 So, my watch says 245 and I like to be on time for things, so we're going to get started. I want to welcome you to this talk today. My name is Shelly Knuth and I'm the Assistant Vice Chancellor for Research Computing here at the University of Colorado of Boulder. And I'm also the PI for the Access Support Project, Access Being the Advanced Cyber Infrastructure Coordination Ecosystem Services and Support. It's a mouthful. So, who am I? You know, I just sort of gave you my big fancy title, but basically I'm the director of the research computing group up at CU and so on campus we provide support for all the large-scale compute resources. I'm also the executive director for the Center for Research Data and Digital Scholarship and I bring that up because it's a really diverse group that combines our research computing with libraries and digital scholarship initiatives and it really helps with bringing data to the forefront of what we're thinking about. I started out as a research scientist. That was sort of my first 12 years of my career. I spent a lot of time going to Antarctica doing meteorological research and that gives me a perspective from the researcher side in user-facing services and there are a lot of people coming in right now and I wasn't, I started on time but I want to make sure that I'm allowed, I don't have too many people jumping in the middle. So I'm also the chair of the Rocky Mountain Advanced Computing Consortium or ARMAC and ARMAC is a group of institutions that have an interest in high-performance computing or data and we are very active and we connect pretty frequently around all kinds of issues related to that. And as I stated, I'm the lead PI for the Access Support Project, formerly called MATCH. How many of you here knew about XSEED? Okay, good, so maybe a quarter of you. How many of you didn't hear of XSEED but have heard of NSF Access? Oh, wow, okay, great. So I want to, for those of you who maybe are unfamiliar with either XSEED or Access, I want to first step through kind of describing what that is and then talk about what our specific project is about. And the reason that I'm here to talk to you today is really about expanding our reach within the communities that we either have traditionally served or not traditionally served in terms of these large-scale computing resources. So Access is the acronym for a project that is the follow-on to the XSEED project. The XSEED project ran for about 12 years prior to XSEED. There was Terrigrid, anybody here heard of Terrigrid? All right, there we go. Terrigrid and XSEED and then now Access are all basically forming an ecosystem of large-scale advanced cyber infrastructure resources across the country. And these resources are independently funded. So a institution such as the Pittsburgh Supercompeting Center will put in a proposal for NSF to get funding to support a system on their campus. And what they will do is say we want to be part of the Access ecosystem. And then what they do is they have to fit certain parameters such as integrating with some of the centralized resources and whatnot to then be formally called part of that Access ecosystem. So there's, this is a map that is actually from XSEED. I couldn't, I was trying to find an updated one and couldn't find it, but I don't think it's changed very much. And I know it's at the best looking map. But there's maybe, depending on your level service provider, there's maybe, I want to say 20 of the level one service providers. So these are some of the biggest supercomputers in the country, the storage centers and things like that. So we have all of these resources. They're all across the country in different locations. And we want to combine the use of those resources and then combine expertise to provide this ecosystem. So you can imagine if you have 20 different supercomputers across the country that each have a different focus area. We might have one that's focused on traditional computational uses. We might have one that's focused on more bigger scale with GPUs, for example. Or we might have one that is supporting cloud. You have all of these different resources and you have expertise at those particular institutions who are feeding into this bigger system. And the whole goal here is really to enhance research discovery. This is funded by NSF, that's what their mission is, is to enhance research discovery. The new access project, it had five access services. The fifth one didn't get funded and I'm not really sure if it's going to. That's an NSF issue and not mine. But there are four tracks. And so this, if you were familiar with Exceed, Exceed had one PI, John Towns. And it was a $120 million project, I want to say. With access, they split it up. So instead of one PI, one institution, and everyone feeding from there, there's four service tracks, different PIs for each of the service tracks. And then there's a coordination office that coordinates efforts between all of them. So the separate focus areas that, again, work together to support researchers are really, we have a group that's making sure that the systems are just running. You know, how are they, and there are people on site that do that, but then there's how are they, for example, are they integrating with the login set up that we have that we provide to all of the resource providers and things like that. So there's just the operations making sure that everything's working okay. There's collecting metrics and measurement. There's making sure the researchers know how to use these systems. And that's where my project comes in. I'm the Access Support Project, and that's really what our goal is. And so we always like to say, if you build it, they will come, but that kind of depends. There's a lot of apprehensiveness around using these high performance computing systems or large scale data systems. As I said earlier, I was a research scientist, and then once I got my PhD, I decided I didn't want to continue that path. And then I entered into the world as a data management specialist. And after doing that for a while, I supported users and was the head of the user services team for quite some time. And the one thing that I heard the most there is, you know, I'm afraid I'm going to break something or I'm afraid to, I don't know how to do this. I don't have time to learn how to do this. Or I don't think that my project is important enough or big enough to work on one of these resources. And it's tough because I can appreciate why people feel that way. But, you know, things have changed. This isn't all about the traditional, quote unquote, traditional users who are the ones that, you know, have all of the processing power and memory in the world. There's tons of different types of workflows that really can utilize these. And so one of our primary missions is to help people use those systems. Because if people don't know how to use them, if people are intimidated by using them, they're never going to use them in their science. It's going to be delayed or it might not even get accomplished. So that's what we're here to do. We're here to make sure that those researchers know what they're doing. And we have a $10 million five-year grant. We're just ending year one. I have my panel reviews next week. It's very exciting. And here's all my colleagues as well. And I mentioned this already, but, you know, who uses access resources? Anyone. Anyone can use them. And we are here to make sure that you know how and that you feel comfortable doing so. One of our primary goals on this project is to introduce more researchers to our community. And then specific, we're looking for people from underserved groups or people from underrepresented disciplines. I am the only female PI as part of the five PIs on the access project. And that is not lost on me. And, you know, coming from the background that I did, I didn't necessarily feel like I was appropriate for, you know, using these systems, much less running them at some point. Imposter syndrome is, you know, rampant as we all probably are aware or I'm the only one, I don't know. But so, you know, and we, in my work with the, I mentioned the Center for Research, Data and Digital Scholarship. There's a lot of groups that are in the social scientists or arts and humanities who, or history, who do use our resources and very successfully. So that's what we're really here to do. How do we get more of those people in? So let's talk about the support project of this. And, you know, my goal here is to, again, sort of showcase what we're doing and perhaps show you that this is a great way to get involved and to begin to expand what we're doing. So the goal across access in general, but for us as well, is to provide equitable and scalable support to best enable research on NSF-funded cyber infrastructure. And the way that we feel to do that is by democratizing access to these resources. The way that we think is best to approach this is through a tiered model. And you can see that in the diagram that I'm showing there. It's really important to stress that it's not a researcher, a user of our systems, is not actually stepping through a pyramid. This is just a way to illustrate this. You can imagine, you know, I've laid the groundwork. I've said, we have researchers, they're coming in, they're learning about what this is all about. They might be feeling intimidated, they're unsure. We wanna decrease their time to science. We wanna incentivize them to use these resources. And to do that, we have to remove some of the incomplexity. And then when there is still some complexity, we have to make sure that they can get those answers. Really, what we're here to do is decrease that time to science. And to do that, we have to have a scalable model. We cannot support an entire national infrastructure of users, thousands and thousands of users, without having a scalable model. So the first thing that we've done is we've leveraged some existing tools in the community, which I'll talk about in a second, to remove the complexity right off the bat. So researchers are not getting too far into the details, and they're able to do the work that they need to without ever really having to actually interact with the system. If they still have issues, if that doesn't help them with their work, we've provided a knowledge base, a collection of resources to help them self-serve to answer those questions. And if that still isn't it, at the top of the pyramid, we have concierge level support where we can help. So let's talk about each of these real briefly. So the first is leveraging existing tools. The two tools that we're leveraging are Open on Demand and the Pegasus Workflow tool. Open on Demand has been around, well, both of these have been around for a few years, that Open on Demand is a really neat interface that's run by the Ohio Supercomputer Center. And what it is is you can log in on a web interface. So you can go to Chrome and type in the, whatever the URL is, and log in to your favorite supercomputer. You have to, of course, be granted access. So for me, for example, I have access to the University of Colorado's system and I can go to our web page and assuming I have an account, I just go into Chrome and I log in and suddenly I'm ready, I have the supercomputer at my fingertips. If any of you have done it another way, for example, writing scripts. When I was in graduate school, I was borrowing scripts from other students who may or may not have known what they were doing because as a research scientist, we were never taught how to use these. And that, of course, was problematic because that can create all kinds of issues. You're trying to figure out what's going on in this code. You're not even sure it's right. Sometimes you hold your nose and you hold for the best. And with Open on Demand, you don't have to worry about those scripts. You can just click a few buttons. There's some templates there and you can run your jobs on these systems and you don't even have to worry about what's going on underneath. With the Pegasus workflow tool, it's about creating or simplifying complex data workflows. So when you're doing any kind of job that you're doing, generally speaking, you're inputting data from somewhere. You're doing a bunch of stuff. You're making some calculations and then you're outputting that data somewhere. And with Pegasus, it can allow you to reproduce those workflows and then integrate with Open on Demand and then again, remove that complexity for the users. So that's just the 10,000 foot view of those tools. We're hoping that we can reach most of the users that way. But if we can't, we wanna make sure that a user doesn't have to spend time submitting tickets, waiting for our support staff to answer. I mean, they're losing time, right? Researchers have to focus on what's important for them. So we provide a knowledge base of information and in this knowledge base is documentation, tutorials, code templates. Some of this was ported over from the old Exceed project. Some of this is created by those resource providers, those people that have the large scale systems, those independently funded projects. Some of them are created by us and some of them are created by you all, the community. And we incentivize you to do that. I'm gonna talk about that in a second because it's super exciting. We also have a question and answer forum. People can go in and ask questions, a community can respond. But with the knowledge base, it's really lovely because you can go in, you can type in some keywords, get a response. If you're submitting a ticket and you say the word Python, for example, a bunch of articles about Python from our knowledge base will pop up. So it's really, it's there to help these researchers self serve. We also support allocations, people's ability to work with the systems. We provide some trainings to help people who choose to learn that way. And then we have affinity groups. And affinity groups are groups of people who have an interest in something. So maybe you have an interest in data management. And you form an affinity group, we do a lot of the logistics around it. We can provide a Slack channel, we provide methods for communication. And so you can report on news, trainings and things that would be of interest to that group. And then if you need the more concierge level support, that comes in with our Match Plus and Match Premier projects. So let's say you're a researcher who has some sort of computational need that you don't have the expertise for in your group. And that can be all kinds of things that could be converting code or porting code from one system to an access system. Whatever the case may be. And with our Match Plus projects, we have students who work with mentors who can fix those computational issues for the researchers. And so a researcher will come to us and say, hey, I have this project. Does it fit with your program? We'll say yes. We'll match them with a student who has similar expertise. We'll match that student with a mentor. Now the mentor could potentially do this work themselves, but it'll be slower with this path forward, but it's also promoting more knowledge transfer throughout the community. And we work with these groups and we support the students. We have money to support the students. We support the mentors through travel grants that we provide. And it really helps these researchers move forward further and further. And this is what's really exciting. You should definitely get that QR code or type in that link. I've talked a bit about our knowledge base and the fact that we have people who are contributing to our documentation. So you might have a social scientist or an engineering student or a faculty member in aerospace engineering, whatever the case may be, who has already written some documentation or is writing some documentation or delivered a training, something that would be of interest to the access community as a whole. They can, it's vetted by a committee, but if their contribution is accepted, it's going into our knowledge base, you can receive travel grant money up to, depending on what you do, up to $2,500. And this isn't just for students, it's for anybody, staff, faculty, anybody can apply for these grants. And it's not just about creating documentation, you can ask and answer questions on that community forum that I mentioned. You could join a committee that we have and get money for travel. You could become a mentor to one of the students that I mentioned. So this is a great program. This is part of the way that we're trying to bring in some of the underrepresented communities because first of all, as a student I'm always, or I was always looking for travel money as an early career faculty member as well, but also if you have a different background and you're writing documentation or you have something like that, you can, you know, and you put that into this access resource database, then you're helping other people who may have the same background as you do to further that knowledge. And then finally, I just want to mention some general opportunities to engage with us. You could create an affinity group. If it's of interest, you can connect with people all over the country. And we provide Slack channels. We'll provide a special question and answer forum in our ask.ci where you all could ask questions specific to your topic area. And, you know, you could have documentation in our knowledge base that would be connected there as well. You can join our mailing list and that'll get you information about our trainings and any other kind of news that comes out. We have a weekly newsletter. If you have an event that you would love to broadcast, you know, we could add it in there. And we, the community grants project that I just mentioned, we're always seeking mentors and students and we have, our committees are full right now, but of course that's always an ever changing experience. But, you know, there's plenty of other ways to engage to get some of that travel grant dollars. And you can always just email me and I unfortunately didn't put my email on the last slide, but I can hand it out as needed. So I'm happy to take any questions. We should have a few minutes for that. So thank you. Pop this up right here. Yes. Yeah. Yeah, so the question, I'm not sure if you all heard it, but the question was there was a group called the Campus Champions in the, that were part of XSEED, is the Campus Champions Continuing? And just for those of you who are unaware, the Campus Champions is this, they were supported, they had some staff supporting them through XSEED, and which they do not in access. And basically they're facilitators. So support staff or researchers or people who are trying to help people use computational resources at different institutions. And I don't remember how big that group is, but it's in the thousands. It's a big group of people. I'm a Campus Champion, and that means at the University of Colorado, if somebody comes to me and wants to know how to connect to the national resources, you know, I can help them. I can get them accounts on the systems to try some things out really low key. So to answer your question, the Campus Champions are still continuing. They're still an active group. You can still apply to become a Campus Champion. I think that the method for doing that is the same, and they don't have staff again to do that, but we have a mechanism for that. We were working really closely with the Campus Champions, trying to understand pain points. One thing that's come up is that they used to have the engagement events at PERC and the super computing conference, and so we're working together to try and continue those through access. The connections between access and the Campus Champions are growing and we're learning what they are, but we're very interested and we talk to the leadership team all the time about how to do that. So in short, the Campus Champions, they're still continuing, they're still going forward, and we're really eager to learn more and engage with them in ways that maybe they hadn't even thought of before. Does that answer your question? Okay, perfect. In the back there? Sure. Yeah, so the question was, can I tell a little bit more about the students who were funded as part of our Match Plus program? So the funding source for that actually comes from my projects. So we had written in money for stipends at least for this first year and then we're working toward with NSF and with other groups to figure out how to fund that and depending, it might still come from my bucket of money. That's besides the point. It's not, the onus isn't on the institution to fund the students for that. They can be undergraduates, they can be graduates, students. It doesn't really matter. What we do there is, if you had a student who potentially could serve in this way, they can go on our website, which is listed there and there's a place where they can apply to become a student and the way that it works is they get put into our database and they'll tag themselves with certain things. So I'm in the database and I'm tagged with, I don't even know what I'm tagged with anymore, but introduction to Slurm, I think like a scheduler or Python. And so then if person X has a research project that they also need Python and your student might come up as an expert in Python and then we would work with them to figure out, do you have the capacity to do this? Do you have the knowledge to do this? And then we facilitate a lot of the work after that. We have regular meetings with all of the students who are participating and with the mentors and with the research projects. And so these are typically about six month engagements to move forward on those. We're doing 10 pilot projects right now, but hoping to keep growing with that. Yes, sure, yep. Yeah, so the question was basically how's this project organized? And if I may, what is it that you do here? So let me step back a little bit to answer that and I apologize for not doing a great job at that in the beginning. So yes, so the access project has four service tracks. One is focused on allocations. So an allocation, you can get an account on a system. That doesn't mean you're going to be able to use a system. You have to say, here's what I want to use it for and here's how much time I need and then people can kind of guesstimate how many allocations to award to make sure the system isn't too overloaded. So there's an allocations team and they really support helping researchers write those proposals and figure out how to get on them. Then there's our support team, which is my group. We're working toward making sure the researchers know what they're doing on the systems. Then we have the operations team and so we have these 20 disparately, independently funded groups across the country. Pittsburgh Supercomputing Center has a system. The San Diego Supercomputing Center has a system. The University of Indiana has a system. There's all these different places that have systems and with operations, they can connect all those systems in certain ways but each one of them has their own system admins on staff. They have their own user support people on staff. Then we have metrics which is making sure that we're all doing a good job in collecting these metrics. So there's four PIs that work together. Then we have a coordination office in John Towns, who is the exceed PI. He's the PI for the coordination office. We all work together as equal groups to help promote these resources. In terms of if we've had previous expertise, I mean we all are based, I think all the PIs are running we're directors of Supercomputing Centers across the country but my group is not an access resource provider. So the Pittsburgh Supercomputing Center is an access resource provider. Like I said, they have expertise on staff but they connect with the bigger system. So they're doing a lot of things they're answering certain tickets but they're doing it through our ticketing system that connects everybody. They are, people can sign in to their resources but they do it through a single sign on that access operations supports. So each of us has our own funding. I have my own grant. The metrics people have their own grant allocations have their own grant, et cetera, et cetera. We meet every week all the PIs meet on an executive council to discuss access wide problems or not problems, great things. And our teams work closely together to make sure that this is all happening but it's much more of an ecosystem. We don't spend time on a day-to-day basis supporting the users. That's for the user support on those sites. We spend time building the structure and the framework that people can work within to get done what they need to be done. Does that make sense? It's a little nebulous, I know, but okay. I think we have time for maybe one more question if anybody has one. Okay. Well, thank you very much and I'll be around and feel free to email me if you have any more questions.