300 likes | 374 Views
Fedora Commons Overview and Future Plans Sandy Payette, Executive Director Cornell University Library Metadata Working Group June 13, 2008. 501(c)(3) Non-profit. A Brief History…. 2005. 2012+. 2001. 2007. Community, Partners, Donors. Moore. Mellon. Mellon. Major Funders.
E N D
Fedora Commons Overview and Future Plans Sandy Payette, Executive Director Cornell University Library Metadata Working Group June 13, 2008
501(c)(3) Non-profit A Brief History… 2005 2012+ 2001 2007 Community, Partners, Donors Moore Mellon Mellon Major Funders Mellon
Use Cases We Support… Scholarly Communication Data Curation and Linking Science Humanities Semantic Knowledge Spaces Preservation and Archiving R.D.W. Connor
Users of Fedora Commons open source softwarehttp://fedora.info/wiki/index.php/Fedora_Commons_Community_Registry • National Libraries and Archives • University Libraries and Archives • Research Institutions • Museums and Culture Organizations • Semantic and Virtual Digital Libraries • National Repository Consortia • Publishing and Professional Societies • Medical Centers and Libraries • Government Agencies
How? Web 2.0 Your Applications Scholarly and e-Research Other DigitalLibraries Archive Asset Management Fedora Services and Middleware Layer OAI FedoraServices Manage API Access API Registry Search RDF Query Search Store Ingest FedoraRepository Ingest Manage Access Etc. Validate Policy ServiceBind Fast Disk Honeycomb Other Storage Layer Registry RDF Index
Working Together with Target Communities… FC – 2008 Community Solution Councils openaccess publication datacuration e-research preserve/archive
Current Collaborative Engagements TALKING data curation open access PLANNING DOING DRSG EOL PLoS JHU ORE e-research IA FIZ archiving Sun NSDL Hull UVA BHL Rutgers Oxford Adore DSpace BNF Portico Aquifer CUL preservation Kuali institutional repo Sakai
Fedora Commons Open Source Projects • Fedora Repository Project • The original Fedora Project • Fedora Middleware Project • Service integration and enterprise-orientation for repositories • Akubra Storage Project • New storage plug-in architecture; transactional file system; • Topaz Project • Incubated project; persistence service based on semantic triplestore • MulgaraTriplestore Project • Independent, but FC supports developer; highly scalable triplestore Fedora Commons open source projects can be integrated together.
Fedora Repository – Key Features • Digital Object Model • Aggregate content “datastreams” in an object… any type of content • Intermix both local content and external content • Relationships among digital objects (via RDF) • Register “content models” for known object patterns • Repository Service • Modular • Web service interfaces (REST/SOAP) • Versioning • Dynamic service binding based on object content model types • File-centric (all essential characteristics in XML files) • RDF-based indexing (semantic triplestore index with query) • Security with pluggable authentication and XACML policies • Journaling (replay all events to create replicas of repository)
BasicBuilding Block: Fedora Digital Object Model (relationships and properties) Persistent ID DC RELS-EXT Reserved Datastreams Audit Trail Policy Datastream 1 Datastream 2 Datastreams (any type, any number) Datastream n
Forming Collections… with relationships isMemberOfCollection isMemberOfCollection PID1 PID5 PID3 PID2 Query isMemberOfCollection “Smiley Stuff “ Collection Object
Digital Objects… information network scholarlyobject hasPart annotates hasPart providesContext PID5 PID4 PID 1 PID3 PID2 Amazon Web Service library content External commercial web content
Fedora Repository Service For creating, managing, providing access to various and sundry objects Fedora APIs (SOAP and REST) Manage API Access API Registry Search RDF Query Ingest Manage Access Validate Policy CMABind Store Registry RDF Index Triplestore File system (Objects) RDBMS (Registry)
Roadmap – Active Themes • Ease of use • Re-use and interoperability • Web orientation • Improved modularity • Pluggable storage • Workflow engine • Messaging • RDF and Semantic technologies • Content models • Solution bundles (including end-user applications)
2008 Notable Themes • Solution bundle – out of box application for repositories • “IR2” (Institutional Repository Application “Take 2”) • Institutional repository moving towards enabling e-research • Turn-key out of box user experience • Means of achieving (2 strategies under consideration): • FC and community developers build a brand new application • Fedora + DSpace collaboration • “Easy fit” repositories behind the collaborative web • Motivation: enable seamless integration with e-research applications (e.g., Zotero) and enable re-use and sharing of repository content • Object Reuse and Exchange (ORE) libraries • Lightweight interfaces on Fedora repository • Lightweight web protocol on Fedora repository (e.g., Atom, SWORD)
Better integration with web and workflows Atom (2008) OAI-ORE (2008) new formats Fedora APIs (SOAP and REST) Manage API Access API Registry Search RDF Query Ingest Manage Access Additional APIs: SWORD (2008) APP (full) ? JSR 170 ? Other ? Validate Policy CMABind Store Registry RDF Index Triplestore File system (Objects) RDBMS (Registry)
Replication to Archival Storage • Replica repositories, each with different underlying storage system; useful for failover, redundancy, archiving • Now: Fedora Journaling http://fedora.info/download/2.2.1/userdocs/server/journal/index.html • Future: Journal Event Messaging via Fedora JMS FollowerRepository LeaderRepository Journal Event Log API events API events Sun Honeycomb Can configure multiple “followers”
Better Service Integration First, we are providing simple messaging (via ActiveMQ in Fedora 3.0) services listen and consumeevents or other messages GSearch OAI Simple JMS Fedora Repository Service Ingest repository publishes events More… Next, lightweight integration with workflow engine(s); orchestration
The Model-Driven Repository: Content Models and Service Models Persistent ID hasModel DC DC DC RELS-EXT Persistent ID RELS-EXT RELS-EXT Audit Trail Model Type Def Service Def Policy hasServiceDef Image Preserve Metadata Custom Metadata Persistent ID Fedora Repository handles dynamic service binding at runtime
Runtime : dynamic services available on all objects that conform to the content model Persistent ID DC RELS-EXT Get-Thumbnail-Size-Image Get-Web-Size-Image Get-High-Resolution-Image Audit Trail Policy Image Preserve Metadata Custom Metadata
Akubra Project … backend storage interoperability Manage API Access API Registry Search RDF Query Fedora Repository Service Akubra Store Plug-in 1 Plug-in 2 Plug-in 3 File system (Objects) RDBMS (registry) New pluggable storage framework Triplestore (Mulgara/MPTStore) Sun Honeycomb Transactional File System