A look at some of the commonest services of Duke Research Computing from a slightly different perspective. Our apologies to Harpers Index.

Duke Research Computing Index

Number of Libraries of Congress that could fit into Duke Research Computing storage devices: 201

Average number of Duke Compute Cluster users added daily: 1

Number of current users of the cluster, August 2019: 1,839

Number of research groups using the cluster: 233

Groups added, August 2018-July 2019: 41

Coffee consumed in individual “orientation and introduction chats” with new users of the cluster, August 2018-July 2019: 10 gallons & 1 quart

Average number of years of computing power used per day on the cluster, August 2019: 23.7

Average number of years used per day in 2015: 9.62

Number of scholarly publications supported by Duke Data Commons storage services, August 2018-July 2019: 2163

Monthly cost of 1 terabyte of storage in the Duke Data Commons, in Starbucks Coffee: 1 “Caramel Flan Latte” (Grande) & 1 “Freshly Brewed Coffee” (Grande)4

Percentage of research groups on the cluster from the School of Medicine: 35%

Percentage of data labelled “sensitive” that are stored on the Duke Compute Cluster: 0%

Percentage stored in the Protected Network & “PRDN”: 100%

Number of Research Toolkits’ “RAPID virtual machines” active, August 26, 2019: 231 machines for 749 users

Number of RAPID virtual machines created since January 2015: 1,389


1 Capacity of Library of Congress estimated from: Ashenfelder, M., 2011. Transferring “Libraries of Congress” of Data | The Signal[WWW Document]. URL https://blogs.loc.gov/thesignal/2011/07/transferring-libraries-of-congress-of-data/ (accessed 8.26.19); Raymond, M., 2009. How “Big” Is the Library of Congress? | Library of Congress Blog[WWW Document]. URLhttps://blogs.loc.gov/loc/2009/02/how-big-is-the-library-of-congress/ (accessed 8.26.19); Annual report of the Librarian of Congress for the fiscal year ending September 30, 2010, 2011. Library of Congress, Washington, DC. NB: the size of a digital Library of Congress is roundly argued about. We used Nicholas Taylor’s estimate from 2011 of 208 TB (Ashenfelder 2011). Duke’s research computing storage devices totaled about 4.2 PB in June 2019.

2 Starnes, S., Mejia, F., Milledge, T., DeLong, M.R., 2015. Use of the Duke Compute Cluster in 2015: data, visualization, and lessons. Poster presented at the Duke Research Computing Symposium, Durham, NC.

3 Number derived from two sources: S10-NIH-2019 bibliography, 2019. [WWW Document], URL https://www.ncbi.nlm.nih.gov/sites/myncbi/mark.delong.1/collections/58717811/public/ (accessed 8.26.19); Annual usage report to NIH, grant 1S10OD018164-01, 2019. Duke University, Durham, NC.

4 Starbucks Coffee Prices and Locations – Menu With Price[WWW Document], n.d. URL https://www.menuwithprice.com/menu/starbucks/ (accessed 8.26.19). Monthly cost of Duke Data Commons storage, $6.67; total cost of the two coffees, $6.65