 Post lunch, stupor, and settle in for some exciting, we hope, news here. My name is John Board. My mix of titles will become relevant to the conversation we have here today. For the last 14 weeks, I've been lecturing at exactly this time to 185 students. In my role as a faculty member, so I am probably going to pace with the microphone rather than stay foot at the podium and treat you as my class. I also have the advantage that I'm totally blinded to the audience by the lighting we have here, so you can mock me, make faces, and the like. So the three of us here from Duke are eager to talk about a two-year process that we've been going through to understand how to better support our faculty and our students in their research and learning missions, especially in their research missions with this. And so I'm joined by Tim McGeary from our libraries, and Rebecca Brower from our research office, and they'll introduce themselves more formally when we get to them. So the overview of this is we're going to take you through this two-year-long four-phase process that we did to try to be very thorough in determining what faculty actually need to help do their research. Before I start that, I think it's useful to give a little bit of historical anecdotes of how Duke has gotten to where it is in the space of supporting faculty in research computing in particular. I've been at Duke for 45 years, starting as a freshman, and for 44 of those years I have been running research computing support for my faculty and student colleagues, because back in 1979 it seemed a really good idea when the department got its first large-scale mini-computer. We could have paid a grown-up to run it, but hey, there's this snivelly sophomore nerd who seems really obsessed with computers. He'd probably do it for free. Let's get him to do it and give him the root password and run everything. So that is how we started. So I ran all computing in the School of Engineering when I was an undergraduate. I wrote the damn email system. That's how you get no numbers after your net ID is be the one who writes the email system. But the very fact that engineering kind of cowboyed this system and got it donated externally, and I kind of only told the administration after the fact that we'd arranged the donation was kind of characteristic of the way computing had to be done at Duke in that era 45 years ago. And I did go away and get my doctorate and come back a few years later on the faculty in 87. And computing at Duke was terrible. No one in the senior administration thought it was important. We belonged to an entity called the Triangle University Computation Center with our friends from UNC Chapel Hill and NC State, as well as some smaller institutions that used one mainframe computer for all the teaching and all the research needs of all three of these tiny little campuses that we have. And for Duke's administration, that was check computing is taken care of. With the rise of, so my arrival at Duke in fall of 87 was coincidental with two significant events in the history of computing. The July-August issue of the Bell System Tech Journal is where Bell Labs released UNIX to the world, which made possible for the first time the creation of computing systems that were independent of the details of the hardware that they operated on and Intel had introduced this god-awful, ugly, worst piece of computer architecture ever designed in the history of man, the 8086 chip, which of course has come along to totally dominate all computing and made possible this whole world of standardized computing building blocks on which all of us now build our systems. I'm just old enough I was the last class at Duke to do programming on punch cards. So yes, part of my digital archives are still many, many boxes of punch cards. We get to the early 80s and these UNIX workstations are starting to come out and the faculty filed a report every year saying the university was negligent in not doing more in this space, and those reports were dutifully filed in the circular file every year. But they did, they, to try to shut the faculty up, if the basketball team happened to do well in the tournament and we got TV money, and this is due. So we did get this money kind of often. That's the, that was the firm budget foundation on which academic computing was run at Duke. The only administrator who cared about computing was the university librarian. So it was the librarian who oversaw all this. And then my very first committee assignment at Duke was convincing them to build one of these network thingies that might not ruin the university if we built it. So years of tragedy continue until about 25 years ago and there is the kind of crisis moment that you seize upon to change things. And after a bit of a faculty riot, I think it's fair to say, and a change in leadership, we finally founded a central IT organization to start organizing services. My role in all this was a whiner that we weren't moving fast enough and so about 20, almost 20 years ago I was co-opted as the associate CIO for Duke to be the voice of faculty inside the IT organization at least to try to make sure we were moving in the right direction to meet faculty needs for research and for teaching. So over the past 20 years we, I've had the privilege of writing all the IT strategic plans, co-authoring them over the last few years. We've had about one a decade and the first one really irritated people when we seized on the mantra third to none. That was Duke's aspiration was to be in the position, I can see that would go over real well today, wouldn't it? But our aspiration was to be in a position to be early followers of what the real schools who were serious about research support and computing support could do. We've gotten better and part of that getting better is listening carefully to the faculty, having the faculty give the university good advice about all its IT investments, not just its academic IT investments. That was one of the critical things that has fed into this process is that we've created an IT governance structure at Duke which is quite remarkable. In the 25 years that we've had a central IT organization, we've only had two CIOs and we've had Tracy Futhi now for almost 20 of those 25 years that we've had a CIO. And she is sufficiently good at her job acceptable, as the Vulcans would say, that I'm willing to do two full-time jobs to be able to work with her on doing these things. And so about two years ago, in the meantime, I also have oversight over research computing at Duke at this point. And we were discussing that it would be, we know that our research computing support is too high-performance computing-centric because that's what you're out of, we're very good at that. There's a subset of our faculty who do need those services with whom we've partnered very well. But everyone else on campus in all departments, there are faculty doing intensive computational things and we did not think that the collection of infrastructure we had in central IT was fully meeting their needs and we knew we needed to expand to have better resources in the digital humanities and social sciences. We needed better and easier to use protected networks in which research on sensitive data could happen. So Tracy began organizing a very thorough review of how IT, OIT initially, our central IT unit, could better support research computing needs of faculty all across the institution. So how do we support the non-HPC crowd? There are lots of them, especially the social scientists and humanities faculty who generally have the added excitement that they have no money. As has been said at some of the sessions I've crashed today, the engineering faculty, the STEM faculty usually have money to buy more or less what they need and have the grantsmanship opportunities to get the money if they need it. But when I started working with our social science research institute before it went through some renaming, we were going, oh, the boundary condition is there's no money, but I'm a PhD student and if I'm going to graduate I have to do this study anyway, so please make that happen. That's a very different world for those of us who are used to supporting STEM faculty. On top of this, no news to anyone in this room that there are one or two more regulations now around the use of data than there have been in the past. Our vice president for research was addressing our equivalent of the faculty senate last Thursday and showed a graph plotting the cumulative number of federal regulations and policies affecting research starting from zero in around 1993 and it's almost a linear plot up to 250 today. And she was being very transparent with the faculty that this is why you're feeling a lot of pain and we know you're mad at us because you want us to make it all just go away and part of our goal is to see how much of it we can make go away knowing that, nonetheless, it is real. Managing storage is hard. We have not one but several storage options at Duke. It depends on the type of data you're using. It depends on the classification of that data and we have the added excitement that we do have an academic medical center and any of you from institutions with an academic medical center know and an associated gigantic hospital network that their IT world is very different from our IT world. At Duke they have separate CIOs, separate organizations. We do have shared services across the whole campus. Our network is now more or less the common network. Some of our storage systems are common but it's still very much one country, two systems and the faculty who really get stuck in the middle of that tend to be the basic sciences faculty in the School of Medicine whose work has nothing to do with human PHI since they fall on the wrong side of the fence. They're hard to act as if all their research was on human PHI and that is a problem. So as one of the reasons we still like having Tracy around is her first instinct is let's go to the faculty and see what the faculty think about all this. And she did not just talk to one or two. This is where the IT governance that we have at Duke is relevant as part of creating this central IT unit 25 years ago, the faculty insisted there be a single IT governance group that talked about all types of IT investments at Duke, not just the research and teaching investments but also the administrative systems that we want all these systems. We have experts in the business school and the law school in the public policy school who can guide the university even on something as mundane as the HR payroll system. And because Duke can make some tragically bad investments in that space in prior decades. So we have this group meet 90 minutes every two weeks all year long including in the summer, although we do drop maybe half the summer meetings, and the faculty consistently say it's their favorite committee at Duke and can they be reappointed please. How many of you have an advisory committee like that? The way you get an advisory committee like that is to listen to them for 25 years and either do what they say or explain carefully why you understand what they are asking for but why we have to do something different for this and this and this. And when they recognize that what they say matters and they was me for a long time because I was a member of this for many years before I became evil and flipped hats and became an administrator instead of a pure faculty member. But we have this group. It has 15 faculty who come. We also have undergraduate and graduate students who are full members of the committee as well. As well as IT staff, we have a very distributed IT structure at Duke. Our central IT is 350 people or so but overall just on the campus side there are about a thousand IT people. Tim has a bunch in his shop. There are IT people in Rebecca's shop. Every school still has some local IT. Most business units have some local IT. So IT is a big coordinated mess at Duke. But the 15 faculty just on our advisory council she did not think were sufficient to speak for the full faculty. So she has the clout on campus to be able to ask and get yeses from 50 other faculty to come participate in a series of seven working groups to talk through in detail the emerging IT support needs for research that we were seeing all across the university. So we created this four phase process that we've been going through. And phase one is what I will conclude talking about here in just a few minutes. This is the airing of the grievances. This is where we get about the 65 or so faculty participating in this endeavor. And you know how hard it is to make faculty talk about what irritates them, right? So many notes were taken, right? But what's critical about the process is if we didn't stop there, just collecting a list of things faculty are pissed about doesn't move the university forward. What is critical are what has happened subsequently in phases two, phase three, and phase four. And just to jump to the end position now that we are actively writing job descriptions for many new people and new services as the result of the thoroughness of this study. But we will get to that when we get to it. So phase one, we had the listening sessions with seven groups of faculty, natural sciences, social sciences, basic sciences, engineering, and then our friends in the humanities made it very clear to us that there couldn't be one group representing the humanities. We needed three separate groups to understand the complexity of computational needs in humanities which we said great, thank you very much. Tell us who they are. We will talk to all of them. So all seven of these groups met a number of times and gathered and went through a semi-structured process to get faculty to talk about what was really holding them up in their research when it came to data and computing as broadly defined as we could make it. Each of these groups then consumed one full session of our advisory meeting group airing what they found with the larger IT advisory committee. And how many dozens of entries were there of things that the faculty had explicitly complained about. Many, many dozens of post-it notes were put up on whiteboards in a couple of poster sessions that we organized and filed and tried to coalesce into some themes and were able to reduce after many months of work into this very simple-looking chart. But this captures ten things that were very important to the faculty and the surprise to Tracy, I think in particular, was how much it was not about what central IT can do anything about. That when you actually ask the faculty to start complaining about what holds up their research, it was all about where's the subject matter expertise to help me understand in my field how to accomplish, navigate the storage systems, navigate the data use agreements, navigate all this cruft that is in my way of just doing the science that I want to do. And so we would be calling it the research IT needs assessment process and the slide decks coming out of that, Tracy made it very clear to put a red X through IT. It became very much the research needs assessment process of which IT was a component. The three boxes on the bottom are the ones where IT, kind of where we were thinking things would go. Of course, faculty want some additional HPC resources. We have an entitlement level for computing on campus and they want that entitlement bigger and they want it spread to PhD students. Those are things we're going to be able to accomplish. Shocking. They want more GPUs on campus. We kind of knew that, but we were able to quantify that through this process as well. And with everyone in every discipline thinking about machine learning, many of these disciplines don't have the expertise to really tune the models well themselves. We have a large outcry for help building very specific pipelines in the data analytics and machine learning spaces. Five, six, seven. We have so many solutions and yet no one has ever heard of any of them on the day that we tell them it exists. And so there is the constant cry that we've heard for all 30 years I've been involved in IT delivery that can't we simplify our documentation and make it more clear what Duke offers. I'm sure the libraries never have this problem either of crystallizing for the faculty the services we have. We will try to do better with the added complication that now some of those services are cloud services as well and yet we're still kind of responsible for brokering them to the campus community. And then a special bullet on storage. All of us have perfect models for storing all the data forever even though the grants only funded for five years. I've heard that theme at this meeting I hear it at every IT leadership meeting I go to. We are also making progress in how to build long term storage options that are affordable, economic and meet the specific requirements of NIH and all the other funding agencies as well as are easy for faculty to use and we are fortunate enough in our IT organization that we get lots of NSF grants. I saw a couple of people refer to CCSTAR programs in some of the sessions I've crashed today. We have had a lot of CCSTAR grants and some right now very much looking into a modern open source file system that integrates metadata as a first principles part of organizing the file system with some tools for automatic migration of data from hot storage to warm storage to cold storage that we think is going to play a role in this and faculty certainly indicated these were needs. But I've skipped the most important boxes because this is where we were kind of surprised that by far the largest human outcry was number one. We need teams of domain specific technical people who can talk to someone in the social sciences someone in the humanities about how to map the problem they're trying to solve to technologies that Duke has available. We didn't have a great answer for that at Duke and this was by far the largest outcry, the most consistently reported thing. It came from STEM faculty, it came from non-STEM faculty but kind of especially from non-STEM faculty I think. And then kind of related to that is we have a lot of training programs that run at Duke to teach people students, faculty and staff about particular technologies outside the context of any particular course. How can we better organize and publicize those so people realize there's a lot of extra and co-curricular education already available on the campus to help students and faculty who are willing to admit they don't know everything about how to take these courses and fill in these gaps themselves. In B, this kind of get, this is Duke code for, we have the separate university and health system environments and faculty, especially faculty who fall in both camps and that's a lot of faculty at Duke have some type of joint appointment between the two entities. How can we make their lives better? And that is one of the more difficult things we have built at, but we have made some progress in this process. And then Box C here is about our Duke's security and compliance posture. You can Google Duke and NIH and you can find that we got ourselves in trouble not too many years ago. The misdeeds of one can cost the university a lot. One 12. And then a big headache for all the faculty who are left who are now on double secret probation from all the funding agencies to prove that the rest of the N minus one faculty at Duke aren't like the one. There's more than one. Two or three now. So that's part of the problem. When you have enough faculty you will have some who misbehave. Since by dollar volume grants from the School of Medicine and the clinical side tend to dominate Duke, a lot of our research policies historically tried to treat all research at Duke as if it were on the core model of research in the health system and write rules that were intending to apply to all grants at Duke, which you give to a history faculty member writing a national endowment doesn't really make sense and yet is ordering a whole bunch of thou shouts about the production of his book I guess that says it has to be done in secure storage in a controlled environment. So a lot of angst about our one size fits all approach to security and compliance. And so the faculty were very vocal in their complaints as we put these together those six high level categories that we lumped the 10 things into it's very clear they're not independent of each other almost all of these influence almost all of the others a lot of hours went into figuring out which arrows don't go into that slide but basically it says everything depends on everything so we can't solve these problems in isolation. So the most important thing that I probably did at that point was recognize this is not an IT problem and thanks to having a relatively new university librarian and a relatively new vice president for research at Duke they agreed wholly with Tracy that this was a university wide problem to solve by all three of these units working together and with that I turned the story over to phase two. You're going to So good afternoon I'm Tim McGarry I'm the associate university librarian for digital strategies and technology in the university libraries at Duke and I'm happy to talk about phase two and three of our process so as John mentioned the three co-sponsors Tracy Futhi, Joe Salem and Jenny Lodge identified leaders and service providers and faculty champions and stakeholders to take part of those six working groups that John described a couple slides earlier. So again we collected another 55 individuals some of whom were in the original group and some of whom I believe were new because we wanted to test the priorities that were identified by the faculty in the first phase with a different set of faculty and so we met weekly for 10 weeks and I will test that is a major accomplishment in and of itself to get 55 individuals a very short notice to commit to a 10 week weekly process I can also test being a part of four of those groups how complicated that got so they identified 39 potential services that can meet the needs that were in those six boxes and to address those and then they consolidated those priorities to 29 and then to 21 and they kept calling them down and trying to pull together overlaps of different groups and then we focused on surveying the faculty and estimating costs and then we landed on the top 12 priorities which will advance to phase 3 so let me talk a little bit more about that process so these slides will be available so you don't have to try to read all of the text here but here is a list of the 39 different service providers their proposals that came from each of the six working groups and we identified them in the boxes and we grouped these priorities and proposals and again I'll also note this is where we x'd out the IT part of the top three boxes because it really wasn't just IT it was really research needs across the university so we did another faculty survey which we had about 58 faculty participate and these were individuals that were part of phase 1 and 2 and we had about a 67% response rate in those surveys and you can see a little bit of how they grouped the priorities there was a focus on the domains that they came from and what was impactful for each of their contacts and the faculty did heed our request to spread out their scores across 21 services we knew that we couldn't do them all so we asked them to pick what were their highest priorities but even the lowest rated service overall was rated highest by some so it wasn't as if we had a very clear picture of well these are the ones everybody agrees on there were priorities that everybody at least some person in the group felt like it was the most important priority so we still had more work to do so we took the 29 distinct services that we whittled down to and we started focusing on what the cost estimates might be and we grouped these together so the top three bubbles represent what was listed as having the highest strategic priority by the sponsors so this would be Tracy and Joe and Jenny and then we decided from those perspectives that those would be the three groupings of priorities that we even move on to phase 3 and identify funding for there was another two priorities that we put on the bubble which we would put in a parking lot to evaluate further and then the other five priorities that were down the lower strategic priorities were listed as ones that perhaps maybe are more school based or maybe has some mixed overall reactions and we could look at those later. Perhaps we need to do some local pilots or some collaborations you'll notice the purple bubble in the middle that says sponsor additional priorities these are ones that the three co-sponsors identified specifically that were not ones that were in the faculty list but had been deemed ones that the faculty would enjoy because they would take out some of the work or some of the gaps that the faculty themselves may not have either visibility for or can't accomplish themselves. So then the co-sponsors then went again and mapped these 29 distinct services again not only by the groupings that they had in the previous graph but the cost that would be associated with estimated costs that would be and then what groupings of those six boxes that we described earlier where do they fit into each of those six additional priorities. And so again you can see some of the sort of sizing estimations that they were working on to determine what kind of funding options we may need to be looking at. So these are the 12 service proposals that are advancing the advanced to phase three and so in particular as John described the most important box about adding people in the domain context areas we're looking to add 15 to 20 FTEs that are spanning across the libraries the Office of Research and OIT and then another one to three FTEs for research programming and support. We're focusing on we're focusing on augmenting the people we already have in each of our organization so it's not a completely new set of skills and a new set of professionals but it's really augmenting the gaps and where the demand has been where we can measure the demand and see what an area we need to be focusing on. And then more intentionally than we have in the past we're going to be building cross departmental virtual teams so that each of our three organizations now have a matrix relationship we've all worked very well and collaboratively together but this is going to be a little more formality in the perspective. I'm not going to go through each of these areas in particular but you can see that the list that John started out with of the kind of high level areas have been more specifically identified into very specific priorities that we're going to be focusing on. Then we regroup these priorities again into three different areas and so one was focusing on better support for researchers by adding personnel, improving coordination the other second category would be enhancing computational services and building capacity for data intensive resources and the third being balancing security and compliance requirements and the flexibility needed to support different types of research and so we've broken those six boxes into these three high priority areas and allows us to think about the ways that our virtual teams will interact together to meet those priorities. So the phase three perspective or phase three of this process was discerning the structures of how to implement it and this is where the co-sponsors again Tracy and Joe and Jenny added some additional participants and partners so they convened a planning team with the financial leads on campus this would be the executive vice president for finance the provost the dean of the school of medicine the dean of trinity college of arts and sciences and they began and then financial experts in each of our areas to begin to work on what kind of ongoing funding options we might have for each of those priority services do we charge grants directly do we consider what our indirect calculations are can we do philanthropy fundraising for this can we do we need to be thinking about how to pitch another allocation funded proposal to the dean so that they share some of their allocations or consider what cost sharing would be across the different schools they developed a multi-year funding and walk-up plan which included as if needed or as needed bridge funding for various pieces because we weren't entirely sure how quickly we might need to ramp up some of this or how long we could push off and delay some of these implementations and then finally we wanted to be sure that any planning we did would have the sponsor backup backing so that the sponsors would be able to continue to implement these in each of the three organizations so they developed a five-year pro forma that the co-sponsors and the co-funders agreed upon and believed that they were going to be supportive of each of those categories so the categories were in research support the storage services, the service navigation the IT services and infrastructure and then securing compliance and the provost and the executive vice president for finance identified that they could come up with an initial investment of a million and a half dollars as the potential to increase over five million dollars over five years and then we've also begun to explore philanthropic strategies with Duke Development as we're currently in planning for a future capital campaign so these are some of the types of positions that we're able to prioritize in the next phase so in the office of research they'll be focusing on data security and research and navigators in the libraries we're going to be increasing and embolliting our research data management program including adding more capacity for data visualization consultations and data licensing and data sets that our faculty are needing to acquire continuing again with the strategy on metadata being core to how we do long term preservation and access and then also thinking about how to expand in our digital humanities and GIS specialties and obviously the list that you see in OIT are ones you would expect we need additional technical personnel website consultation software development business analysts and support and security analysts that we need to be thinking about given the demand that's rising and then the domain expertise focusing specifically in the areas that each of the faculty groups identified as being critical for their domains and then the types of services that we're going to be implementing increasing our cloud infrastructure and storage to connect active and published research what is that bridge that we need to be building between where the research has been ongoing and where we need to be thinking about long term building protective enclaves for our research data again in different contexts because not everybody needs the same types of protections for the data they're using but how can it be contextually appropriate collaboratively licensing strategies to enable that we're doing campus wide and context limited data sets so the libraries have been doing data licensing for some time that apply for the entire university but when it doesn't apply university wide it's harder for us to do the licensing and faculty are kind of left out outside wondering how do they get the access to the data they need for their research so thinking about different strategies for how to do that so we have the right organization doing the work focusing on also semi-autonomous and reusable sub-clusters to support faculty startup packages so instead of giving faculty specific packages that are just for them how can we create our computing structures so that it meets the needs of what they need for their startup so that when they're not using it or when they're moving on to different things it can apply to another faculty member and continue to be reusable and sustainable and again as John mentioned earlier we all started this from the perspective of how do we enhance the computing cluster that we were already pretty strong at and I'll turn it over to Rebecca to talk about how we're doing implementation so while John talked about 45 years I get to do two months so my job is a little bit easier so my name is Rebecca Brower I'm not sure if we're kind of a unique office so I might take just a minute because I'd love to hear if anybody else has something like this at their institution but I basically employ people that have previously been faculty members or staff so they've done their own research or they've been the right hand person for a researcher and we try to make it easier for the researchers that are at Duke so whether that is sort of navigating them through the complexities of policies or resources or figuring out how to do something that's one thing we do but we also interface with other research support offices so the contracting office, the IRBs and things like that I think I was telling John yesterday that I like to think that we wear research support mullets we're service up front and we're compliance and regulations in the back so I'm going to just talk a little bit about what we've been doing for the past couple of months and then also just looking ahead so a lot of this has already been alluded to so I'm going to be brief we went ahead and have established an implementation working group so we have a leader in each of the three realms OIT, the libraries and research and we each brought a couple of others into the working group so in my case I brought someone who was well versed in the data security elements as well as somebody in scientific integrity so that they can help us think about how we're going to implement these priorities we've been meeting together as a virtual team I think there are about 12 to 14 of us that are leading the charge at this point we're doing a couple of things right now number one trying to engage some of those school level resources so these are going to be largely IT resources also thinking about engaging the central unit so this is your contracting office, privacy IRBs, things like that and we're also starting to re-engage faculty sponsors and that's not up there but we really need to make sure we have those faculty champions as we're thinking about implementation so the things on our to-do list and actually these are due Friday if anyone is in the audience that's working on a job description we're prioritizing hiring so we have I think seven jobs that we're going to be hiring for in the spring we've prioritized it not just by writing the job descriptions but we asked our sponsors to help us remove barriers to HR I don't know what it's like at your institution but having a job description doesn't mean that you're hiring the next day it can be months we said we can't have that happen please help us remove those barriers and so that's good I think we think our job descriptions are jobs posted in January we're also connecting in with other service providers that may not have been included in phase one two and three so if you hadn't heard about this whole initiative at Duke you know about it now because we're making it a priority the working group we actually have a subgroup of the working group that's looking specifically at metrics so one of the things we want to make sure we're clear on is what does success mean we're going to be talking with our faculty ambassadors about that what does it mean and then we can start to map out the milestones understand whether we are meeting demand are we at capacity just really make sure that we're tracking everything so that as we go through the months and years on this do we have the staffing we need and the resourcing we need to meet the priorities as we had described them and then finally especially with us being we have worked together I mean this is not entirely new but we've never worked together with these very specific priorities in mind so we know that we're going to have to add some governance to this whole model so I think that Tim obviously was talking a bit about this how we have the 12 services that have been defined the 12 priority areas when we were in phase one frankly when we started to hear what the pain points were I have to say on some things we just got going we knew it was a no brainer we had to get moving on some of these I'm not going to read all of them I'll tell you I think one that's really important in the top center there is we're working on our protected enclave and I think this speaks to number one it actually has a little bit to do with security and compliance but also just making it easier our protected enclave at Duke previously was very very onerous a lot of barriers to getting in so we can do better so we are actually in the midst of developing a two tier protected enclave that has very tight security controls but if you are dealing with regulated or protected data we're going to layer administrative controls on top of that we're actually we're in pilot right now and we'll be launching in the spring so I'm really excited about that that's a huge advancement and then we have several other offerings that are well on their way including a self service and it can be to find things so we are working on a finder tool similar to what's available at Cornell I'm not sure if you all are familiar with this but you basically answer a couple of questions about your data classification your compute needs and your storage needs and then you're offered up here the technical tools that might meet your needs we're working on something very similar and I think it'll be great when we have that available not just to kind of find on your own but to actually infuse that into part of process at Duke so when we know someone will be at that point in the technical tool let's slip this finder under their nose so next steps so we already mentioned that we're doing our hiring that's exciting the other thing we're doing is we have these 12 priorities when we started meeting as an implementation team we realized we were already putting in the libraries you were already putting the pedal down on this area and OIT was doing this we're looking at our existing sort of processing of the priorities and figuring out how we're going to all work together especially in some areas where we have a lot of overlap I know that Tim and I are working on a data licensing service and we're going to have to just figure out where this is all going to lie in sequence so that's been a big it's going to be a big thing over the next several months and then also just learning how to work as a virtual team I think we are one thing we haven't mentioned but we recognized when we were meeting that we all had day jobs already and so we thought you know it's great we're going to carry these out but there's no way that we could put this on top of our on our existing roles so we are going to be hiring probably a person that's going to be serving as kind of an administrative director or program manager that's what this is going to be their entire job is to drive this forward pull us together when needed and really just continue to drive forward the priorities so the takeaways so I'm guessing some of the things that we've said today are not going to be surprises for you John also mentioned this may feel particularly unsurprising if you are coming from an academic medical center where there can be additional pain points and barriers there so I think that I'd be curious to hear you know if you all have the same experience that we've had in terms of what you heard from your faculty but another big takeaway and this is if you're planning to do this at your own institution there were three things that were key getting that really really sincere faculty input every step of the way I think is absolutely critical the leadership engagement our leaders are very curious and so they're asking us questions they are holding us accountable this is a huge priority for them and without this this never could have happened because there was pressure I will say to make it work and there is an ongoing requirement for this to be a priority for all of the groups and then the other thing that I know is near and dear to my heart because I feel sometimes those of us that are required to implement things get left out and you're handed a steaming pile of garbage that you have to go do and you said I didn't get to weigh in on whether this was going to be feasible and I think all of us were engaged early on so we could really start to think how could we actually make this work so having all three really important frankly at every single stage and then the other takeaway in case it's not implied it is time consuming but I think it's extremely hopefully very beneficial for the researchers so we're looking forward the next 18 months are going to be where we really focus on I think understanding our metrics and how things are going and then we'll be making decisions every single year to continue the staffing and to continue the investments so I think that we're going to leave some time for questions so we're here from University of California Riverside this is really solid and really inspiring piece of work thanks so much well your top priority over there 15, 20 FTEs really doesn't surprise me at all I've been working many years with researchers well we've been working with computational social scientists, computational biologists computational whatever this after that your work would make their life a lot easier but on the other hand I would like to have I wonder if you consider there are lots of biologists psychologists social scientists who does not have a computational thing in front of them they have large data sets they have a research question and who can answer those questions so I wonder if any of your 15 to 20 FTEs would be taken that type of work yeah I think so we also know that we can write all the job descriptions we want the number of people who can do the job of being Superman and Batman simultaneously is limited and will work for what a university can pay for them so we've also put a lot of thought into the ecosystem of developing people who can do this so we expect especially in some of the disciplines you just named to leverage in a formal way PhD students and others through fellowships and other opportunities to create some aggregate FTE out of the young domain experts in the field who have grown up with the technology even undergraduates can be a huge asset to a senior faculty member in a humanities department to navigate how to use in the grand scheme of things not terribly complicated technology but it's complicated to the faculty member so we have worried from day one about where this domain expertise actually exists and expect it to be really the gamut of full-time people down to post-doctoral and graduate student fellows and even undergraduate fellows helping us to deliver this ecosystem yeah the the majority of the staffing that are coming to libraries will be part of a center we created a few years ago called the Center for Data and Visualization Sciences and their mandate is to do that is to be that baseline foundational giving consultation, trainings, workshops and again echo John's point we will train faculty when they are willing to come and be trained but we are primarily focused on graduate students and building up that skill set so that research assistants the graduate assistants can then support their faculty and also have that career trajectory for themselves and have that skill base for themselves one of the one organic thing that happened at Duke that isn't in this project but it's going to be a benefit it's going to benefit from this project but also will be a benefit to us as we created an intellectual center called the Center for Computational Thinking which has also been examining not only how do we inject computational thinking into different disciplines but how do we create curriculum and co-curricular activities around it so that we can boost the holistic experience and so I think that's another avenue we're taking Thank you very much Hi, I'm David Millman from NYU that was great, thank you very much we're in the middle of something similar to this so I'm looking forward to copying your slides down but the question is and I think you may have said this or I took better notes from start to finish I mean you're not going back 45 years but when did you convene the first set of faculty I think it was February February 22 2022 and actually I think if you look in the sketch we actually have links to the full reports both of phase 1 and phase 2 so I know for sure some of the details are in there Great, thank you, that was great Good afternoon, David Drivesback from Florida International University this is a great presentation I'm curious as you assess your entire IT infrastructure of people where you are now how much progress do you think you can make with the existing IT folks to develop them to support researchers just in general as it relates to training and so forth as you develop the community Yeah, we've already done well it depends on the discipline we already have quite a number of our IT staff who are very comfortable working directly with faculty in the kind of traditional HPC space because you know the faculty who are in that space tend to be kind of geeky themselves and so they're able to you know to communicate with the bit breathers we keep in the basement to understand how to optimize the networks but we can't let them out to the humanities faculty so this is so growing in my head is research under research computing we are trying to grow the team of people who are really good at going out and talking to faculty from all disciplines and translating their needs into the services they have it is a rare skill set to find someone who's burst technically enough to understand our services under the cover and can disarm a faculty member into speaking candidly about what they're really trying to do rather than I think I need this cloud service today what's the actual workflow and we have a few people like that that share Katie is scalable in her current form so we need to replicate that person a few times over but we've at least seen demonstration proofs that we can do that and independent of this initiative we already have there's already been some campus-wide organizational assessment of where do IT staff need to be and you know we all as John said we have a distributed campus but we also created some more centralization services for example merged a big chunk of their IT staff into central IT because they weren't really college based and what was retained were the context specific technical staff that really needed to work with faculty and it gave the other more solid foundational IT staff an opportunity to be in a larger grouping a larger network that helped that distribution to be more possible and the libraries of course are similar in that as well so we could be focused on the specialties that the libraries can deliver more specifically. Hi Jill Sexton, NC State University I was wondering well first congratulations I'm really thrilled to see you get this kind of buy-in especially since you're a kind of neighbor of ours and I hope to be able to leverage your reports on my own campus I was wondering if you could share a little bit about the cost and how you raise the money so I'm assuming these are new positions and any advice you can give us to making the argument that this is a worthwhile investment I think we would all benefit from. I think some risk went into the process when our provost changed halfway through because we knew our old provost bought into this and I think there was some nervousness but our new provost immediately bought into this idea as well and viewed it as critically important as did our AVP. I mean one of the most remarkable things about this is we are writing job descriptions now and starting the hiring process before we actually know where all these people are going to live because we got the marching orders from the most senior leadership you've convinced us this is critical to the university we know we might be arguing for months over some of the details of who lives where but go ahead and hire them it's too risky not to proceed with the hiring so go ahead and do that and trust that the process will eventually home them in a sensible place but it's because we had the buy-in from senior leadership more or less from day one that this is a critical thing to do Tracy wasn't going to waste our time if there wasn't a chance of it actually having effect. Specifically to where the funding is coming from so the initial wave of funding is coming from existing resources so strategic funds that were identified as and so you can imagine they're predominantly one-time or at least opportunities for one-time injection that we'll be able to walk on in the distant future the other model they're looking at is what kind of grant tax can we apply so we're actually we're looking at stealing an idea from UNC Chapel Hill which had applied a tax to start up their data science program so we're looking at that and what would the calculation be and how much can we raise again that's one that they want to make sure that we'll pass muster with the federal funding agencies that we can actually apply that and then the third aspect is strategies and philanthropy and so what we've been discussing with the central development organization is again starting with the faculty so having faculty ambassadors write their impact statements about what the impact is to their research is making sure that when we make big requests from donors that we apply a 5%, 10%, 15% sort of addition to that request to say this is what is needed to support the research and sustain the research that's going so our dean of Trinity talked about being in a living room where getting ready to pitch a big research equipment need in a particular area where this family has a history of wanting to be a contributor to and said to them this is great this gift will go and buy this piece of equipment but we still need the people to run the equipment and sustain the equipment and so it's going to need another 10% gift for that and that donor understood that and that kind of compelling case is what we're also going to be working on. One of the keys of keeping the faculty coming to these IT advisory committee meetings for 25 years is we always guaranteed to end on time so we will end on time and thank you for your attention.