[FFA_CEBAF_Collab] Sirepo

[Jay Benesch]

Colleagues, 

I bring this to your attention because of the material beginning on page 26.  I quote the first part of that below the arxiv info. 

Jay 

https://arxiv.org/abs/2203.10377

Nurturing the Industrial Accelerator Technology Base in the US
A. M. M. Todd, R. Agustsson, D. L. Bruhwiler, J. Chunguang, S. C. Gottschalk, A. Kanareykin, A. Murokh, J. W. Rathke, M. Ruelas, V. Yakovlev, K. Yoshimura

    The purpose of this white paper is to discuss the importance of having a world class domestic industrial vendor base, capable of supporting the needs of the particle accelerator facilities, and the necessary steps to support and develop such a base in the United States. The paper focuses on economic, regulatory, and policy-driven barriers and hurdles, which presently limit the depth and scope of broader industrial participation in US accelerator facilities. It discusses the international competition landscape and proposes steps to improve the strength and vitality of US industry. 

Comments: 	contribution to Snowmass 2021

9.4 First example: Sirepo
Sirepo is an open source framework [24-28] for bringing scientific, engineering or educational 
software to the cloud, with a GUI that works in any modern browser on any computing device 
with sufficient screen size, including tablets. The Sirepo client is built on HTML5 technologies, 
including the JavaScript libraries Bootstrap and Angular. The D3.js library is used for 2D graphics, 
while VTK.js is used for 3D. The supported codes and dependencies are containerized via Docker, 
an open platform for distributed applications. RadiaSoft has developed open source software and 
expertise for building, deploying and executing scientific codes in Docker containers, and the 
corresponding images are publicly available. These containers are compatible with the Shifter 
containerization technology at the NERSC supercomputing center, which enables a Sirepo server 
to automatically launch jobs at NERSC.

A free Sirepo scientific gateway is available to the particle accelerator community [26], 
providing a broad selection of supported codes. The accelerator tracking codes include MAD-X, 
elegant, Synergia, OPAL and Zgoubi. Presently under development, the MAD-X sequence file 
format will be used as a common format to enable rapid code benchmarking and sequential use 
of multiple codes. The loosely coupled cloud-based architecture of Sirepo enables coupling with 
other sophisticated software and systems. This is another reason for the enterprise approach. At 
NSLS-II for example, the DAMA group is integrating Sirepo/SRW with their BlueSky [29,30] 
software for experimental control and data management.

Sirepo has been designed to transcend the limitations that discourage many scientists from 
working with GUI-driven applications. Sirepo can import the necessary input, data, or 
configuration files for the codes that it supports, so experts can quickly transfer their simulation 
results to a GUI user. Likewise, the GUI can export a zip file with everything needed to run the 
identical simulation from the command line.