 Thanks everyone for coming to this next reclaim open online session. I'm joined here by Alex and Ben and We're gonna jump right into it. I'm gonna let them introduce themselves get started here. All right. Yeah, I'll go first I'm Alex. I'm currently a student at Dartmouth College and last year. I graduated from Skidmore College That's where I met Ben and that's where I worked on the project that I'm going to be presenting today and really excited to be here Hey, I'm Ben Harwood. I'm an instructional designer, learning experience designer, instructional technologist, wearer of many IT hats at Skidmore College in Saratoga Springs, New York. And yeah, it's great to be here with Taylor and everybody from Reclaim Hosting as part of the presentation. Let's move into slide one And a little historic context. That's something I think about every day and wearing different hats at my job. But, you know, if I had to create a quick snapshot history of web hosting In higher education because we're all doing this really similarly, I think across our institutions. So in the good old days, you know, we used to offer these tilde spaces Which is a web space for employees, faculty, students and staff at universities. And so the way we've been able to accommodate requests from the community members who say, Hey, can you guys set up an environment for me? Has evolved quite a bit over the years and it went from, you know, basically just giving them an Apache Lamp space on some server On the campus network and, you know, it was determined after rethinking that security wise that it probably wasn't the best way to do things anymore. So we have moved everything over to Reclaim Cloud now, which includes both our community WordPress and also we have the Domains Initiative as well. So, yeah, so we had Alex came to our group, Learning Experience Design and Digital Scholarship Support. It's a mouthful, but that's the name of our group. We're in the library. So we run Brightspace. We do. This is our recording space here, which is in Well, we're doing renovations here. We're putting some sound panels up. But yeah, we're a hub on our campus to try in the library, trying to help out a lot of community members. And Alex wrote to us with this project and it just seemed to be perfectly timed, you know, to have this new space and Reclaim Cloud. And yeah, so here we go. You want to talk a little bit about your project, Alex? Yeah, absolutely. Okay, so I was kind of struggling to come up with a name for this presentation because I want to talk about a bunch of things. But what I really want to show is kind of the goals that I want to have for this presentation. So what I want to do is be able to teach you something new about the role of the open web in physics research, present my work, so how I identified and solved a problem within my community, share my story as a student, learning how to share my creations on the open web using Reclaim Cloud, and then hint at the role that cloud services will have in emerging technologies at the end. So first we'll start off with like how this all came to be. So I want to start this off with a case study, something that I find really, really interesting. So a couple of years ago, humanity accomplished probably one of the greatest heats in human history. We took a picture of a black hole. And how that happened was the problem, the biggest problem there is that the black hole that we were taking was so far away, so many light years away that the analogy is to take a picture of it would be equivalent to taking a picture of like a nickel on the surface of the moon. So you need an extremely high resolution telescope, one that doesn't even exist. So how humanity came up with doing that is we turned the entire globe into a massive telescope by combining signals from one another. And having them all these different telescopes around the world and having them all work in tandem at once to essentially simulate one massive telescope. And that led us to take this super high resolution image of a black hole super far away. But all that data, there were so many, many terabytes and terabytes of data that it was actually faster to put all the hard drives onto planes and fly them all to a single data processing center because there just wasn't enough, there was the cloud infrastructure didn't exist yet to handle that much data coming in and out of the cloud and going across the world. So it was five petabytes of raw data, which ended up being put onto a thousand pounds or thousands of pounds of physical hard drives. So at Skidmore, I was doing astrophysics research, it wasn't nearly as cool as taking a picture of a black hole, but we had a similar problem. So our focus wasn't on black holes but was on large scale galactic structure. So we were looking at the cosmic web, which is kind of fitting the open web and the cosmic web. Basically, we can think of the entire universe as a web of massive galaxies and strands of matter that connect them. And so we were looking to use graph theory algorithms to try and simulate and try to calculate what the most likely large scale structures there are in the universe. And what this actually means is we had these massive Excel spreadsheets with galaxy data of coordinate locations and mass and different concentrations of different elements and we would run algorithms on that. So in the early days, how do we manage all these thousands of lines of Excel spreadsheets and how do we share them with each other? How do we collaborate? Well, we load them up onto physical hard drives, we pass them around, we had a whole fleet of hard drives, we had like a sorting system for tagging all of them. What could go wrong, right? I mean, everything's stored on physical hard drives, passing them around. What could possibly go wrong? Well, COVID happened. And we all went home to Zoom University. And I mean, what was the solution here? I mean, we could mail the flash drives to each other. We were using email as like a temporary solution to email different files at once. Yeah, so obviously something needed to change. And the obvious solution was to use Google Drive, right, to use any sort of basic cloud hosting service at the start. So Google Drive lets you share any repository, any folder and share it with people automatically. But is there such thing as something that's too general, right? Google Drive is a one size fits all. And is there a need and a space for a custom tailored solution just for our problem? Okay, so that's the guiding problem here is that data is really hard to deal with but really important. And what we were doing is we were uploading everything onto Google Drive and then downloading it onto our desktop. We would have the raw files on our desktop and you would literally go into your Python script and you would change the location of your script, make some changes to that spreadsheet, upload it back up to Google Drive, ping your coworkers and your lab mates. They would download it to their computer, change the name on the Python code, you know, for a constant collaboration and for fast feedback this just wasn't working. So that was the problem I identified and then I wanted to set out and build a solution. So I built the data hub, which is an API for data management tailored to the needs of the physics researchers and myself at Skidmore College. So I wanted to create a place that was a central store to have all of our data and it would be accessible by anyone at any time without having to download any files. And there had to be some way of programmatically getting the data so that you didn't have to get an entire spreadsheet which might be every galaxy in the universe. Let's say you wanted to only get the galaxies in a certain cone or a certain galaxy cluster. And then there also had to be a lot of things that come with a normal website like permissions, accounts, organization, collaboration, scalability and security. So how I built it is not really the focus of this presentation. I kind of want to show the impact and what Reclaim Cloud was able to do for me. But the whole point is that this was a totally custom built application from scratch. It wasn't something that fit into like a WordPress website or something that could easily be deployed on what the Skidmore was using before Reclaim Cloud. Because I built this completely from scratch, I coded it from scratch, I was using JavaScript and TypeScript. But really the important thing is what it actually does. So each researcher myself had an individual account and you could upload and download that data directly from Python by using an API. So essentially it was really just an API with some extra things sprinkled in like logging in and accounts and different blog posts for like an overview of how to use the website. So I built this website. I had it on my machine. I had it running in my terminal. And I was like, all right, well, it's done. It's ready to be deployed. It's ready for my physics professors to use it. It's time to find a home for it. And I thought long and hard and I was like, wait, I know exactly where I can put this. So because you can't have it running locally, you have to put it somewhere. How can I put it somewhere where my professors can use? And I had the perfect idea, which was the server box underneath my friend's bed stuffed in a box with all the wires dangling out. And honestly, I mean, it kind of worked for a while. It was port forwarded so anybody could access it. I mean, they'd have to type in like 68.28 and whatever. But obviously that wasn't a permanent solution. I just say I love this solution though as a as a student who did the same thing for a way less impressive project. When I was when I was in undergrad, it's you got to start somewhere and I bet you learned a lot about servers if you didn't already by even doing that. Yeah, certainly in the wintertime as well. I mean, you're getting a lot of extra heat there free heat, right? Yeah, exactly. I mean, my friend should be paying me for the heating service. Yeah, so essentially I was like, all right, you know, this is great, but I really want a permanent solution. I want something better. And my friend is bugging me about all the computing resources that this is taking. So I talked to my research advisor and she put me in contact with Ben and I told him I project. But we had a problem is that I built this completely from scratch using tools that didn't fit into the current architecture, because it wasn't a lamp stack. It wasn't all these acronyms that I had never heard before. It wasn't a WordPress website. It was completely custom tailored. So how do we host something that's completely custom? And the answer to that is containerization. And so this is really the magic of reclaimed cloud in my eyes is that Docker allows you to essentially package your application along with the operating system that it needs to run into a tiny little component called a container. And the container doesn't contain the operating system itself, but what it runs inside of is this think of it kind of like a virtual machine where you can have one operating system shared by multiple containers. And so that lets you create a custom application completely from scratch and handle the number one problem that every developer has of it runs on my machine, but it won't run there. Well, as long as you can make it run in a container, you can make it run everywhere. It doesn't matter what machine it's being hosted on as long as you properly containerize it. And so this gives you total freedom for student projects. It means that once you can containerize your project, it can literally go anywhere. And this is what reclaimed cloud actually looks like once you get into it. You essentially have I have my Docker engine here. I've got my database in one container and I have my actual application in another container and you can set everything up so that they communicate together. And it's just hosting itself. And this is accessible to anybody on the Internet at the IP that you can see there. And the coolest thing about reclaimed cloud for me is that the usage scales with how many people are using it. And so when nobody's using it, which is pretty often, it will use almost no resources at all. And then it can dynamically update how much CPU and memory and how many cloud, this is what it's called, is using based on how many people want to query the website. And so what that means is that I don't need to pay for the absolute cap of what I'm going to have it use. It will just scale to that. So the one or two minutes that my professor takes to upload a thousand galaxies, that's a lot of computing resources, but it's only for a couple of minutes. And so having the ability to scale like that means that not only can a student upload a project, but if it blows up into the next social media website or the next huge project, it will scale with it. You'll never have the problem of, oh, too many users are using it, it's going to go down. So what are the next steps with this project? Well, this project wasn't really meant to be a permanent solution. And so it's not going to be at Skidmore for the next 10 years or so. But it was a good project to kind of encourage my physics professors that, hey, there are other opportunities out there. There are other programs out there that are more custom tailored than using something like Google Drive. But what's actually cool is that I'm still finding uses for the data hub all the time. And right now I'm actually using it at Dartmouth in the new physics lab that I'm working on. This is a quantum computer right here. Well, sort of, this is the refrigerator that the quantum computer goes inside of because quantum computers have to be stored at almost absolute zero. Like this place, when it's cooled down, is one of the coldest places in the entire universe. But we want to see how hot the temperature is because, you know, what if it goes from five Kelvin to 10? Like, oh no, everything's going to go wrong. It's suddenly the temperature of the outer space universe and that means everything's going to get heated up and destroyed. And anyway, we are using the API to stream the temperature of the fridge into the API and then consuming that with a dashboard application that just is the front end for my API. But really what I also want to talk about is that cloud, reclaimed cloud, student projects that require custom architecture. That's all here to stay. And I have a presentation with Jim and Meredith in another session about my art fair piece where I go into detail about the smart microscope. After my introduction to using the cloud with Reclaim Hosting and the Skidmore Data Hub, I figured this idea of being able to have any scaling level of infrastructure that is just available to a little computer anywhere, like connecting to the cloud means that we can use it for all sorts of things. So this is the smart microscope. It is a digital microscope that is eventually going to use the cloud to run AI algorithms in real time on the images that you're looking at. And we're going to deploy those AI models in the cloud so that we don't have to have an entire GPU in every single microscope. We can have each microscope be weak computing. They all have Raspberry Pis, but they can all connect to a massive cloud data center. And once again, with something like Reclaimed Cloud, only when somebody is running those AI algorithms is the resource usage going to spike and then it will go back down to its baseline. And so we don't have to pay a million dollars a month to allow for that cap to be reserved at all times. And so that's kind of the end of what I had. I wanted to thank everybody for attending both the in-person session again when I was at Reclaim Open for anybody who's watching this recorded session. I really want to thank Ben Harwood for taking me under his wing and kind of, I mean, this was new for him too. So he could have easily just said, oh, we don't offer that, you know, whatever. But no, he took me under his wing and he said, we're actually partnering with Reclaim Hosting. Let's see if they have something. And then Taylor actually, I met with you when I was trying to get this set up for the first time. And you walked me through it and you got everything set up for me and you kind of introduced me to the platform. So I wanted to thank you. I wanted to obviously thank the Skidmore Physics Department for giving me a reason to build this in the first place and using it once I built it. And then Reclaim Hosting as a whole for making cloud hosting available and accessible to students. Thanks so much. This is such a cool project. And I think especially, I love how, and you touched on this, Alex, but I kind of love how this came together with Ben, like with, you know what, we don't have something that fits that, but let's explore what that looks like. That is my favorite kind of ed tech is that kind of exploratory. I remember, I don't remember up the top of my head exactly when the three of us met to kind of talk about cloud stuff initially. But I want to say it was like a year and a half ago, but I could be wrong. But it was, it was great to me to kind of see like, hey, you know, we have this thing we want to do, we think it would be possible. And, you know, my expertise is pretty, pretty specifically in the like, well, I can run things in Docker, but I am not a developer. So I don't know how to tell you how to ship your application in Docker really. So that's a question I kind of had is, like, what did that process look like for you? Did you already have it sort of running in Docker in some form, or was that kind of step one of like, all right, we're going to run this in ReClean Cloud. First, we got to get it in Docker. Like, is that, I'm just kind of curious. Yeah, let me, can I switch to sharing my screen? Yeah, absolutely. So just pull up, let me see if this, this is spontaneous. So hopefully this works out. But I will probably just pull up, GitHub is making me sick. Yeah, so what I'm going to show is essentially all that is required in order to at Docker working. So I'm going to load up Docker desktop, and I'm going to load up the actual code for the project. Yeah, and that's, and you did touch on this, but for folks who are watching this who are unfamiliar, Docker desktop would just be a way for Alex to run this on his local machine, which is kind of the great portable aspect of Docker. Like, like you said, you can run, once you have it running in Docker, you can pretty much run it anywhere, almost exactly the same way there. There's some exceptions in terms of some some applications require a little bit more configuration, but in general, it's pretty much exactly the same. So that is an advantage to and to be clear, not, you know, even ReClean Cloud specific, we try to make Docker really easy to use in ReClean Cloud, but you could take this and run it on any number of services. Yeah. And so this is my local Docker environment. And what's amazing is that if it works here, it'll probably work on ReClean Cloud. So right here I just have a database container. I don't have the whole entire Skidmore catalog here. But as a developer, what you do to get something containerized is you just describe in what's called a Docker compose, which is how you get multiple containers to talk to each other. So you essentially say, what containers do I need? What sort of applications is my full website going to use? In this case, I'm using the actual application itself, which is the Skidmore data hub, and I need the database. So that's Postgres. What's cool is that Docker has a massive image library of ready to use containers. So I don't have to do anything to get Postgres. I just have to say I want it and it will get it from Docker. This is my custom Docker container, which is the Skidmore data hub. And so the only thing you have to do is write something called a Docker file, which just describes essentially what your server is going to do once it is created. So literally all I do is install the dependencies onto the server. I run the build step and then I do some asset management and then I start the server. And I'm kind of hand waving it here, but as a developer, you essentially just figure out what would you do if you had to download your project onto your computer for the first time. You need to install your dependencies, build your application, run it, maybe mess around with where your various folders are. And then you just describe to the Docker container, OK, these are the things you're going to do and you have a container and from there you can have it work with other containers. You can have it run on different operating systems. And once you have that figured out and it's working in Docker desktop, then you're ready to deploy wherever you want. Yeah, and I can see in that Docker file in particular, because it's a Node application, you're able to even base it off of an existing Node container, which is great because making your own container for Node would be a lot. Well, maybe not that bad, but it's a lot of work that you can skip basically and build off of. So, cool, that's really awesome. I have a whole, I could probably ask you all questions, the two of you questions for a long time, but I'll try to keep it sort of brief. Ben, I'm kind of curious, sort of, your thoughts on like project life cycles and things like this. So, I know Alex, you said like, hey, this is probably not a permanent thing, like I worked on it, I'm gone now. I don't think anybody's picked up the mantle. And, you know, sometimes that's not really a problem, like that's not a bug, it's a feature in this case, I think. But Ben, like, what's it like to sort of, I don't know, create an environment that can support this work? Because I've definitely seen projects where it's like, well, if we can't do this for 10 years, we're not going to do it at all. Right, and I don't know if, I resist, I tend to be resistant to that line of thinking, but I'm kind of curious your thoughts on that. Yeah, I mean, these are, I mean, we're working in higher ed here. I mean, this is a, you know, for profit, you know, a business type project here. You know, I think having this recording is a really good case study for us to refer back to, and I'll refer back to, you know, Alex's project in the future. I mean, we talk a lot about, you know, is it a service that can scale, and I think it is. I mean, in terms of IT's hand here, we're providing an environment that is, you know, accessible to community members and, you know, drawing on synergies, collaborating with other people at the institution. You know, I would expect in the future, we'll get similar types of requests from, you know, the sciences, computer science department, just having an environment where, you know, students can tinker and learn with the understanding that, yeah, nothing's permanent, nothing's forever, but the service will stay, will continue to provide it. And I think it's, yeah, I mean, eventually we'll, I'm sure the site will take it down. It won't be needed anymore, but the proof of concept, the idea that this can be created is there and it's going to grow. And ultimately, it's about learning, right, both for Alex for you a little bit and kind of getting to terms with some of these cloud-based tools. But I think, again, you mentioned this, but I also think probably the physics department in general is saying like, look, there is, this is a workflow that is a game changer potentially. I'm curious if you have any insight into, you know, going from, hey, we're downloading spreadsheets and like running some Python code against CSV is very different than accessing an API, right? And luckily, if they're still using Python for that, like Python does have some nice libraries to make that easier. But I'm kind of curious if, was it a big lift to get like from like the students and teachers perspective to change over that workflow, or was it like, all right, well, we can kind of figure it out. Like we're doing Python code anyway. I don't know, I assume that wasn't like an overnight change, but I'm kind of curious if you have any thoughts on what that was like as observing it from your perspective. Yeah, so the beauty of being a developer is that you're also kind of marketing as well, right? Like not only do you just build an application, you have to make it work. That's the development side. But you also have to make it so that other people can actually use it and want to use it and make it easy to use. And so what I did is I created a bunch of wrappers around the API so that you essentially don't have to know that the API even exists, right? Like once you have code, you can make it as you can keep on adding abstraction layer on top of abstraction layer on top of extraction layer onto the code to the point where it's just like you can replace the line that says read CSV with just The same read CSV, but Alex is read CSV. And then underneath the hood, I can do whatever I want. And so I get to kind of choose how people interact with my application by giving them the wrapper or the facade or the contract and all these sorts of names mean the same thing. It's just how you interface with my application I get to choose. And so that means I get to be not only a developer, but also somebody who gets to talk to the professors and talk to different students and also work with it myself and say, what is the easiest way of doing this? And what that means is that by default, have it just be the simplest thing possible. You know, Alex's library dot read CSV, all they have to do is change it from pandas to Alex's library. And then also giving the opportunity for, okay, well, once you get comfortable with that, there are these advanced options that you can use such as writing your own SQL query to only get a to only get the data that you want. So let's say you just wanted to get galaxies around this cluster, you can do that. But that's for them to discover and figure out on their own once they're already comfortable with the default level. And so there's this whole entire like mind game kind of how do I make my product as simple as possible for people to use, but also have the extra options and the extra flavor that people get excited about and can add once they are comfortable with the basics. And that's, that's the beauty of a custom solution like that you built right is that you get to define that and say, I know exactly how we use this in class. And I'm going to make it a literal drop in replacement, you know, at least to start and then people can discover functionality from there if they if they want or need it. That's, that's awesome. Anything else you all want to talk about here we got a little bit of time, I am kind of curious about, you know, we did have you in the art fair session but how was the conference for you I'm kind of curious about, you know, it were a month out now from the in person experience and it was, it was fun for me to be a part of how was the conference for you all, how did you find yourselves there. Yeah, Ben, do you want to do you want to talk about your experience and then I can kind of talk about how I, how we got into this or actually how about you how about you talk about it from your perspective how how we got into this because I talked about it from my perspective in my other session but I want to hear what it was like for you and what was like at the conference to Well, yeah, I mean it was his first time I went to a reclaim open type event and it was it was just great to connect with, you know, colleagues and people I've known for a long time. And to really, you know, it was the first real in person conference I've been to since you know the pandemic, and it was, it was, it was really well run and you know really good sessions. So yeah, I look forward to coming to those events again and every once in a while unfortunate enough to bring a student along to this type of event and it was great you know Alex really, you know that was your first you know first conference and you got to go to an ed tech type of conference and so not exactly in your specialization area but you know I think it was pretty cool for you to. Yeah, do you want to talk about how how we got invited and like what that was about the whole thing with. I mean, so, you know, I think, you know, I can't really say from reclaims perspective so much but it sounds like this type of type of project is, you know, I'd imagine you're starting to support more of those, you know, for your, your customers right Taylor. Yeah, I remember when we were looking at you were on you, the both the two of you were on my short list of like I need to get, I need to get this at reclaim or a reclaim open because I want to see I love to see projects like this that are like we we built something exactly for our, you know, our class or community or whatever and so that yeah I wanted I think I sent you I think I sent Ben in email, I don't remember when it was but it was pretty early in my like, Have you considered submitting a presentation you know. Yeah, and what actually ended up happening was I did this project a year, like I was already graduated for a year I was at Dartmouth already. And my friend is Stan, who was one year younger than me. You know, Stan and Ben were working together. And I think you two both decided to submit or like we're talking about reclaim cloud about submitting my project or about this the data hub. And I had no idea that this was happening and so I went to visit Stan for his birthday. And out of nowhere he was like, Hey, you want to go to a conference. And I'm like, Excuse me, like, like what he was like, Yeah, we we submitted your project from a year ago to the to the conference or they asked for it or whatever and it got accepted and yeah you're going to go and here's the best part. As soon as I graduate, I'm leaving and going back home to Estonia so you're going to go on your own just you and Ben. And I was like, Okay, I mean I've never in my life have I just had such a cool opportunity just plopped on me. And so I think even that weekend I we I met up with Ben again, and we talked about it and we were like, Yeah, let's let's freaking do this you know like, you know, if it comes out of the blue like this, you might as well do it right. So I prepared a little presentation. And I was I didn't really know what to expect when I when I came to reclaim open but it was just, it was such a cool experience it was it was so great to see so many people who are so passionate about things that I didn't even know existed on the internet, because they just aren't in my sphere of what I focus on. And just even on day one seeing people just so excited to be there and so interested and so interesting and with so many stories to talk about and all these different technologies that people use and just seeing that the internet has affected so many people in so many ways, not just me and my friends the way we talk about it software development sphere there's an entire I mean obviously there's an entire world out there is an entire cosmic web of the internet. And it was so cool just seeing a totally different sphere of that, and being able to learn from people and then contribute to people and share my own, you know, teaching people about the way that it's used in physics research, the way that I'm doing it using it now for this microscope project. And then I had some really crazy like coincidental. Like, there was another doctor there who could have like, use the mic I don't know it was just everything you'd expect from a conference you know just crazy opportunities and just, you never know going in what you're going to get. Well, well, well said. I really appreciated having you at the conference and Ben I appreciate you having there, having you there too. And having both present and and coming back for the recap session so thank you so much. And thanks for everyone for for watching. Thanks everyone.