M. Donszelmann

DIM, a portable, light weight package for information publishing, data transfer and inter-process communication

Abstract The real-time systems of HEP experiments are presently highly distributed, possibly on heterogeneous CPUs. In many applications, there is an important need to make information available to a large number of other processes in a transparent way. For this purpose the “RPC-like” systems are not suitable, since most of them rely on polling from the client and one-to-one connections. DIM is a very powerful alternative to those systems. It provides a named space for processes to publish information (Publishers) and a very simple API for processes willing to use this information (Subscribers). It fully handles error recovery at the Publisher and Subscriber level, without additional software in the application. DIM is available on a large variety of platforms and operating systems with C and C++ bindings. It is presently used in several HEP experiments, while it was developed in the DELPHI experiment and is maintained at CERN. We shall present its capabilities and examples of its use in HEP experiments in domains ranging from simple data publishing to event transfer, process control or communication layer for an Experiment Control Package (SMI++). We shall also present prospectives for using it as communications layer for future experiments control systems.

Architecture and performance of the DELPHI trigger system

Abstract This paper describes the trigger system of the DELPHI detector at LEP. It reports on the most relevant aspects of the hardware and shows the software strategies that have been developed to optimize its use. In 1993 the structure of the trigger in four decision levels has become fully operational and data collected during this period have been used to study the trigger performance. Various final state channels such as μ + μ − , e + e − and hadronic events were selected and their trigger efficiencies were calculated as a function of the polar angle θ. The results obtained indicate that, for any of the event topologies considered, the DELPHI trigger efficiency is independent of the θ angle, and, furthermore, the attained efficiency values are determined to be very close to 100% within an extremely good precision. This is a consequence of the high redundancy presently provided by all the DELPHI subdetectors. In addition to this analysis, events containing isolated particles either in the barrel or in the forward regions have been selected to evaluate the trigger response to single particles. Hence, trigger efficiencies for single particles have also been computed for charged tracks as a function of the momentum and for photons as a function of the deposited electromagnetic energy.

The DELPHI distributed information system for exchanging LEP machine related information

Abstract An information management system was designed and implemented to interchange information between the DELPHI experiment at CERN and the monitoring/control system for the LEP (Large Electron Positron Collider) accelerator. This system is distributed and communicates with many different sources and destinations (LEP) using different types of communication. The system itself communicates internally via a communication system based on a publish-and-subscribe mechanism, DIM (Distributed Information Manager). The information gathered by this system is used for on-line as well as off-line data analysis. Therefore it logs the information to a database and makes it available to operators and users via DUI (DELPHI User Interface). The latter was extended to be capable of displaying “time-evolution” plots. It also handles a protocol, implemented using a finite state machine, SMI (State Management Interface), for (semi-)automatic running of the Data Acquisition System and the Slow Controls System.

Design and performance of the DELPHI data acquisition system

The data acquisition system of the DELPHI experiment at LEP allows the recording of data from the front-end electronics of 15 different detectors housed in about 100 Fastbus crates. Each detector has the possibility of monitoring its data in parallel with the global data acquisition as well as of running in a stand-alone mode. A description is presented of the design of the system and the performance achieved so far. Emphasis is given to the readout system, and an overview is provided of the systems for control and monitoring. >

Basic concepts and architectural details of the Delphi trigger system

Delphi (DEtector with Lepton, Photon and Hadron Identification) is one of the four experiments of the LEP (Large Electron Positron) collider at CERN. The detector is laid out to provide a nearly 4/spl pi/ coverage for charged particle tracking, electromagnetic, hadronic calorimetry and extended particle identification. The trigger system consists of four levels. The first two are synchronous with the BCO (beam cross over) and rely on hardwired control units, while the last two are performed asynchronously with respect to the BCO and are driven by the Delphi host computers. The aim of this paper is to give a comprehensive global view of the trigger system architecture, presenting in detail the first two levels, their various hardware components and the latest modifications introduced in order to improve their performance and make more user friendly the whole software user interface. >