 Hello, and welcome to our CNI Fall 2020 Project Briefing. My name is Alyssa Guzman, and I am the Digital Scholarship Librarian at the University of Texas Libraries. And I'm here with my colleague, Larry Yang, Principal Software Engineer. And we're going to be talking today about our project to containerize our digital exhibits for scalability and sustainment. So here at the University of Texas Libraries, we have very large collections we hold over 10 million volumes, as well as broad and deep primary source archival collections. So these are some examples from our architectural collections, early maps from the Americas, from our Benson Latin American collection, as well as our online PCL, Pericastonida Library Maps Collection. We have been digitizing content and sharing it online since the late 90s when we started digitizing our maps collection and creating simple HTML pages. Over the last 20 years or so, we've created many different legacy exhibits, each created using different technologies based on the time it was created. These exhibits were very difficult to migrate and maintain because they were all created in different systems and some not even created in a content management system at all. Some of the assets included metadata associated with them and some of them didn't. And we didn't have a comprehensive plan for sunsetting exhibits or sustaining them into the future. So in 2017, I became the Agile Product Owner for our website Refresh Project and I had to grapple with the legacy of how many exhibits we had in different CMSs. This is an example of one of those exhibits from our architectural collections. I took this screenshot from the wayback machine from 2011 and it looked largely like this in 2017 when I took over the Web Refresh Project. Since 2017, we have established a digital asset management system and we have, alongside that dams, we have a series of other repositories and our digitized content continues to grow. We have things that are digitized in-house at ET libraries. We have post-custodial materials as well as collections that have been contributed by ET scholars. And one of the things that we really valued from our original approaches to creating online exhibits were the varied approaches that each curator took to highlighting their content and we wanted to maintain that flexibility in framing the content and the purpose of the exhibit moving forward. In 2017, when we were beginning to establish the digital asset management system and we were undergoing the initial stages of our website redesign, we did need an intern solution for creating digital exhibits, ways to highlight small selections of content. So we decided to subscribe to omeka.net and we created what we call collection highlights which are small-scale digital exhibits of 10 to 15 items providing just a taste or a sampling of a larger collection that may or may not be digitized in full. This is an overview of part of our repository landscape. So we have our digital asset management system and the associated user-facing collections portal developed in Blacklight. We have our Latin American digital initiative site which holds all of our post-custodial materials and then we have a variety of other repositories of content like the archive of the indigenous languages of Latin America, our collections and our Texas data repository as well as our institutional repository, Texas Colorworks. And we were looking at wanting to be able to draw content from all of these different places to include in our exhibits. So in 2020, we decided to launch Spotlight as an enterprise solution for digital exhibits customized to our needs and we migrated over all of our omeka.net collection highlight exhibits as well as some other exhibits that were created at the Benson Latin American Collection and then looking forward in 2020, we'll be working on migrating all of our legacy exhibits from the old website into Spotlight but some of those materials need to be moved into an asset management system before they can be migrated to Spotlight. You can view our Spotlight site at exhibits.lib.utexas.edu and I will give a demo of our Spotlight site at the end of this presentation after Larry talks about the infrastructure of the site but first stage you wanna talk a little bit about how we're thinking about exhibits moving forward. So each of our exhibits has an exhibit compact document and that's something that the exhibit curator creates. It includes a description of what is in the exhibit, the purpose of the exhibit, who the audience was and most importantly, an enduring point of contact for the exhibit and a backup point of contact if the primary person leaves the organization and we piloted this approach with our Omega.net sites and it was very, very helpful when we were migrating things over to Spotlight. We actually did have a retirement and we had to contact the backup person with some questions. And then moving forward in our documentation, we're looking at options for what happens when we choose to sunset an exhibit and recommendations for each exhibit curator about how to preserve their content in perpetuity. So we do recognize that the creation of these exhibits for many staff represents an active scholarship and it's important for them to be able to preserve a place where they can access their work in the future and link to it from a CB or an online portfolio. So we have some recommendations for how they can deposit their content into our institutional repository which is called Texas Scholar Works. And those options include turning the content of the exhibit into a PDF and depositing that or doing some web archiving and scraping the exhibit and depositing that into Scholar Works. And for people who are in temporary staff roles like graduate research assistantships, we're actually recommending that the exhibit be deposited into Texas Scholar Works at the point of publication so that in the future when that person leaves the organization, they still have a permanent link that they can put on their CB to their work. And with that, I will turn it over to Larry to talk about the infrastructure of Spotlight. I'm going to briefly share the technology behind our exhibit portal and the infrastructure that's needed to support that. About two years ago, UT Libraries made the determination that we're going to use containerizations as our main platform to host and build our applications. So there are two view of containerizations. First from the system administrator's perspective, system admin sees using containerization will allow better utilization of resources, run more applications with less resources. Another one is from the developer's perspective, developer believe that they will have more precise control over how the applications build and run. The official answer from Docker is that containerization allows developer to package up all its dependencies and libraries that's needed into one single container and the containers can then be run pretty much anywhere whether it's locally on local infrastructure or cloud infrastructure like AWS will Google and the container should perform and run exactly the same. And we see some major benefits doing running applications with containerizations. First one kind of mentions that we're getting efficiency with less resources. Traditionally, application runs on servers and then comes along, virtual machine is developed, applications then run on virtual machine and multiple virtual machines run on the server. Over time, we see many of virtual machine is not fully being utilized because some application isn't as heavily used as expected. So VM are sitting idle and containerization comes along and try to solve that problem. So multiple container can run on a single VM and multiple VM can still run on a server. And the hope is that that would then better utilize the resource. In our case, we see the value of that. And another one is from more from the developer's perspective that same piece of code that developer had written would perform the same way anywhere where the container is being run. We often hear that when something goes wrong with that application, developer will respond that the application runs fine on his or her machines but the code didn't work on the server. With containerization and the standard of the configuration and the control inside the container, this piece of code will perform the same regardless of where it's being run. So the same error that would occur on the server would occur locally on developer's machine and it will be the same for his machine, her machine and my machine. And with all this, we see the benefit of able to deliver requests and turnaround requests much faster to our clients and stakeholders and increase a speed of delivery for all our processes. And that kind of lead into the next point of IT manager often push for everyone and everything to do it better and faster. And with containerization, we see that we can achieve that. Here's a basic overview are our existing ego systems. So we still have some VMs on the right-hand side in the box of yellow and the left-hand side is Docker containers that we knew we built. So DMs is our digital asset management system and RGIS is our geo-coded assets. Both of these systems allow user to ingest data and manage metadata. These are still some time ago and we're looking into transition that into container as well, but for now they're still in the VM and those are our source system to maintain our assets and metadata. And then both of those process then goes through a custom publishing process that's coded by our developers. These are our Python script running inside containers that will take the source data and push to our triple IF server in our solar cloud. If you don't know, not familiar with triple IF, triple IF is a standard protocol that help display media contents and that could be images, videos and audio is a pretty powerful protocol. And then we use solar for our static search and that's where our metadata goes. And with triple IF and solar cloud that then power our front-end portal. And in this case, we have collections portal running BlackLine, geo-data portal running geo-BlackLine and spotlight runs our exhibit portal. And this setup allowed us a very robust way of managing these applications. Should there be more low than anticipated, we could bring up any one of these circle have a redundant copy to handle the low and technology behind it will be able to low balance all those traffics. So if we're expecting a heavy hit on our exhibit portal, we could bring a multiple copy of that and to facilitate the traffic that's needed. And all the communication in between are all through API and that allow us to be able to swap out any of the pieces as needed in the future. More detail into our exhibit portal. The container itself is a BlackLine running spotlight plugin and these are built and developed by Stanford University. And then we added Mirador viewer as our triple IF viewer. And then we include MySQL plugin to talk to our database and then also SAML code for our single sign-on for authentications. And lastly, it will include any sorts of look and feel customizations from our stakeholders. And all that together is the single container that we built for our exhibit portal. And it itself had three external dependency. It uses a solar service for static search. We have external storage to manage all our assets and host our assets. And then we have the MySQL database to kind of key track information related to the exhibit. I mentioned containers running on dockers. It's a technology and it has a whole science behind that. And the main part is the Docker file. This is the heart and soul of the container. This is where developer can go in and customize every single line of code on how the container is built and run. And all that, we are trying to create a self-sufficient environment for our business users by standardizing the publishing workflow, having triple IF as the publishing targets that allow the publishing to be done the same way regardless of what kind of asset it is. And then we built everything around our solar and our triple IF and that allow us to stand up the front end portals independently from these assets. And when we bring up exhibits using our spotlight containers and that fit nicely into the existing ecosystem by utilizing all existing assets. So all that technologies and processes and workflow have the single goal of empower our librarians and curator to, so they are able to create and publish exhibits and collections without any sorts of IT interventions. We launched the exhibit portal at the annual August. Now a few months later, there are over a dozen exhibit up and running in that portal. Now back to Alyssa for a demo of our exhibit portal. All right, thank you, Larry. Now I'm back to give a brief demo of our spotlight site. Okay, so this is the homepage of our exhibit site. Again, you can view it at exhibits.lib.utexas.edu. We have a number of exhibits that were migrated over from our Omeka.net subscription as well as some brand new exhibits that were created directly in spotlight. We've decided to organize our exhibits by the purpose or type of exhibit. Our collection highlights are the small scale sampleings that were brought over from Omeka. And we anticipate continuing this approach to creating small scale introductions to different aspects of our collections. It was very successful in allowing us to give a broad overview of different types of materials without overwhelming our digitization queue. And then we also have exhibitions which are broader explorations of a specific topic. And in the future, we'll add another category for teaching collections, which will be materials that have been collected together specifically to accompany a certain course. Okay, when you click into an exhibit, there is a homepage for each exhibit which includes content. And then you can add different items to the page using these widgets as they're called in spotlight. And then you can create subpages for each aspect of your content. If you click onto a particular item, you see an item detail page with the Mirador 3 viewer and all of the metadata that was imported from our collections portal via AAIF. Now, not all of our items were imported in the spotlight via AAIF, but this particular one was. There's also a browse feature with some basic sorting and fastening options. And this is sorting through just items in this particular exhibit. A number of our exhibits have also been translated into Spanish. So the interface itself has been translated into Spanish and then Spotlight has an internationalization feature that allows the curator to go in and translate the different metadata fields and of course all of the content of the exhibit into Spanish. Right now we've only configured Spanish for our installation of Spotlight, but we plan to add support for Portuguese very soon. A Spanish and Portuguese are the two primary languages that our users in Latin America speak. Now I'm going to navigate over to the administrator dashboard and show you what that looks like. So this is an unpublished demo exhibit because I would like to show you what it looks like to import an item from our collections portal. So this is the curator dashboard of this particular exhibit. I'm going to navigate to the items tab. So this exhibit has 30 different items in it. We decided to maintain the ability to add items in different ways in our Spotlight site. So you can manually upload something or you can upload a CSV of materials. This is particularly important for collection materials where the entire work is under copyright, but we're choosing to share maybe the cover or a single interior page under fair use and then of course AAAF are things that have been published from our digital asset management system to our Blacklight site. So this is an example of one item from our collections portal, some sheet music. If I scroll down in the metadata, I can grab this AAAF manifest and copy it over. And heading back over to Spotlight, I can just paste it here and add the item to our Spotlight site. And it appears here. We also added the ability to delete items to our site because it was important for us to be able to remove something if something went wrong with the import from the CSV, but we don't have the option turned on to edit any metadata within Spotlight. And now the item has been removed from the Spotlight site. And lastly, I just wanted to put up our contact information. If you have any further questions about our installation of Spotlight or our approach to creating digital exhibits. Thank you again so much for watching our CNI project briefing and we look forward to hearing from you.