TR-2020-05

Scientific Computing Equipment (SCE) Decommissioning in the University Community, 2017-2023: A Survey of the Coalition for Advanced Scientific Computing (CASC)

Hakizumwami Birali Runesha; Mark J. Dietrich; Andrew A Chien. 8 July, 2020.
Communicated by Andrew Chien.

Abstract

As part of the NSF-funded "Extending the Productive Life of Scientific Computing Equipment" project, we conducted a survey of members of the Coalition for Advanced Scientific Computing (CASC) about their experience and plans for decommissioning scientific computing equipment (SCE). The survey was introduced at CASCís April 1, 2020 virtual meeting, attended by an estimated 88 members. Twenty-two (22) responses were received -- a 25% response rate.

Several important conclusions can be drawn from the survey:

● The amount of decommissioned SCE is large and growing.

○ Equipment decommissioned in 2017-2019 has aggregate compute power comparable to a system in the Top 500 worldwide (roughly #490 on the June 2019 list).

○ Equipment planned for decommission in 2020-23 has aggregate compute power comparable to a system near the top of the Top 500 worldwide (9 petaflops, roughly #25 on the June 2019 list).

● Roughly 90% of equipment to be decommissioned was or would be disposed of in a way that ends its scientific life, rather than being used in any scientifically productive way.

● On average, the NSF funds only 30% of this equipment at respondent institutions, while the institutions themselves fund 60%. Both the institutions and government funders should want to get more value from their investments.

● Local limits -- for example on data center space or power, or the availability or cost of warranty support -- drive most decommissioning.

● There was universal interest in extending the productive lifetime of this equipment.

Respondents identified a number of concerns about extended operation, primarily around its cost and environmental impact. This survey documents the need for good alternatives to decommissioning and characterizes requirements for good options:

● Continue to provide computing services that advance scientific inquiry,

● Meet the real financial constraints around keeping the SCE operational,

● Have minimal negative environmental impact (i.e. they do not create e-waste or increase carbon emissions).

The survey confirms the value of exploring new mechanisms for extending the productive lifetime of scientific computing equipment.

Original Document

The original document is available in PDF (uploaded 8 July, 2020 by Andrew Chien).

Additional Document Formats

The document is also available in PDF (uploaded 29 July, 2020 by Andrew Chien).

NOTE: The author warrants that these additional documents are identical with the originial to the extent permitted by the translation between the various formats. However, the webmaster has made no effort to verify this claim. If the authenticity of the document is an issue, please always refer to the "Original document." If you find significant alterations, please report to webmaster@cs.uchicago.edu.