Perry R1
1Australian National University, Canberra, Australia
Biography:
Ryan Perry is the Director of the Australian Data Archive (ADA) at The Australian National University, where his work bridges applied social science and digital research infrastructure. With a background spanning survey methodology and social psychology, Dr. Perry’s core focus is on improving how sensitive social data is managed at scale. He manages the delivery of major national data platforms, and leads critical initiatives to advance data sovereignty, data governance, and metadata standards. His work aims to design repository infrastructure that ensures complex research data remains trustworthy, secure, and accessible to the broader research community.
Abstract:
To scale trustworthy digital preservation, repositories must move from policy-compliance to encoded trust. This presentation explores how the Australian Data Archive (ADA) implemented custom middleware to automate OAIS compliance.
The Open Archival Information System (OAIS) provides a foundational framework for digital preservation – defining what an archive must do, but not how to implement it. Historically, archives bridged this gap via written policies and manual staff compliance. However, relying on human intervention for high-volume operations creates scalability bottlenecks and policy drift – the divergence between documented governance and daily repository practice.
To operationalise OAIS, ADA developed the Archival Deposit and Preservation Tool (ADAPT). Functioning as orchestration middleware, ADAPT translates abstract trust standards into software constraints. It enforces a "preservation-by-design" architecture: programmatically separating Submission Information Packages (SIPs) from Archival Information Packages (AIPs), automating cryptographic fixity, and generating machine-actionable provenance logs.
Crucially, the goal of this infrastructure is not to replace archivists. While ADAPT effectively manages standard, high-volume data deposits, attempting to automate the long tail of exceptions is impractical. This talk will present strategies for implementing code-based trust and identifying where automation ends and human expertise takes over.
