preserving and sharing digital data greg colati, director, archives and special collections may 11,...
TRANSCRIPT
Preserving and Sharing Digital Data
Greg Colati, Director, Archives and Special Collections
May 11, 2012
2
The Problem with Digital Assets
Silos Separate technical systems, metadata, and
organizational schemes Difficult to repurpose or share
Duplication Same data in different systems out of
synchronization Duplicative but not redundant storage
Risk: No accurate way to understand and manage the
scope and volume of digital assets No way to insure long-term availability
How Do We Manage Digital Assets?
3 http://www.dcc.ac.uk/resources/curation-lifecycle-model
Digital curation in a managed environment
4
How Do We Share Digital Assets? EASY is not an attribute! Multiple audiences, casual to serious Multiple interests Multiple levels of sophistication Multiple needs
No ONE delivery system can provide for all potential uses
What We Didn’t Want:
An “integrated” solutionAnother information silo
8
6
What We Needed: A Flexible Digital Repository Service
Preservation-oriented, trustworthy digital repository Centrally managed Supports long term management of digital assets
Set of services that support current and future needs Open APIs for local development Management tools Presentation tools
7
Repository Attributes
1. Simple and complex digital objects and relationships
2. Schema agnostic: any metadata schema allowed
3. File format agnostic: any file types allowed4. Normalization for all metadata and supported
content file types5. Global, persistent identifiers6. Rules-based access and management7. Flexible presentation options
8
What We Needed…
Independent, self-describing objects existing in a managed system with No particular administrative workflow or end-user
application No particular catalog or organizational model
9
What We Chose
Flexible Extensible Digital Object Repository Architecture
(Fedora)
http://fedoracommons.org/
What Fedora is Not
NOT a relational databaseNOT a software applicationNOT an integrated solutionNOT an information silo
6
11
Fedora is…
A generic foundation upon which many kinds of applications can be created.
A conceptual framework about digital information and relationships
12
Repository-based Activities
Repository (Preservation) Layer
Management Layer
Presentation (Access) LayerAccess, Discovery, Authentication, Authorization
Curation Activities; Administrative, Descriptive, Technical, Rights, etc., Metadata Mgmt.
Information UniverseRe-use and exchange
Policies, Controls, Preservation Activities
Dig
ital
Obje
cts
Meta
data
Fedora
13
Current Project Calendar
Phase I: Establish and implement core functions (July, 2012)
Phase II: Advanced tools and services, develop storage and hardware requirements for production system (January, 2013)
Phase III: Implementation of v1.0 production release for UConn content. (July, 2013)
Extending the Model:A Statewide Repository Service
15
Repository Organization
16
Management Layer
Presentation (Access) Layer
Information UniverseRe-use and exchange D
igita
l O
bje
cts
Meta
data
Management: UConn
UConn
Partner Framework
Repository (Preservation) LayerPolicies, Controls, Preservation Activities UConn Repository
Partner Others…
External aggregators and direct links
Management: OthersManagement: Partner
17
Durable Content, Shareable Data
Data Center Data
UConn Data
Cultural Heritage
Data
Other Digital Data
State Pubs/archive
s
18
Pay for Services/Storage …Cost-recovery on an as-yet-to-be-determined model. Potential approaches include: Membership ($/year based on a formula) Subscription ($/GB/year) POSF (Pay Once Store Forever)
Other approaches or a combination of approaches are possible.
Sustainability Model (In Development)
19
Governance Model (In Development)…Volunteer your time Advisory Committee made up of participating
organizations Reviews development plan for infrastructure and
services Suggests funding opportunities Creates working groups for specific areas and tasks
SLAs created between the repository and UConn units document rights and responsibilities.
MOUs created between UConn and participating organizations document ownership, rights, and responsibilities.
20
Challenges Ahead Secure development funding for statewide
services Find partner institutions Current partners include:
UConn State Library State Data Center (MAGIC) Connecticut History Online (CHO)
Develop a sustainable governance and funding model
21
Interested?
Greg ColatiDirector, Archives and Special Collections
Trylon and Perisphere graphic from the Hagley Library and Museum http://www.hagley.org/hagley-wf-02-rgb.jpg