A near-online approach to archive systems

摘要

In the face of growing data archiving needs, a better approach than tape based storage is needed. We describe how a MAID based storage and data system can be designed specifically to ideally meet the many requirements of archival data. These include the need for large scale dense capacity, cost competitive with tape system, high object read throughput and write ingest rate, data retention, data immutability, compliance support, security, and content retrieval.