 ahead and get started. Thanks for joining us today. I'm Cliff Lynch. I'm the director of the Coalition for Networked Information, and I want to welcome you to this project briefing, which is part of week three of the CNI Fall 2020 virtual member meeting. Week three, just to refresh your memory, focuses on infrastructure, technology, and standards. And certainly this presentation fits very nicely into that set of themes. I want to note that as well as the live project briefings for week three, we've also released quite a number of prerecorded videos, and please avail yourself of those. As far as the session today, it is being recorded and will subsequently be publicly available. There is closed captioning available. Please turn that on if it's helpful to you. There is chat running, and please feel free to use that as we go through the session. There is a Q&A tool at the bottom of your screen, and you can use that to queue up questions as they occur to you at any point during the presentation. We will deal with all the questions at the end of the presentations where Diane Goldenberg-Hart from CNI will moderate a Q&A session. So with that, let me introduce Kirsten Leonard, the Executive Director at Palney, and Jill Morris, the Executive Director at Palsy. And if you can keep all of those straight and say them real fast in the right order, congratulations. They are going to talk with us about their progress on building out Haiku, an IMLS-funded effort for cross-consource field community infrastructure. Basically, this is a multi-tenant thing that can be branded as part of individual institutions. And it has, I think, at least potentially really attractive scaling and cost implications, especially for smaller institutions. So I'm delighted to have this update, and I'm gonna turn it over to Kirsten, who will start the presentation. My thanks to both Jill and Kirsten for joining us today. Over to you. Thank you very much, Cliff. So welcome and thank you for coming. As Cliff said, the project is building Haiku for consortia, really largely for any groups and partnerships of any kind. We aim to optimize the sharing of the infrastructure and expertise to run an institutional repository not only across our institutions, but also across groups. So we're excited about this project and the agency that it brings to our libraries to meet institutional repository needs at this challenging time. We also wanna share with you how working together in this way, even if you currently think that your institution is self-sufficient, will support a better outcome for all. As Cliff said, I'm Kristen Leonard, the executive director of PALNY, which is a private academic library network of Indiana. My partner in this project is Jill Morris, the executive director of PALNY, the Pennsylvania Academic Library Consortium. We're frequently known as the PALS in this project, which definitely, I think, aligns with our perception of ourselves in all of this. Today, we're gonna give you a very brief overview of our two consortia and what led us to the solution in working together. We'll give you an update on the status of our project and the development goals and future project plans. And we have a call out for continued collective action to claim our agency and development direction to leverage our joint expertise and to reduce costs. For the staffing for the Haiku for Consortia project, Jill and I are members of a larger team, including our project owner, Amanda Hurford, who is the PALNY scholarly communications director and Gretchen Guggen, who's Pal C's digital projects and communications manager. In addition, we have Nicholas Stanton-Werk, who is a PALNY coordinator. That's somebody that we've bought out some of their time from one of our institutions. He's from Anderson University, and he's supporting the project by serving as the liaison to our pilot participants. He does some testing and writing documentation, workflow, best practices. We're also joined by Notch 8, which is a software company out of San Diego, who is our lead developer. And we have several design, testing, and requirements teams with members from library staff across both of our consortia. We are supported, as Cliff said, by an IMLS grant running through the end of 2021. And our project is building not work that Stanford, DPLA, Endura space, and others have done to develop the formally titled Hydra out of the box, Hydra out of the box. We are focusing, in our instance, in leveraging the core capacity for an open source, multi-tenant IR solution that can be developed to easily scale. So collaboration is at the core of our project. Here you see the partners writ large. All of the members to this project have brought different skills, experiences, and viewpoints that strengthen the project. For example, Palsy brought significant IMLS grant knowledge and they participated in the Haiku Direct Pilot. So they had that experience about the tool. Palney brought significant vision, use cases, and usability knowledge as we've been struggling to find a solution for a long time. We've received significant support from others, including Jeremy Friesen from Notre Dame. And through participating in the Haiku Samvera community, we've been both able to surface the needs of a broad range of institutions that haven't participated in that space or been represented before. And we've also benefited from learning from this wider community. So a little bit more about our two consortia. So you see the pals there together side by side. Palsy on the left. We couldn't have more two different organizations, I think, in the way that we're laid out. We have different sizes and number of institutions. 71 for Palsy, just 24 for Palney. Ours are mostly private academic libraries. Whereas Palsy has a wide range there, as you can see ARLs, in addition to small schools as well. But our consortia staff sizes differ. Our structures vary. Our funding models vary. We are both 501C3s, but that's the similarity. Here you see in the animation slide, this is a graph that OCLC created when they were doing a study on consortia. So it shows the institutional missions of the member institutions in our two consortia. Palsy again is on the left and they have a group with a wide variety of those institutional missions. The whole gamut, career research and level arts colleges with different levels of intensity. And on the right is Palney, the smaller group. And we're pretty much aligned along the level arts education axes with just some career professional education and less intensity in research. So of course a number of IR solutions already exist in the marketplace. So why would we take on this kind of project? Both Palney and Palsy have been seeking IR solutions on behalf of our members for several years. And the characteristics of an IR that were sought by both groups was a solution that met the needs noted in the boxes that you see here. We could not find an existing solution that would support and energize our staff collaboration, significantly reduce and contain costs to run an IR, be easy to use. Again, especially in Palney's case, we've got small staff at our institutions and who do not have time to use complex or learn and to use complex systems. And we also wanted something that could be easily further developed that was future facing. In addition, we've got increasing pressures, both internal to our institutions and external forces that have highlighted gaps in the current options. The need for increased cost controls, perhaps the biggest due to higher eds, growing budgetary challenges. But also the purchase of digital comments by Elsevier was another tipping point that had many questioning the impact that further vertical integrations in open access space, vendor lock-ins would have about controlling the ability to manage open access content and make sure that that was openly discoverable. So as a result, Palcy and Palney developed a shared vision across all of our institutions for a scalable IR solution that a consortium, multiple consortium, groups of different types could partner to run. We have the values, the vision and the values listed on the slide. One of the things I wanted to point out there as well as the good enough technology solution, I think that's a really key point, especially when you're talking about this large number of institution, but also because of the opposing nature of some of these goals. Haiku has the potential to check all of these boxes where other solutions only had pieces and parts of that vision and with a limited ability to develop. So what is Haiku? Haiku is a product maintained by the Samvera community and it's one of their two main projects alongside Hyrax. Haiku is the standardized installation of all the basic tools of the more configurable Hyrax repository software. And Haiku is designed to meet the largest possible set of needs and scale across our institutions with a central administration and support. So that means that the functions, the object types, the metadata all come out of the box and don't need adjustment of the underlying tools. And Haiku is installed and configured as a single application but can support the hosting of multiple tenant repositories. And here you see an outline of what multi-tenancy looks like. The multi-tenancy is the most important factor behind our decision to use Haiku. Working together to support repositories is our first thought, our second thought or last thought. It's improving that infrastructure to support this collaboration is effectively our core goal. And multi-tenancy means that it's easy for us to create new tenants that have their own URL, their own collections, their own individualized branding. Our project is supporting academic libraries within a single infrastructure across all of our groups. And also at the end, I think there's a note at this and we also have a demo on our website that highlights just how easy it is to spin up a new tenant. All right, great. So now it's my turn on Jill Morris. I'm the executive director of Palsy, the Palsy Consortium. As Kirsten mentioned here at Palsy, we've been working on Haiku in some capacity since 2017 when we first piloted Haiku Direct as a part of that project run by DPLA, Stanford, and DuraSpace. And then of course now together in this project with Palmy. So our Pals project has really been defined in three phases, the first of which was designed to get a multi-tenant instance of Haiku up and running to determine whether it could meet the immediate needs of our users and of our members. And longer term whether it had the potential to fulfill a more fully developed vision that Kirsten mentioned. The second phase of this project is where we are now and it's heavily focused on building out and in some cases fixing group-based and individual tenant aspects of a multi-tenant administration of Haiku. Our anticipated future phase three of this project is where we hope to finally remove any barriers for large-scale adoption for our institutions within our two consortia but also beyond. And our methodology or our approach to the various phases of this project bring together input from many libraries toward a shared vision and prioritized needs which can then be translated into Haiku, high racks, open sources environment with a focus on a collaboration first mentality which I think is different from what else is happening in the IR space. And we hope ultimately our contributions in this space can be widely benefited from. So in terms of our project's goals and digging a little deeper into this phase of the project we identified a number of development priorities and are focusing in on these specific bullets here. First is really developing out collaborative workflow support again in support of that larger vision that Kirsten mentioned. We're wanting to provide improved ability to administer Haiku as a group especially related to roles and permissions required for the diversity of users and the permissions-based roles that we anticipate meeting as well as our desired level of flexibility and consortial level administration. So development is currently in progress here and together with Nachi we've identified and fixed some of the issues that had previously existed in the multi-tenant implementation and we're now beginning to work on the application of these roles and permissions at consortial or super admin level to create greater efficiencies and scalability. We're developing support for electronic theses and dissertations and OER content with the development of two corresponding work types which in Haiku that's just sort of metadata templates and the way objects are treated for the types of content that they are. So we're working on those two. We've completed the OER work type and we're working on ETDs. We do plan to create themed user interface templates for the diversity of IR use cases we know exist. So this work will improve sort of that front-end display depending on what the purpose of the IR is for individual institutions and we want that to be customizable at the tenant level. We're also planning for DOI integration and hoping to embed the work that's been done in the community and other projects here as well as cross-tenant searching, multi-tenant search. And then lastly and really most importantly our goal is to share all of this back with the community at large and make sure that others have an opportunity to implement the work that we've done so we're not duplicating efforts. In practice, here's an example of how we tackled the broad challenge of improving collaborative workflows. So first we asked our cross-consortial product management team which is a group of volunteers across both consortia to list out and describe the kinds of scenarios or use cases in which they might want to use Haiku as an IR. Some examples are shown here on the screen. For example, faculty self-submissions or journals hosted on an IR, theses and dissertations in archival collections. To create the specifications for collaborative workflows we listened to the product management group and then we created sample workflow scenarios to help guide the development of our efforts. We developed user role narratives and then based on those we developed a roles and permissions matrix. Then ultimately we took all of that together to develop what was a landscape visual view of how our individual tenants and roles could work together within the consortium to support the collaborative workflows that we wanted. The staff required to run a Haiku installation can be as small as one manager or can involve multiple groups of managers, editors, depositors and viewers of content which can get really complex. But we needed a system flexible enough to support the diversity of the workflows that were envisioned. So here's an overview of our timeline. We initially began in July of 2019 at the start of this grant phase and together our two consortia developed out a project plan. And as we did that we started a requirements gathering and design process that would run in an iterative fashion using that cross-consortial product management team that I mentioned. To date this group has really helped us to gather specifications for the work types that we plan to develop as well as specifications for the collaborative workflows and provided feedback on the user interface template designs. So now this group is sort of on hiatus at this point as we go into the next phase which is the development sprints. We began those in April together with Notch 8. And to date we've completed the creation of the OER work type which is exciting for both of our consortia as we both have affordable learning initiatives happening. And OER is one of the use cases that we see as a possibility for each of us. We're working on ETD development soon. We've gathered those specifications and we'll begin development on that process soon. We've refined user permissions, roles and groups for multi-tenant administration and we'll develop further support for multi-tenant efficiencies and scale. So sort of taking the management of those user roles up to that consortial level is the next step. We've sort of done the two pre-steps that needed to happen in order to support administration at a consortial scale that didn't require us to log into each individual tenant. And along the way, because this is an open source project we've of course identified an uncovered bugs that were sort of prerequisites that we needed to fix in order to address some of those other issues. We've also implemented bulk racks which is a bulk uploading tool. We know many of our institutions need that in order to migrate content. And we're working on functionality to hide tenants and also to be able to turn features on and off so that there's flexibility depending on the use case of each institution. We did begin a testing phase with some production level pilots that occurred starting this fall. And we're still testing in that testing phase right now. So we hope to have more information on that soon. But we started by actually asking our institutions who are piloting to describe what success looked like for them to describe their goals to us so that we could get a good handle on what was needed in order to be able to adopt Taiku within their institutions. And we're gathering data along the way. And all of that will help lead into how we develop our governance and business models over the long term. So I'll say more about that in just a moment. But lastly, of course we'll work on documentation and close out the project by contributing code back to the community. So just a few words about business models. As Kirsten mentioned, one of the major reasons for this project was to develop the potential to contain costs. And our vision is to jointly offer the IR services desired by our member libraries at greatly reduced costs and to support a variety of consortial and individual library use cases. So to do that, we need to develop a business model for this collaborative IR and identify what's needed to sustain the effort from a governance standpoint as well. We're asking questions like, what should staffing look like and how should it work? How much can this scale? What should we do ourselves versus relying on others? And can we add additional partners or additional institutions in order to spread the costs? We came up with a couple of different scenarios that we thought could work and be modeled out in order to support the use of Haiku at scale. I'll just focus in on the first one, which is our preferred approach and the one we hope will work. But it's really a deeply collaborative approach where we're sharing the administration, the management and the expertise across both of our consortia in order to provide that scale. We see that as the most beneficial and hopefully we can make that work. We did identify a couple of other alternatives which would be less deeply collaborative, but would still give us the ability to share costs and provide services at scale as well. So there are a number of factors that go into figuring out what the costs of this will look like when you're considering a consortial implementation. And some of them are fairly fixed costs where some of them are variable depending on how much use we see of these services. So some of them listed here are AWS, our hosting and server costs, support and maintenance contracts. Right now we've got a contract in place with Notch 8 to help support our production service. And so there's costs that's associated with that. Software development. So any development that we wanna do, if we're able to split it across both of our consortia that of course gives us additional benefit to sharing those costs. Staffing and management, we're figuring out what should that look like? What do our members need from us and how much time does it take to support one of these implementations? And then how many institutions will share all of that? There's up to a certain point, you can imagine that the more institutions you have, you will be able to reduce costs per institution up to a certain point of those fixed costs. So we're looking at all of those things and just sort of based on some of our initial estimates and models, sort of a sweet spot that I see at least in terms of the estimates that I've looked at so far, we each need to have around 10 to 15 institutions participating to really get to that ultra low cost and benefit that we aim to support for our libraries. So again, this is an open source community and software project. There are limitations to our current software, but we really do see those as opportunities for future development. And also we'll just again shout out to ask others to get involved and help us work on these things. The biggest one for us here at Palsy anyway is just the opportunity to continue to build on the flexibility of the metadata configuration per tenant. And right now, you make some compromises in a multi-tenant situation or installment of Haiku where there's some less flexibility per individual tenant, but there is work that's happening in the space to allow for additional flexibility and customization at the individual tenant level. So these are some great opportunities in order to enhance Haiku and something that we hope to see the community working on in partnership with us, hopefully in the future. So a couple of years back, three years ago now or so, CNI published its Rethinking Institutional Repositories Report, which identified a number of different challenges and opportunities. And to sum up sort of what we've been working on, there were a number of times in that report where it was referenced, the need for consortial scale. And we just strongly believe that, instead of all working in silos, we think this is a really great opportunity to leverage the diversity of our institutions and to work on open source systems like Haiku that have the potential to provide sort of previously unavailable scale and also really increase the access to underserved and less well-resourced communities. And we know that budgets are gonna be an issue going forward for everybody across higher ed, regardless of where you are. And so this is a great opportunity for us to work together and to support the community with more options. So I'll close with sort of a call to action to join the Haiku community. We'd love to encourage your institution to get involved. Haiku itself is in use by several other projects and there is a space for a wider variety of needs to be addressed with Haiku. The British Library in particular has created a shared research repository and we've been collaborating with them and sharing a lot of good information to help us shape each of our projects. Haiku is also the focus of another grant project called Advancing Haiku. And we've participated in the interest group in the Sambara community as well and that's a great place to learn and to get involved. So with that, I will close and hope if anybody has any questions that we can address them. Thank you so much. Thank you, Jill. And thank you, Kirsten. What a great project. And thank you so much for the shout out for the CNI report there. Indeed, it is always beneficial to leverage these resources communally and we appreciate you calling attention to that in particular. At this point, the floor is now open for questions. So I invite our attendees, please type your questions into the Q&A box at any time and our speakers will be happy to address those. I know that Clifford had a question. So while we're waiting to hear from our attendees, Cliff, you wanna go ahead? Sure. I'm really curious. I didn't see a mention of research data management and how Haiku could connect up to institutional strategies for that. Sort of leaving aside the ARLs who are already pretty far down that path and have enough scale that they have a lot of other options. I'm wondering how your other members are thinking about whether this is gonna have a role for their research data management needs or whether they're thinking of another portfolio of external services, you know, dryad or something like that to fill those kinds of needs. I can maybe just start with what I've heard from our institutions and then Kirsten, if you've got other commentary. I think that for the most part, at least initially the folks who've indicated interest in Haiku have not been strongly interested in the data aspect of this, the research data aspect of this, although inherently Haiku is very flexible and could ultimately support that use case. So there is likely a place for that conversation here, although it's not been sort of the initial driver, I guess, of our involvement in this project. There are also questions that we still have to answer about how it scales to support research data. And so at small scales, it's probably okay as it is today, but at larger scales, I don't know that this is the right solution for that use case. So we're gonna continue to look into that. It is something that we hear about very frequently here at my consortium and there's a lot of interest in it, although like, as you said, some of our larger institutions already have other solutions in place and we're not looking to necessarily reinvent wheels. So, Kirsten, I don't know if you have any other. We have, I mean, again, our institutions are very small and so there's that aspect of good enough. How many tools do I need to learn? Can I even get one set up? What else is going on in my institution? And so when we looked at the use case of institutional repository, that was writ large. It was essentially any data or any piece of information that we're trying to store on behalf of our own institution. But the balancing point for that is that the more you throw out this one tool, what does that do in terms of cost? And so we've even talked about in terms of media, so large scale video content. Is this really the place that we wanna store that? What's the actual cheapest way to store something like this and is it a discovery piece? What is the aspect of it that they want? And in the initial conversations, it's really, we don't have the bandwidth to even figure out what the best place is to put this. You've got this tool in front of us, can't we just put everything in there? So I think with Jill, it's initial levels of discussion about this as it becomes more of a discussion on these smaller campuses about where does this go? And they're being asked by their administration to figure out a solution for that. I saw there was a question in there from Clem. That's really about Notch 8. So Notch 8 is our development partner. So they're the one actually doing the coding on our behalf. So we build out the specifications, we meet with them weekly, set up sprints and but they're the ones actually doing the coding. And Notch 8 does a great job of supporting the community. I think they're the product owner actually in the Sambaric community for the Haiku project. Notch 8 is also really cool because they have basically a training avenue for new developers as well. And so we feel like they're really fully part of the educational enterprise that the rest of us are as well. But they're doing all the heavy lifting in the coding. I hope that answers your question there Clem. Thanks Kirsten. And thanks Clem for that question. I don't see any more questions coming in through the Q&A at this point and we are at time. So I wanna thank Jill and Kirsten one more time for coming to CNI and talking to us about this collaboration where eager to hear more as your project progresses. And thank you so much to our attendees for making time out of your day to be with us here at CNI. If you would like to hang around and join the conversation I think Jill and Kirsten will stick around for a little bit longer. I'll go ahead and turn off the recording but any attendees who would like to have a chat ask a question, I will turn off the recording now just raise your hand and I'll be happy to turn on your microphone with that. Thanks everyone. Take care. Bye bye. Thank you.