Elements: Kingfisher: Storage management for data federations

Storage space is a critical – and finite – resource for data-intensive science. Within dHTC environments like the Open Science Data Federation (OSDF), systems often react poorly when free storage runs out. The solution is explicit management of storage space as a resource, beyond traditional methods (quotas).This project aims to develop a Framework for distributed Storage Space Management (dSSM) based on the following four principles:All usage of storage throughout a distributed system is explicitly allocatedWorkloads explicitly state their storage requirementsAllocated storage space has a size limit, reclamation policies, access control rules, and an ownerStorage space allocations are recursive; an owner of a space can partition it into sub- spaces and assign allocations to others