Innovation Heroes

TRANSCRIPT: The Real Superheroes Behind a Supercomputer

December 9, 2021

Ed 

[theme music plays] This episode of innovation Heroes is brought to you by the Samsung Galaxy Book. Visit shi.com/samsungmobility to learn more. Welcome to SHI's Innovation Heroes, a podcast exploring the people and businesses making a difference in our constantly disrupted world. I'm your host, Ed McNamara. [dramatic music plays] The first computers looked very different than they do today. They used to be these giant machines housed in huge rooms on university campuses, and only a small percentage of the population had access. Fast-forward to today, and we have computers that are so tiny, they can fit in our pockets, or on our wrists. But even with the innovations of the past several decades that have helped make personal computers smaller and smaller, supercomputers still exist. Advanced Computing, also known as High Performance Computing, can help researchers immensely in projects where a high degree of computational resources is needed. This is driving major advancements in research, data, AI, and other cutting edge use cases. Rutgers University is home to just such a computing cluster. But as is usually the case, it takes more than just great hardware to make the most of this advanced tech. Because if there's one thing that the evolution of personal computers, and really the Internet itself has taught us, it's that even greater things can come about when there's a focus on equity and teamwork - because teamwork is how we can best leverage tech to the highest possible level.

 

Barr 

So the supercomputer...it's a tool, right? It helps people be able to get work done. But in order for a tool to be useful, it's got to have the support, and the people, and the knowledge. And so absolutely, I would say the real heroes are the people.

 

Ed 

The pandemic has shown us that accessibility is going to have a very different meaning in the future when it comes to staying connected digitally. But there's a silver lining here. There's been an undeniable opportunity for expansion and advancement at nearly every level, including supercomputing.

 

Barr 

The majority of people who interact with a supercomputer are doing it remotely. So it's either from their office, from their home. You could be halfway around the world, you still can access it no matter where it is.

 

Ed 

Some of the people at Rutgers who put the super in supercomputer work in the Office of Advanced Research Computing, or OARC. They help to provide university researchers with essential computing, networking, storage and data handling capabilities. We're incredibly lucky that our guest today is Barr von Oehsen, and the Associate Vice President of the Office of Advanced Research Computing, joining us today from his office at Rutgers University in Piscataway, New Jersey. The projects that Barr and the OARC team get to help with are pretty out of this world, literally, for some of them. They're paving the way for best practices by leveraging some of the most advanced computing capabilities. But they're also making sure that there's engaged teams at the helm providing equitable access and fostering true collaboration. Because sometimes it's the most basic things that could be the most important. Welcome Barr, thank you for joining us on Innovation Heroes.

 

Barr 

Thank you, Ed. I really look forward to the conversation.

 

Ed 

As do I. [music fades out] So I know you have a really big, multifaceted job, so I'm not even going to try to explain it myself. [chuckles] I'm going to simply ask you, what's the elevator version of what you and the OARC do at Rutgers University?

 

Barr 

So honestly, I struggled with this all the time, because in my role at the university I speak with, you know, the Chancellors, I speak with Deans, I speak with researchers, and I speak with network engineers. And everybody always asks me, "So, can you explain what you do?" And so depending upon who I talk to, I've got a different explanation. [laughs] But because I report both to IT and to research, in a lot of ways this office fits right between those two worlds. So we spend a lot of time building out infrastructure in the support of research. And our goal is to enable researchers to be more productive in getting their research done, and to be more competitive in getting their grants. And so in a nutshell, that's what we do.

 

Ed 

How do people find you to even start having those conversations?

 

Barr 

So when I when I first started at the university, I actively went out and knocked on people's doors. Now, Rutgers is a huge university. So we've got 70,000+ students, we've got 25,000 faculty and staff, and we were a fairly small group when I first got here. You know, really just five people for such a large university. But, you know, we all got out and we had conversations, and I didn't really want to build any, you know, technology solution for anybody until I really understood what the needs would be. And-- because, you know, there are a lot of people that have the mentality, "I will build it and they will come." My feeling is, I want to understand what their needs are so that I'm guaranteed that they will come [laughs]. Because it doesn't always work out that way. So because I work with central IT, and I worked with the research office, we were able to tap into their listservs, and, you know, send out email to people across the university. But a lot of it is word of mouth. You know, at first, when we started, we did just a few people. But over time, it's grown. And now we're starting our sixth year and it's just amazing how many people come to us asking for our help. We've got well over 3,000 people using our services now and attending our training sessions. We've got over, you know, 300 different departments, and centers, and institutes, you know, sitting in and going through all of our training. So, word is getting out. And because of that, we've been able to also grow our team to help support a much broader community than when we first started in 2016.

 

Ed 

I do appreciate that there's a marketing aspect to what you do here. So as my mom says, I also work in computers in a very different way than you do. [Barr laughs] Once you find these people and they and they come to you guys, what are you helping them do?

 

Barr 

So it depends upon their area of research, right? So. So for a majority of the people that we have conversations with, most of it is around data. Data is everywhere, and the research community is interested in looking at that data. And so it could be medical data, it could be marketing data, it could be digitized collections within libraries, genomics data, it's just-- and it's growing at such a phenomenal rate that people are really struggling on how to be able to manage all that data, and how to be able to do the analytics, and the statistical analysis on that data. And so we spend an awful lot of time having conversations with several groups across the university on strategies on how to manage that data, and how to be able to do the analytics on the data. Now, with that being said, artificial intelligence has just exploded over the last few years. And I mean, you just watch TV now, right? I mean, all of the ads, they all talk about, "AI this, AI that." And it's true in the research world, too. So now they're not only interested in, you know, figuring out how to manage all that data and how to do the statistical analysis, but now they're trying to train AI systems on all that data. And that gets really challenging on how to move the data around, how to access it. You know, it's just a very challenging space to be in. And so we spent quite a bit of time working with people on how to go through that process, and what are the tools that are available, and help people be able to get access to the tools. And, you know, the team-- the part of the Office of Advanced Research Computing, the Research Science and the Technical team, they have skills that enable them to go and have those conversations and help sort of lead people through and make decisions on what are the best workflows to come up with in order to get their research done. [music plays]

 

Ed 

So when I think about all of the people who could benefit from having super computing power to help them analyze data, there must be so many amazing projects that you get to be a part of, or at least help people work through. Do you have one or two of the coolest or most surprising projects to have come your way recently?

 

Barr 

So as I mentioned, we work with everybody. And you know, every time I bumped into a new researcher, they tell me what they're doing, and it's just so way cool [laughs]. I get excited about it. [Ed chuckles] We work with everybody, but, you know, just a couple of examples. One is-- so, libraries. We just submitted a grant with libraries. So their collections, which typically used to be in books, and people would go in there, they're all digitized. And they want to be able to make them much more accessible to the general public. And so we're working with them on creating sort of an AI interface to access these collections and allow people to go through and interact directly with them. And it's actually fascinating just to have a conversation with them about what they're trying to succeed in doing, and then how do we train these AI systems, and all that. It's just-- it's fascinating. And so that was a lot of fun to work on. Another one was in dentistry. What I found out was that when they put an implant in the tooth, it actually creates a lot of stress on the jaw. And when you chew, it actually hits that implant and then spreads it out. And they did a full model of just the implant. It's fascinating. It just, I love just having these kinds of conversations with people because, you know, this is stuff that I just don't think about all the time. And then when I sit down with them, I go, you know, "That makes complete sense." And, you know, "How can we help you with, you know, trying to work through some of these problems?" So, it's a lot of fun.

 

Ed 

So obviously, this show is called Innovation Heroes and going in I thought maybe the hero was going to be, you know, the supercomputer, or the super computing power that you have at your disposal, but they haven't even talked about that. [chuckles] I'm really getting the impression that there's a lot of human muscle behind the scenes that keep the whole thing running. Like, who are the real heroes here?

 

Barr 

So the supercomputer, you know, it's great, and it helps...it's a tool, right? It helps people be able to get work done. But in order for a tool to be useful, it's got to have the support, and the people, and the knowledge. And so absolutely it-- I would say the real heroes are the people who are working very closely with the research team. So within OARC, the way we've got it set up is we've got the technical people, so the technical facing people, the people that build out all the infrastructure, support it, and make sure that it's running 24/7, 365. And, you know, with research, you never know what you're working on, right? Because it's cutting edge. And, and so the fact that this team is able to work very closely with the research community, and come up with solutions, and build, and support, it's just phenomenal. And then we've got the research facing team, so the research scientists. And they all have, you know, the educational background, most of them have PhDs, but they've all done research themselves. So they really understand how to get research done, how to write papers, working with students, but they also understand the technologies. And so they're the ones that sit down very closely with the research community and understand what they're trying to do in order to compete and complete their research projects. And then come up with the design of workflows that that work best for those research groups. And what we're finding now is that the research groups are looking at these people as an extension of their own research groups, and they're being included on grants now, they're being included in meetings. And so everybody is really looking at these team members as being essential to supporting research at the university. So I agree, I think they are definitely the real heroes in getting this done, and supporting, and coming up with novel ideas and solutions to support the research community. I mean, the worst thing you could do is say no to a researcher, because they're going to find some other way to get it done. And a lot of times the solution they come up with, it's not a very...it's not the best solution [laughs]. And often offers security holes. And so we don't want to do that. And so the team is really prepared to have conversations and say, "Look, you know, here's what we can do now. Let's think about future funding opportunities that could lead us to exactly where you want to be." And so far we've been extremely successful in doing that, so...and the trust from the community that that's, that's the big thing, right? You know, when I first started in 2016, everybody looked at us and said, "You know, you're not gonna-- you're not going to last." [Ed and Barr laugh] But-- just because they sort of know the history of Rutgers. But the team has gone out, and they've built trust in the community. And that's incredible, right? When you when you build that trust, and now we've got people coming to us and asking for our support. And so it's been a great ride so far.

 

Ed 

This episode of Innovation Heroes is brought to you by the Samsung Galaxy Book. Visit shi.com/samsungmobility to learn more. [music plays] We've all had to adapt to different ways of working, living, and playing, and we know more changes lie ahead. So in order to keep up, you need to be able to do more, and do it better. The Samsung Galaxy Book is more than a laptop. Loaded with features that help you look, sound, and do your best, it offers true mobility so that you can work comfortably from home, the office, and everywhere in between. Plus, it's designed for a worry free workday, with the security, durability, and protection you need so you can keep your mind on your work without any distractions. Built to be the brightest, lightest, and fastest, the Samsung Galaxy Book features 100% color volume, and an incredible 1 million-to-1 contrast ratio. And thanks to its lighter than ever slim chassis and 11th Gen core processor, heavy lifting is made easy so that you can make quick work of work. Start doing more and doing it better today. Check out the Samsung Galaxy Book at shi.com/samsungmobility. [music fades]

 

Ed 

[music plays] Frankly, I'm kind of struck by the diversity of the groundbreaking projects Barr encounters on a daily basis. But the OARC faces the same problems we're seeing across our industry: for instance, talent hiring and retention. [music fades]

 

Ed 

I'm thinking there are IT managers out there who are envious of the team that you just talked about. It seems like you have a group of highly skilled IT professionals who can quickly identify the practical application from a research perspective of the computing power that they provide. So how do you identify and cultivate these pretty special human beings, and get them into a team that supports that, you know, one, the community that you talked about, but then, you know, develops, you know, a trust with the community that they're supporting?

 

Barr 

So it's challenging, right? And there are groups at universities all across the country who are in the same situation. We find it very challenging to find people who are able to step right in and be able to take on these kinds of roles. So it's not your traditional, you know, "I know how to stand up a server, I know how to set up a Cisco switch, I know how to..." It's not that. They have to come in, and they have to really have a deep understanding of lots of different types of technologies. So that could be advanced networking, so, you know, how do you build out the network capacity to be able to move all the data, or get access to the data? It includes data storage, it includes the supercomputing, it includes cloud services, and also communication skills. We can't stress that enough. [laughs] Being able to communicate and understand what a researcher is going through is most of the work that we do. It's just having that ability to connect with a researcher, with the students, and really understand what they're trying to do, and then come up with solutions. So when we put out job openings, when we announce them, we may only get about five people that that apply to them, because it's such a specialized area. And so this is something that we as a community have been struggling with and trying to come up with solutions. Another one that we struggle with is diversity. So typically, this community is not a very diverse community. And so we've tried to come up with better ways of figuring out how to be much more inclusive in training people. And part of the problem is that our area of specialty is never really seen as a career path for people. So when they go through, you know, high school or college, they don't think, "Gee, I'm going to be supporting supercomputing, or building out next generation technologies, or offering core services." And so we have to be better as a community to educate people on the fact that this is a decent career path, and an exciting career path. And so we've been having conversations with a lot of the community colleges within the state, and with the four year colleges, and trying to come up with ways that we can sort of build pipelines. You know, sort of bridging the gap between some of the areas that they're focusing on within the university system, to the new technologies that we're trying to push out. And what we're finding is that industry is suffering from exactly the same thing. And so often  we find people that have the personality, the skills, and  we train them on our services, and then industry comes in and offers them a better job. [laughs] But you know that that gives us another connection into industry. But it's-- you know, we're trying to figure this out. And it's a challenge, but just the conversations I've had recently with a lot of the different schools around New Jersey, I'm pretty excited about creating some of these career paths for students who might be interested in taking on this role.

 

Ed 

In our post, you know, pandemic hybrid world, has there been a shift into the way we think about accessibility? And, you know, basically, how does that play out for something as big as a supercomputer? And how do you make sure that, you know, people are getting the accessibility that they need to in order to do their research?

 

Barr 

So the nice thing about supercomputers is they're always accessed remotely. So you don't have to be sitting in the room. In fact, you wouldn't want to sit in the room with it, because they're so loud. When people are running their jobs on a supercomputer, the sound of the processors just humming along is so loud that you actually have to wear earplugs in the data centers to be able to be anywhere close to them. And also to cool them, the room is freezing. And so, you have to wear a winter coat  if you want to spend any time in there [chuckles]. And so the majority of people who interact with a with a supercomputer are doing it remotely. So it's either from their office, from their home, you could be halfway around the world, you still can access it no matter where it is. And they don't really even care where it sits. But the interface to the supercomputer is what's changed. So originally, the way people interact with a supercomputer was that you bring up what's called a terminal. So, you know, believe it or not, our laptops and desktops still have that capability, and you could still use them. But most people have stepped away from that. And so most of us have now switched to these virtual desktop environments that allow people to interact directly with a supercomputer. But again, you know, it doesn't matter where they sit, as long as they've got a decent network connection they can fire this up. They've got all their applications as buttons on the desktop, and, you know, we try and give them the iPhone experience, right? They've got a little icon on their desktop, they can click on it, it brings up the application they want, and then be able to run the statistical analysis on it. So ,that's never really changed. What did change because of COVID was people who tend to be in wet labs or, you know, labs on campus, who run experiments where people actually have to be in the room-- and, you know, they've got the gloves on, and the hoods, and everything going, they run these experiments-- all of a sudden, were sent home to work remotely. And as you know, you just can't run experience experiments remotely like that. And so we had a surge of people who switched from running these experiments in their labs to actually using modeling and simulation. So they would use, you know, these equations to help define the experiments that they were running, and then be able to set up the simulations within our supercomputer, and then run the job and get results. And they would be able to compare with past experiments to see how accurate they were. But a lot of people who run experiments just weren't familiar with our services. So we did spend an awful lot of time that the beginning of the pandemic to help people transition from desktop to, you know, experimental, you know, experiments in the lab, to being able to use a supercomputer. So that took extraordinary amount of time. The team did a great job. And again, this goes back to your heroes, just stepping up to the plate and just being able to work through all of this. [music plays] And I mean, just the number of emails and phone calls coming through, it was just phenomenal. And they were able to keep up with it and help people through the situation. So, it was amazing. So, absolutely couldn't they were the heroes.

 

Ed 

For some of the real techies in the audience, you-- it just occurred to me, you must have, you know, somebody who's using more resources far and away than anybody else had previously. You know, in terms of the computing power that they need. Off the top of your head, can you recall, you know, what that project might entail?

 

Barr 

Certainly astronomy. You know, they're modeling galaxies [chuckles]. And, you know, that just takes an incredible amount. Anybody who's doing physics tends to take up quite a bit of compute time. So astronomy, physics, those are huge. Anybody who's doing weather modeling, right? I mean, weather is so complex, and it's gotten even more complex  with the change in the climate, and everything's so unpredictable. And the number of variables that have to be fed into these projects are just massive. And so for people that really need that kind of computing, we actually help them get onto national supercomputers that are, you know, a magnitude or more faster and more powerful than anything that we can provide here at Rutgers. So if you look at the types of jobs that we run at Rutgers, it's sort of the small-to-medium size, but those really, really large ones, we know the research labs across the country really well. We have partnerships with them. And so when somebody, you know, we work with need something that's much more powerful, then we'll help them put in an application to get access to those national resources. And so it's actually, it's really a nice setup, because, you know, we're helping people transition from the desktop into our supercomputer on campus. And then people's needs grow, the data grows, and all of a sudden, you know, they've learned everything they could possibly do on our system. And then we help them transition to the larger system. [music fades out]

 

Ed 

Barr, there's so many different aspects of what you and your team are doing. My last question to you, what makes you excited to go to work every morning?

 

Barr 

It's just working with so many different research groups and knowing that we're helping and making a difference in the way people are getting their research done. And how productive, and hopefully, they're becoming more competitive with getting grants because of some of the services that we offer. You never know what's going to happen day to day, right? I could come in the morning, and I have my plans laid out for everything that I'm going to do. And I get one email, and it changes everything. Some people describe our office as a startup company, right? So, we've got the makings of a small company with all the technologies and all of the expertise on all levels, right? We've got a business office, we've got, you know, research scientists, we've got, you know, research IT people, we've got, you know, system administrators that, you know, all work together to support the university. [music plays] But, you know, that being said, our office is often described, as, you know, we're building the plane that were flying at the same time. So it's a bit crazy [laughs]. And finding the right people that can then work in a crazy kind of environment like that is always a challenge. So that adds to that, you know, how do you find the right people? But it's exciting.

 

Ed 

Barr von Oehsen, Associate Vice President of the Office of Advanced Research Computing at Rutgers University, thank you so much for your time.

 

Barr 

Thank you very much.

 

Ed 

I would go as far as to say it's not just a supercomputer that they have at Rutgers. There's a real community that's being built. And I think there's a lot to be learned here. Barr and his team are giving business leaders an example of how to efficiently manage these kinds of high-powered resources with teamwork and innovation. And even in the shortages of our post pandemic world, at least these things are never in short supply. [music fades out]

 

Ed 

[theme music plays] Thanks for listening to this episode of Innovation Heroes. Next time on the podcast, I'll be speaking with Alexa Pavliuc, a doctoral researcher at the Oxford Internet Institute. She uses data visualization to analyze communication networks. In other words, she helps paint a bigger picture of the interactions in our digital communities. She's using her research to find ways to make the online world a safer, more welcoming place. We're taking a break for the holidays, but we'll be back in the new year, so tune in again in 2022. You won't want to miss it. If you enjoyed this episode, then consider being our hero. Smash that like and subscribe button to Innovation Heroes wherever you get your podcasts. Innovation Heroes is a Pilgrim Content production in collaboration with SHI. Our producers are Tobin Dalrymple and Jessica Schmidt. With production assistance from Carmi Levy, Jane Norman, Ronny Latimore, and Amanda Scheffer-Cavanagh. I'm your host, Ed McNamara, and I'll be back with more amazing stories in 2022. See you next year! [music fades out] This episode of Innovation Heroes has been brought to you by the Samsung Galaxy Book. It's not a laptop, it's a Samsung Galaxy Book. Visit shi.com/samsungmobility.

Podbean App

Play this podcast on Podbean App