US ATLAS production and analysis

D. Adams
06jan03 1355

Introduction

The U.S. has and will continue to have an important role in the production of ATLAS data. As useful data accumulate, the U.S. will also have an important role in defining and providing the environment for analyzing these data.

The project described here serves as major contribution to ATLAS program and guarantees that U.S. scientists will have easy access to ATLAS data.

Goal

The goal of this project is to deliver a distributed production and analysis environment that is easy to use and provides all the functionality required for the ATLAS experiment including complete tracking of the provenance of all produced data. Users (including production managers) produce event data by selecting an appropriate transformation, identifying an input dataset and then submitting a job, e.g. through a web portal. The environment also provides interactive analysis capabilities that are smoothly integrated with any relevant analysis frameworks (e.g. ROOT).

The framework will be GRID-based to provide democratic and economic access to large datasets and extensive computing resources.

Project

This project produces requirements, identifies tools to meet those requirements and then integrates these into an easy-to-use environment for use within the ATLAS experiment. It supports installation and maintenance of this environment at U.S. sites and carries out production tasks assigned to the U.S.

In the common case that identified tools do not already exist or need modification, this project makes requests of the tool maintainers or other relevant projects. The development of anything beyond temporary prototypes is outside the scope of this project. In particular, it is expected that this project will identify tools appropriate for development under the auspices of the U.S. ATLAS software and facilities projects.

This project must be tightly coupled with the international ATLAS plans (including LCG and GDB) to avoid duplication of effort and to guarantee that that fruits of the project serve as contributions to the ATLAS program. The project will make use of the ideas and products from other U.S.-based efforts including GriPhyN, PPDG and GEAC and will serve as conduit to connect these programs to international ATLAS.

Data production is already in place and the current production environment will not be abandoned but will serve as the starting point for this project.

WBS

Scenarios

Management

There are already many relevant U.S. activities that are under way and the goal here is to consolidate these and provide recognition for these efforts. An inclusive board of directors will be selected from participants in these projects and other interested parties and will meet regularly to discuss and approve the project status and plans. The plans include the choice and scheduling of architectures and releases, identification and selection of components, scheduling of deployments, and particpation in production runs.

A project manager is responsible for tracking the project and ensuring its success. This (level two) manager is appointed by the U.S. ATLAS computing project manager.

The project manager may at his discretion appoint level three managers but it is expected that the project will be small enough that this will not be necessary for most categories. Within each level three category most of the effort is contained in a series of tasks. These include architecture, release, component, deployment and production run. A manager is appointed for each task and this manger is reponsible for guaranteeing that the task is carried out in a timely manner and documenting its success or failure.

Resources

This project will make use of computing hardware installed at the ATLAS computing centers and so most of the required resources are people and travel. Most identified software development tasks will be passed on to other groups but effort will be required in the following areas: management, design and integration, deployment and support, and data production. These will require staffing at the level of 2-3 FTE.

It would be ideal to quickly identify new funding to support this project which is a missing piece of the current U.S. ATLAS computing project. However, it is important to start this project as soon as possible and we propose to use resources presently allocated to the U.S. ATLAS software project until new funds are identified. Approximately 1 FTE is already dedicated to efforts that could be properly credited to the new project.

dladams@bnl.gov