
the dependability and distribution properties of the architecture are offered in a transparent way to the user (both incremental dependability and incremental performance are offered on a service by service basis).

The proposal has both a generic component and an implementation component. The generic component contributes to basic concepts in the areas of dependability, multipoint communications, system administration and open distributed processing. The implementation component serves to demonstrate the validity of the concepts; the use of an experimental site and pilot sites is intended to enable their correctness to be established.

An open workshop presented a number of demonstrations, including the remote banking facility implemented in the DELTA-4 technology. The workshop also demonstrated, at system level, the effectiveness of the cloning of data and processes in achieving a dependable system. Results have been obtained in the area of system security based on fragmentation and scattering supported by DELTA-4 distributed technology.
More work is required on the real time and performance aspects of this advance in distributed systems.
As an open system, DELTA-4 has three important properties:

- Implementations based on "off-the-shelf" heterogeneous computer systems are possible; the fault-tolerant properties of the architecture are in no way dependent on the use of proprietary, fail-safe or self-checking processors.
- The architecture conforms to the OSI model and implementations are able to coexist with and interwork with systems communicating by the use of current standard ISO/OSI protocols.
- The dependability and distribution properties of the architecture are offered in a transparent way to the user. Both incremental dependability and incremental performance are offered on a service-by-service basis.

The proposal has both a generic component and an implementation component. The generic component is already contributing to basic concepts in the areas of dependability, multi-point communications, system administration and open distributed processing; these concepts are of value to other ESPRIT projects and to the standards community. The implementation component serves to demonstrate the validity of the concepts; the use of an experimental site and pilot sites is intended to enable their correctness to be established.

Coordinator

Ferranti International plc
Address
Simonsway Wythenshawe
M22 5LA Manchester
United Kingdom

Participants (12)

BULL SA
France
Address
Tour Bull 1 Place Carpeaux
92800 Puteaux
Centre National de la Recherche Scientifique (CNRS)
France
Address
7 Avenue De Colonel Roche
31077 Toulouse
Fraunhofer-Gesellschaft zur Förderung der Angewandten Forschung eV (FhG)
Germany
Address
Sebastian-kneipp-straße 12-14
76131 Karlsruhe
INSTITUTO DE ENGENHARIA DE SISTEMAS E COMPUTADORES
Portugal
Address
Apartado 10105, Rua Alves Redol, 9
1017 Lisboa
MARI Applied Technologies Ltd
United Kingdom
Address
Boldon Business Park Witney Way
NE35 9PE Boldon
NATIONAL RESEARCH COUNCIL OF ITALY
Italy
Address
Via G. Moruzzi 1
56124 Pisa
Renault Automation
France
Address
8-10 Avenue Émile Zola
92109 Boulogne-billancourt
Sema Group SA
France
Address
16-20 Rue Barbes
92126 Montrouge
UNIBANQUE-CREDIT AGRICOLE
France
Address
40 Rue D'oradour Sur Glane
75015 Paris
UNIVERSITY OF NEWCASTLE UPON TYNE
United Kingdom
Address

NE1 7RU Newcastle Upon Tyne
United Kingdom Atomic Energy Authority (UKAEA)
United Kingdom
Address
11 Charles Ii Street
SW1Y 4QP London
Université de Grenoble I (Université Joseph Fourier)
France
Address
385 Avenue De La Bibliothèque
38041 Grenoble