Research Storage Funding Model
Research Data Storage
As of July 1, 2017, Harvard Medical School is operating its Research Data Storage service as a Specialized Service Facility (core). This cost recovery model, combined with generous subsidies for On-Quad faculty, allows HMS to continue offering state-of-the-art, world-class systems for research data storage to our user community.
HMS Information Technology provides HMS and affiliate researchers with a spectrum of options to store research data. Under the Specialized Service Facility model, users will be charged for their use of certain components of those services at the actual costs of operating those services. The School encourages the prudent use of resources through the continued funding of long term storage, temporary storage areas (scratch), and research data management consulting services, allowing us to offer those components of the service for free. Additionally, the School provides a monthly Storage Credit to On-Quad principal investigators to offset the costs of data storage, sufficient to cover the needs of the majority of labs.
Why is research data storage operated as a cost-recovering service core?
Harvard Medical School has for many years provided advanced computational resources to both On-Quad and affiliate faculty at no direct cost to researchers. Historically, the rate of data stored at HMS has doubled annually; it is therefore no longer sustainable for the School to wholly fund the demand for storage. Over the past several years HMS has made significant investments in our data storage and computational systems to better serve the needs of our researchers. Since 2012 HMS has invested more than $25M to grow the capacity for storing scientific data fifteen fold. Reflecting both the value of those services as well as trends in research, the demand for these resources continues to grow. Due to the exponential growth in large data sets, many other research intensive peer institutions (including Broad, Cornell, UPenn) have had to adopt similar funding mechanisms as it becomes increasingly difficult to keep pace with escalating costs associated with storing large data sets.
Who will this impact?
Any researcher storing data on systems operated by HMS’s Information Technology organization. However, the majority of HMS On-Quad faculty has only modest storage needs and will see no financial impact due to subsidies offered by the School. We are partnering with PIs whose research generates the largest amount of storage with a focus on better organizing and managing data in order to lessen the future financial impact.
What is the cost per TB of data?
The cost of storage depends on the specific technology used to hold the data. The rate per TB is based on an assessment of the costs of hardware, software, operating expenses, and personnel. For users not on quad, the rate includes a facilities charge. To calculate the monthly costs per user, usage is measured daily and averaged over the course of a month. For a description of research storage options and associated fees, visit Research Data Storage Services. On-Quad users are defined as HMS faculty members that have a quad appointment and are submitting grants and recovering IDC through HMS. Research service cores at HMS are eligible for On-Quad rates and subsidies.
How do I know how much storage my lab is using today?
A monthly storage usage report for each lab will be distributed, which exhibits the monthly average total storage for all users mapped to the PI. The mapping of users to PIs for the purpose of usage reporting is a new practice, and we anticipate modification requests over the first several months as the information is reviewed and refined.
How does the School subsidize the service?
On-Quad principal investigators receive a monthly Storage Credit of $500 to offset the cost of storage for their labs. This is equivalent to 20 TB of data on the Isilon Research Storage systems.
When will the new model become effective?
The new model will begin July 2017.
How will the School support the transition to the new cost structure?
The HMS Office of Finance has approved a multi-year transition period for On-Quad faculty to allow for full acclimation to the new model.
To ease the impact on faculty the School will subsidize storage costs for On-Quad faculty for three years according to the following schedule.
- Subsidy Year 1: July 2017-June 2018 - 75%
- Subsidy Year 2: July 2018-June 2019 - 50%
- Subsidy Year 3: July 2019-June 2020 - 25%
- Full cost recovery begins Fiscal Year 2021
How should PIs budget for storage on new grants?
For On-Quad faculty it will only be necessary to include storage costs in grants if the new demand will bring the costs above the Storage Credit offered by the School each month. If that is the case, all new grant proposals should include funding for storage. Please consult with your department finance managers and/or SPA for budgeting storage expenses to future grants and the appropriate language to include in your application.
What action can faculty take now to avoid or minimize charges?
PIs should review the Monthly Storage Usage reports for accuracy and completeness. The reports will be used to maintain transparency of the total storage consumed by each PI and his or her workgroup. HMS Research Computing will offer consultation related to data management for PIs interested in understanding how they can avoid or reduce future charges. For more information please visit Research Consulting and Long Term Storage Service.
How frequently will the cost model be reviewed?
The cost per terabyte of storage and the size of the HMS funded Storage Credit to PIs will be reviewed annually by the HMS Office of Finance. IT will continue to ensure we offer the most effective solutions for on premise storage and implement competitive cloud storage offerings moving forward. Future improvements in pricing will be directly passed on to users.
Who can I contact with further questions?
Send all questions and feedback to HMS_Research_Storage@hms.harvard.edu. IT will be triaging all questions to the appropriate individual to provide the support needed in a timely way.