As data volumes continue to accelerate across scientific and research domains, infrastructure decisions are becoming inseparable from research outcomes. The ability to store, access, and preserve data at scale, without introducing friction for researchers, has become a defining challenge.

In this recorded session, leaders from Dana-Farber Cancer Institute discuss how the institution is addressing that challenge in practice. The conversation explores the evolution from an all-disk and cloud-reliant model toward a more balanced architecture that incorporates multiple storage tiers, including tape, while maintaining a unified and transparent user experience.

A recurring theme is abstraction: researchers interact with data as a consistent resource, while underlying systems handle migration, tiering, and long-term preservation. This separation enables infrastructure to evolve, whether due to cost pressures, hardware constraints, or new technologies, without disrupting scientific workflows.

The discussion also touches on less visible but critical considerations: the unpredictability of cloud economics, the importance of metadata in managing large datasets, and the long-term implications of data retention strategies in research environments where deletion is often not an option.

Rather than presenting a single solution, the webinar offers a grounded perspective on trade-offs, operational realities, and architectural flexibility. For organisations managing data-intensive workloads, it provides a useful lens on how infrastructure can adapt alongside both technological change and scientific ambition.