Three hidden costs—stranded capacity, operational overhead and data-movement risk—are inflating research storage bills, Floyd Christofferson said, especially where petabytes of unstructured data span on-premises and cloud systems.
Those costs build when file-system metadata stays tied to individual storage platforms, forcing teams to keep data on expensive high-performance systems or duplicate and move it in ways that disrupt users and burden IT staff.
shadow IT adds another layer of fragmentation: grant-driven hardware purchases and researcher-run storage outside central controls can weaken visibility, classification, security and compliance across institutions.
A shared global namespace, Christofferson argued, can unify data and metadata across heterogeneous storage, automate placement by policy and let AI and HPC workloads access files without repeated copying.
For budget-constrained R&D groups, the broader message is to manage data as a virtual lake across existing infrastructure rather than keep buying new siloed systems.