IIIF Redesign Project Charter

I. Problem/Value Statement

Problem Statement

The Harvard Library maintains a technology infrastructure for delivering images and metadata to users in platforms like HOLLIS, HOLLIS Images, HOLLIS for Archival Discovery, Harvard Digital Collections and CURIOsity. This infrastructure is largely based on standards and technologies compatible with the International Image Interoperability Framework. The IIIF environment hosted by the Harvard Library is also used by the Harvard Art Museums, HUIT Academic Technology (with integration into Canvas) and HarvardX, and is considered critical university-wide infrastructure for teaching and research.

In recent months the service has experienced an increasing number of outages and performance issues, revealing the need for a major upgrade. As such, the Harvard Library intends to modernize its IIIF-powered digital image delivery infrastructure to take full advantage of the latest technologies and specifications, and deliver a best-in-class infrastructure to Harvard.

Business Value

As we anticipate an increase in the use of our digital collections in the context of remote research and teaching, it is critical that our digital image delivery infrastructure is performant, resilient, and scalable. To support a world-class program of teaching and research, it is critical that images are delivered to patrons seamlessly, and the service is available with minimal to no service disruptions. We also expect to see innovation in the use of digital content in research and teaching, including a desire for more interactivity and engagement with content. This work will position Harvard to support new online services (such as controlled digital lending), and innovative uses of our digital content, such as annotation creation or applications of machine learning and artificial intelligence.

II. Vision and Approach

The redesign of the Harvard Library digital image delivery infrastructure will be addressed in two phases: Discovery and Implementation. The Discovery phase will include consultation with  Digirati, a UK firm with extensive experience in building similar environments for large cultural heritage organizations.   The Implementation phase will be led by LTS in collaboration with other Harvard IIIF partners.

The objectives of the Discovery Phase are:

1. To conduct a high level Harvard-wide review of IIIF services and tools, the IIIF architecture, the technologies that have been implemented, and how IIIF services are connected, related, or dependent across Harvard;

2. to identify, document and understand the use cases and needs across Harvard departments that could best be served by IIIF services and tools; 

3. to assess the extent to which Harvard’s current IIIF infrastructure is meeting Harvard’s needs, and identify gaps, performance issues, or sources of instability or risk;

4. to develop a plan to modernize and rationalize IIIF services across Harvard; and

5. to evaluate Digirati’s Digital Library Cloud Service (DLCS) as an option for Harvard’s IIIF architecture.

The primary objective of the Implementation Phase is:

To develop a performant, resilient, and scalable multi tenant infrastructure for IIIF image delivery at Harvard. The implementation will involve three sub-projects:

• IIIF Asset Delivery
  • Objective: Develop a scalable, stable performant and multi tenant solution for image delivery;
    
• IIIF Services
  • Objective: Develop a scalable, stable performant and multi tenant presentation service layer for IIIF assets;
    
• IIIF Viewer / UI
  • Objective: Launch a pilot Mirador 3 installation to support scan/deliver and controlled digital lending workflows;
    

III. In Scope/Out of Scope

In Scope

• Re-evaluate the current image server platform (Loris vs IIP vs Cantaloupe) for features, performance, stability, and maintainability

• Analyze the mechanics of relocating delivery files (JP2s) to a new storage platform

• Plan for a multi-tenant IIIF image infrastructure for use by departments across Harvard

• Plan for decoupling IIIF and delivery services from the Digital Repository Service

• Plan to utilize Mirador 3 for the Harvard Viewer, with minimal customizations (use Mirador themes and plugins wherever available and necessary)

• Explore implementing an oEmbed service for easy embedding in any web environment

• Simplify authentication approach to rely on existing Harvard IAM services and open standards

• Plan to upgrade the Presentation API service.

• Plan for support of latest versions of IIIF Image, Presentation and Authentication APIs (with backward compatibility where appropriate). 

• Define performance metrics to be used for acceptance criteria in the implementation phase of this project

Out of Scope

• Architecting and implementing support for streaming audio and video services

• Architecting and implementing an annotation storage and management service

IV. Deliverables/Work Products

Discovery Phase

• Documentation on items in scope for discovery phase as listed in section III, including works plans, risk assessments, use case summaries and user stories

• Documentation, licensing, training, and other knowledge sharing  etc. to allow Harvard to maintain the solution 

• Implemented DLCS instance with a separate IIIF stack for experimentation with integration scenarios

Implementation Phase

• Re-evaluate and the current image server platform (Loris vs IIP vs Cantaloupe) for features, performance, stability, and maintainability and implement a performant, resilient image server
  

• Relocate delivery files (JP2s) to a new storage platform

• Implement multi-tenant IIIF image infrastructure for use by departments across Harvard

• Decouple IIIF and delivery services from the Digital Repository Service

• Implement a pilot installation of Mirador 3 for the Harvard Viewer, to support scan/deliver and controlled digital landing workflows
  

• Implement an oEmbed service for easy embedding in any web environment

• Simplify authentication approach to rely on existing Harvard IAM services and open standards

• Upgrade the Presentation API service.

• Implement of latest versions of IIIF Image, Presentation and Authentication APIs (with backward compatibility where appropriate).
  

Definition of “Done”

The Discovery Phase is considered completed once:

• Stakeholders accept final documentation and implementation plans for IIIF infrastructure redesign for all items listed in the In Scope section (section III) of this document

• Features are prioritized for Fall semester delivery

• Success metrics for implementation phase are defined

Implementation Phase

• Criteria for success during the implementation phase will be defined during the discovery phase

V. Stakeholders/People

Project Sponsors

• Stuart Snydman, Managing Director of Library Technology Services, HUIT Library Technology Services & Harvard Library, Executive Sponsor

• Harvard IIIF/Mirador Working Group, including:

• • Suzanne Wones, Associate University Librarian for Digital Strategies and Innovations, Harvard Library

  • Claire DeMarco, Associate Director of Digital Strategies and Innovation, Harvard Library

  • Franziska Frey, Chief of Staff and Senior Advisor for University Library Strategy, Planning and Assessment, Harvard Library

  • Alan Wolf, Managing Director, Academic Technology Services, HUIT

  • Rashmi Singhal, Director of Arts & Humanities, Research Computing, HUIT

  • Jeff Steward, Director of Digital Infrastructure and Emerging Technologies, Harvard Art Museums

  • John Alper, Director of Technology, VPAL/HarvardX

What organizations, departments, or people will benefit from this work?

• Those who support services and technology related to image delivery, including staff from the Harvard Library, Harvard Art Museums, HUIT Academic Technology Services, HarvardX

• Harvard faculty and students who rely on digital content to support courses. 

• Scholars at Harvard and world-wide who use digital content for their research.

• Patrons and visitors to the Harvard Library and Museums who want to use and appreciate the world-class digital resources at Harvard.   

Who is funding the work?

Harvard Library is funding the supplemental resources assigned to this effort.

Who will accept the work?

• Stuart Snydman, Managing Director of Library Technology Services, HUIT Library Technology Services & Harvard Library, Executive Sponsor

• Harvard IIIF Advisory Group, including:

• • Suzanne Wones, Associate University Librarian for Digital Strategies and Innovations, Harvard Library

  • Claire DeMarco, Associate Director of Digital Strategies and Innovation, Harvard Library

  • Franziska Frey, Chief of Staff and Senior Advisor for University Library Strategy, Planning and Assessment, Harvard Library

  • Alan Wolf, Managing Director, Academic Technology Services, HUIT

  • Rashmi Singhal, Director of Arts & Humanities, Research Computing, HUIT

  • Jeff Steward, Director of Digital Infrastructure and Emerging Technologies, Harvard Art Museums

  • John Alper, Director of Technology, VPAL/HarvardX

Who is the project manager?

Vitaly Zakuta, Senior Systems Librarian, Library Technology Services

Who will be involved in doing the work? 

Discovery Phase:

Implementation Phase:

The implementation of the new IIIF design will be done in three sub-projects, each with its own team. The following teams will be working on the implementation: 

• IIIF Asset Delivery Team (see Implementation Teams page for more information)
  
• IIIF Services Team (see Implementation Teams page for more information)
  
• IIIF Viewer Team (see Implementation Teams page for more information)

VI. Schedule

VII. Other

Additional constraints, dependencies and risks will be added to this section as the discovery phase of this project proceeds. Some initial information is stated below.

Constraints

• Digirati is contracted through June 22, 2020 and may not be available to consult or participate in implementation beyond that date. 

Assumptions

• The goal is to deploy this updated infrastructure to support remote teaching and research in the Fall of 2020.  

Dependencies

• DRS Metadata Services enhancements to support IIIF manifests are a prerequisite for production implementation of the service.

Risks (description, plan, impact, owner)

• Risk: Tight timeline to have new production services delivered for Fall Quarter

• Remediation: During the discovery phase it will be decided which high priority items are in scope for delivery in Fall Quarter. Other times may be delayed to Winter 2021

• Risk: Digirati’s DLCS platform utilizes technologies that LTS has not supported. Implementing DLCS will require engineering and operations staff training, at the very least.

• Remediation: During the discovery phase LTS will engage on a pilot to implement a test installation of the DLCS platform to find out whether it can be supported by LTS staff and whether it can interoperate with Harvard infrastructure without heavy development commitment

• Risk: Disentangling content delivery from DRS will be complex with many dependencies. There is a risk that much of the Discovery Phase will be spent on learning about DRS, when a goal is to move away from the DRS context, and not on designing a new environment. 

• Remediation: During the discovery phase LTS and Digirati will keep to a high level overview of the existing Harvard infrastructure in order to facilitate quick progress on designing a new solution

Glossary

DLCS - Digital Library Cloud Service

DRS – Digital Repository Service

IIIF – International Image Interoperability Framework

Mirador – IIIF Viewer

VI. Schedule

VII. Other

Additional constraints, dependencies and risks will be added to this section as the discovery phase of this project proceeds. Some initial information is stated below.

Constraints

• Digirati is contracted through June 22, 2020 and may not be available to consult or participate in implementation beyond that date. 

Assumptions

• The goal is to deploy this updated infrastructure to support remote teaching and research in the Fall of 2020.  

Dependencies

• DRS Metadata Services enhancements to support IIIF manifests are a prerequisite for production implementation of the service.

Risks (description, plan, impact, owner)

• Risk: Tight timeline to have new production services delivered for Fall Quarter

• Remediation: During the discovery phase it will be decided which high priority items are in scope for delivery in Fall Quarter. Other times may be delayed to Winter 2021

• Risk: Digirati’s DLCS platform utilizes technologies that LTS has not supported. Implementing DLCS will require engineering and operations staff training, at the very least.

• Remediation: During the discovery phase LTS will engage on a pilot to implement a test installation of the DLCS platform to find out whether it can be supported by LTS staff and whether it can interoperate with Harvard infrastructure without heavy development commitment

• Risk: Disentangling content delivery from DRS will be complex with many dependencies. There is a risk that much of the Discovery Phase will be spent on learning about DRS, when a goal is to move away from the DRS context, and not on designing a new environment. 

• Remediation: During the discovery phase LTS and Digirati will keep to a high level overview of the existing Harvard infrastructure in order to facilitate quick progress on designing a new solution

Glossary

DLCS - Digital Library Cloud Service

DRS – Digital Repository Service

IIIF – International Image Interoperability Framework

Mirador – IIIF Viewer