The PRISM Project at UC-San Diego and the Qualcomm Institute is the continuation of a long-term program to install high-bandwidth networking capabilities on the UCSD campus and surrounding area. By supporting advanced facilities to transfer large amounts of scientific data, the PRISM project extends the earlier OptIPuter and Quartzite development projects. Our team engaged with the rollout of the PRISM network by conducting interviews, observational sessions, and policy analyses with key stakeholders and laboratories poised to integrate with the new network, particularly with an eye toward obstacles and barriers to adoption. Our findings can be summarized as follow:
- New technologies impact current scientific practices and workflows, sometimes in disruptive ways. Scientists have reservations about enrolling into a new technology, despite the promise of benefits, if they perceive that the impact on their own scientific work will be too great.
- Scientific laboratories do not have clear incentives to be the experimental subjects of infrastructure research, and are hesitant to absorb the experimental risk of networking research into their own work. This is to say that enrolling into an experimental network technology internalizes the risk of failure into their own scientific data and computational resources.
- Campus network projects and upgrades embed assumptions about the campus being the relevant unit of observation and analysis for development, implementation, service provisioning, and evaluation. Our interviews revealed that campus may be less relevant to scientific laboratories, and the level of service is mismatched for their needs. Instead, scientists tend to focus on the hyper-local context (activities within their own lab) or the extra-campus world (communicating and transferring data to colleagues and institutions outside of the campus).
- Defining and measuring success criteria of visionary programs that include “creating new data imaginaries” (understood as inspiring scientists to develop, exchange, and use a broader range of scientific data and related tools to explore new inter- and trans-disciplinary questions) and “promoting new types of collaboration” requires both qualitative and quantitative approaches to evaluation. The unfolding of such high-level goals have uncertain temporalities and trajectories, and will likely defy standard metrics of success and regimes of evaluation.
PRISM Social Analysis Project Team
- Matthew Bietz
- Cory Knobel
- Geoffrey C. Bowker
- Deborah Forster (Qualcomm Institute)
- Steve Slota
- Andy Echenique
- Courtney Loder