Final half day of Access 2014.
The last stretch.
## RDF and Discovery in the Real World(cat)
- Karen Coombs, Senior Product Analyst, WorldShare Platform
The Web of Data: Things, Not Strings
Way for search engine to be the most relevant e.g. May 2012: Google Knowledge Graph provides more knowledge in search results.
Traditionally in bibliographic description have title, format, author, but nothing about the author e.g. created, influence, appears in
Need to make use of Linked Data , and to share data. Means to publish structured data, describing things, and relationships between things. You need to identify things, and make them linkable.
Used identifier of domain with namespace and number. The ID .
RDF
RDF is a standard model for data interchange on the Web, consisting of triples (subject, predicate, object). Subject is worldcat link, predicate is schema.org, object is viaf link. Object can be the subject in another statement.
RDFa = RDF as attributes embedded in HTML
JSON-LD = bridging the gap between Linked data and traditional API JSON
Vocabularies
Utility e.g. RDF = RDF Concept Vocaublary, RDFS Resource Description Framework Schema
descriptive e.g. DC Terms
Bridge using bibliographic data and linked data e.g. dbpedia
Results
Can show related data that is not recorded in bibliographic data using WorldCat Discovery API provides data for application, RDF library see OCLC-Developer-Network github account
## Under the Hood with OpenStack
* Steve Marks, University of Toronto
* Amaz Taufique, ScholarsPortal
Ontario Digital Research Cloud (ODRC) = shared digital resources for digital collections, archival resources, research data.
Goals
- lower cost
- highly scaleable
- replicated
- open technologies and standards
- integrated (with gtanet, Orion (sp?))
- hosted in Canada
- secure
Needed high capacity, low cost, online storage approx $300/TB so proposed 1.2PB online storage for OCUL schools
Storage hardware: high density desk storage services (DSS), Each drawer 48TB (12 x 4TB). Swift doesn’t like RAID, but need controller with remote management. DSS capacity: 48-432TB. Lots of infrastructure purchases for proxy, authentication, processing, etc.
OpenStack Object Storage (Swift)
* evenly distributed throughout system
* standard hardware, RAID not required
* account/container/object structure (not file system, no nesting) plus replication (N copies of everything)
* Proxy Server inherently knows where the object will reside through an algorithm (and will succeed it reports writing at least 2 copies)
* use API to request object through proxy server, first one to respond returns object
* needed to monitor amount of traffic from replication, and in case of failure (hand off)
GTAnet Pilot
* understand how to design and implement effective network topology
* Ryerson, York, UofT
* model and record network traffic between four nodes (fourth to see how handoff would work)
* proxy hit 1 GB limit very quickly
* analysis on network, traffic, setup, ongoing support, etc.
* considerations: bandwidth? jumbo frames? extend VLANs across network? low ongoing, setup cost? have ORION POP on site?
Orion created virtual, private paths.
Implementation
- can use command line tools
- Swift Browser: GUI for more user friendly approach for basic functionality
- Cloudfuse – layer that mounts Swift objects in file storage on Linux box
- Fedora repository e.g. Islandora to work on cloud
- docker-wordpress configured to work with Swift
- SwiftFS – with node filesystem on top by parsing delimiters
- OwnCloud – exposes dropbox-esque functionality to sync local files, some editing, exposed on WebDAP
- Archivematica Storage Service to integrate Swift
- Newspapers collection for OurOntario.ca
Next Steps
- deplay all the hardware
*beta - develop end user tools
- repository integration
- Hadoop cluster and text mining
## Panel: Have you Tried Turning it off and Back On Again? Rebooting Library Technology Conversations
* Gillian Byrne, Associate Chief Librarian, Ryerson University
* Andrew McAlorum, Head, Digital Initiatives, University of Toronto
* Tim Ribaric, Brock UniversityGraham Stewart, University of Toronto
* Steve Marks, University of Toronto
* Graham Stewart, University of Toronto
Traditional IT Functions e.g. payroll, accounting, ILS – compliment existing models, don’t influence planning services. Seen as external, neutral, not involved in decision making, ripe for outsource, cloud
Traditional IT concerns:
- stability
- security
- reliability
- infrastructure management
- ROI
Seen as always seeing reasons to say no, the road blockers. Negative stereotypes of themselves and users.
Web Scale IT = entirely new services, business models, which means role is to lead innovation, drive decision, define services
In Libraries = engage faculty & students, integrated services, collaborative online spaces, etc.
What does a Web Scale IT team look like? librarians, developers, sysadmins, culture of empowerment, blameless post-mortems, systems and services must be humane
Human Systems:
systems that enhance the safety, content, knowledge and freedom of both the authors and users of the system.
Admin Decision Making about IT
If admin can understand it, then it seems easy. May be tendency to ignore, over reliance on peers and vendors. That everyone else has it.
Decision to go with LibGuides based on a content decision. Not enough interaction between liaison librarians and IT.
Many reasons IT decisions are made. Good IT doesn’t always win. Admin needs to be transparent about their decisions.
Values
What things cost: everyone has a price. Open source software doesn’t have a price tag, ergo it is not given value in the organization.
No in house CMS is a suitable replacement for LibGuides, because LibGuides.
Sharing (or library leviathan that will swallow us into the pit of despair) e.g. using LMS plugin for EDS, which became a hosted solution instead of running it locally
Innovate and create stuff, but suddenly it gets gobbled up and being sold back out.
Do your institution’s IT values match the core Library values your institution stands for? Do you advocate for Open Source as much as you advocate for Open Access?
It is very difficult for me to consider putting money into a thing that disappears outright the day after your contract expires. Rather put resources into building a thing that you can keep and control.
Approaches to Collaboration
Need for increased conversation. How do we bridge this gap? communication (shared information) collaboration (shared resources, stakes)
Communication: How do we facilitate meaningful conversation across: institutions? departments? functional areas?
Collaboration: How can we promote shared stakes in light of different institutional priorities? difference incentive structures?
Approaches to Projects for Better Communication
Agile: customer collaboration and respond to change with working software focusing on individuals and interactions. Focus on relationships and collaborate with users.
Scrum: What’s the vision? take functionality wishlist, chunk into 2-4 weeks sprints, daily meeting to do check-ins, produce
ITIL: you don’t have to do everything in ITIL. Many things in ITIL that will help. Take the parts that you need.
* Issue Tracking
* develop service catalogue (what services does your unit offers? how are they defined?)
* change management (vet and approve changes and new IT projects, proposals in writing following template, changes assessed for prioritization, technical feasibility, strategic alignment, members representative of all org. levels, exercise in risk management)
* service level agreements.
* continual service improvement with tight feedback loop
It’s about the people, not the technology. IT is a service profession.
Break
May or may not be completely awake still.
