ScholarWorks Update - September 2018
David focusing on ScholarWorks
Now that things have settled down on the ULMS project, David will be devoting the lion’s share of his time for the foreseeable future to the ScholarWorks project, focusing on interface and branding issues in Hyrax, providing support for OJS and related applications, some project management and communication tasks, and assisting Kevin wherever possible, including with ongoing DSpace support. This increase in staffing should help us move faster on various aspects of the project.
Hyrax implementation
Much of the last six months has focused on the pilot phase of our Hyrax implementation. The purpose of this phase of the project was to thoroughly test the various aspects of the DSpace-to-Hyrax migration and implementation process -- including the data migration scripts, Amazon Web Services (AWS) architecture, and Shibboleth authentication -- with a single pilot campus, San Marcos, in order to unearth and resolve any unexpected problems.
The pilot phase accomplished that goal by uncovering a major, unexpected issue with the load balancers in our AWS architecture, which prevented us from implementing Shibboleth authentication. After tracking down our Shibboleth problems to the load balancers, and several attempts at resolving it, we eventually concluded that we needed to bring in consultants to help rearchitect our AWS environment. Over the past few weeks, we’ve been working with consultants from our AWS provider, Infiniti, to reconfigure the Elastic Beanstalk (EBS) service within our AWS environment in order to solve the load balancer problem. We’ve set-up test servers to test new configurations, and have already begun working with Infiniti on a review of other aspects of our environment as well.
In the meantime, we’ve set-up San Marcos on a stand-alone server outside of our intended EBS architecture in order to continue testing further aspects of the migration and implementation. We now have Shibboleth authentication working for San Marcos, and they have begun reviewing their migrated data and testing various aspects of their Hyrax deployment. Our plan is to move San Marcos into production while we continue to workout our EBS issues, with the goal of moving them back into the intended architecture at a future date.
Unfortunately, this problem has delayed the project by four to six months, rendering many of our initial timelines moot. Once we complete the pilot, we’ll have a much more realistic sense of the timelines and costs involved, and can begin giving campuses a schedule for migrating their DSpace instance to Hyrax.
Hyrax 2.1 upgrade
In May, the Samvera community released Hyrax 2.1, which includes new features for collections, the implementation of a new universal viewer to better support images and multimedia, various UI/UX upgrades, and other changes. We updated our deployment codebase to version 2.1 in August after Hyrax released version 2.2. Our goal is track upstream changes to one minor version behind.
Migrations
DSpace to Hyrax
In addition to the migration work for San Marcos, we completed the migration for Long Beach’s McNair Scholars Program, one of the smaller current DSpace repositories, in July, and began migration work for Sacramento in August. For a collection the size of Sacramento’s (~4,200 items), the process can take upwards of three weeks.
CONTENTdm to Hyrax
In March, we began coordinating with San Jose and Fresno on possible options for migrating CONTENTdm to Hyrax. A Samvera interest group organized around CONTENTdm migrations is also underway. Much of the focus here is on the CDM Migration Tool developed by the University of Victoria. Further work is needed to extend that code into a full-blown migration process.
Islandora to Hyrax
In May, we began working with Moss Landing on migration strategies for their current Islandora instance. Similar to the work for CONTENTdm, the focus here will necessarily involve developing migration scripts and a process to move the item metadata and files into Hyrax. Our DSpace migration toolset seems the best candidate for that work given the similarities between DSpace and Islandora data. Moss Landing has been able to export their items and deposit them in one of our AWS buckets.
Symposia to Hyrax
We met with Long Beach in July to discuss migration efforts from their current IR system, Innovative’s Symposia, to Hyrax. Many of the same issues discussed above are relevant here as well. The planning for this is in its early stages.
Publishing
We currently host beta services of Open Journal Systems (OJS) and Scalar for campuses to explore. We’ve seen a marked uptick in interest in OJS and related services for conference proceedings and monograph publishing over the past couple of years. We’re currently in the process of setting up true test and production servers for OJS and will be re-evaluating the other services for eventual production deployment as well.