CASCADE, an Evolutive Construction Set to Build Distributed Multi-Platform Real-Time Data-Acquisition Systems

                Abstract

The paper presents CASCADE and puts a special emphasis on the concepts and the design principles which aimed at achieving a high degree of homogeneity and scalability while leaving openings for the future. A review is made based on the experience gained with the present applications and the most recent developments and on the suitability of CASCADE to recent requests for new applications.

Finally, some conclusions are attempted on the validity of this approach and on its applicability to the demand for data-acquisitions systems in the years to come. The availability of standards, the power of today's technology combined with a modular design has allowed CASCADE to be developed as a consistent set of software building blocks enabling physicists to design data-acquisition systems for a wide spectrum of applications. The current implementation supports configurations based on VMEbus processors running OS-9 as well as UNIX workstations interconnected via VICbus and Ethernet. In addition to the basic data collection, it provides facilities such as parameterised multi-level event building, local and remote event monitoring, data recording, run control and error reporting capabilities. The NOMAD experiment's data-acquisition system, for which CASCADE was originally developed, will be presented. The use by the Energy Amplifier project, the adoption of CASCADE by other experiments, the integration of new platforms such as the Lynx'S operating system, the SCI Scalable Coherent Interface, the CES Fast Data Link and of new services such as event-builder farms will also be described to illustrate the flexibility and the potential of this approach.

The CASCADE package eases the building and the maintenance of distributed multiple platform data-acquisition systems by providing uniformity across the entire architecture. The high degree of modularisation of its components facilitates the integration of the most recent advances in technology. The capabilities of its building blocks in terms of connectivity helps the extension and reconfiguration of systems as imposed by changes in the requirements of the experiments. The debate about the mutual exclusion of generality and performance should certainly be reviewed in light of what today's hardware and software technology permit. This is particularly important when the size and complexity of experiments become so large that the development and integration of special data-acquisition packages dedicated to subsets of the whole system become virtually impossible and unmanageable. Use of such solutions should remain limited to exceptional cases motivated by a particular demand in terms of performance or service, rather than imposed by hardware and software environment considerations.

In that sense a package like CASCADE, although not applicable to experiments with extremely high rate and complexity such as the ones foreseen at LHC, could provide flexibility and consistency in the way many data-acquisition systems for small and medium size experiments could be built.

Submitter's  Name:          Claude Bizeau
Submitter's Institution:    CERN ECP division
Address of Institution:     1211 Geneva 23, Switzerland
Submitter's EMAIL address:  ype@dxcern.cern.ch
Submitter's Telephone number and/or FAX number:
                            Tel: +41 22 767 3194
                            Fax: +41 22 767 5490

Authors of the Paper each with Institution (and Experiment Affiliation, if any):

C. Bizeau, W. Bozzoli, D. Burckhart, R. Divia, F. Lamouche,
J.P. Matheys,Y. Perrin, J. Petersen, A. Saravia, L. Tremblet,
P. Vande Vyvre, A. Vascotto, P. Werner
ECP Division, CERN, 1211 Geneva 23, Switzerland