Data production of a large Linux PC Farm for the CDF experiment

TL;DR

This paper describes the design and implementation of a large Linux PC farm for the CDF experiment, achieving high data throughput for particle physics data collection.

Contribution

It introduces a custom control system and network architecture that enables stable, high-rate data production on a large PC cluster for scientific experiments.

Findings

Achieved a stable data production rate of 2 TByte per day.

Designed a scalable system meeting 20 MByte/sec data rate during Run II.

Successfully integrated hardware and software for high-throughput data processing.

Abstract

The data production farm for the CDF experiment is designed and constructed to meet the needs of the Run II data collection at a maximum rate of 20 MByte/sec during the run. The system is composed of a large cluster of personal computers (PCs) with a high-speed network interconnect and a custom design control system for the flow of data and the scheduling of tasks on this PC farm. The farm explores and exploits advances in computing and communication technology. The data processing has achieved a stable production rate of approximately 2 TByte per day. The software and hardware of the CDF production farms has been successful in providing large computing and data throughput capacity to the experiment.