The New Stack Podcast

From DB2 to Real-Time with Aerospike Founder Srini Srinivasan

Episode Summary

Aerospike Founder Srini Srinivasan had just finished his Ph.D. at the University of Wisconsin when he joined IBM and worked under Don Haderle, the creator of DB2, the first commercial relational database management system. Haderle became a major influencer on Srinivasan when he started Aerospike, a real-time data platform. To this day, Haderle is an advisor to Aerospike. "He was the first one I went back to for advice as to how to succeed," Srinivasan said in the most recent episode of The New Stack Maker series, "The Tech Founder Odyssey."

Episode Notes

Aerospike Founder Srini Srinivasan had just finished his Ph.D. at the University of Wisconsin when he joined IBM and worked under Don Haderle, the creator of DB2, the first commercial relational database management system.


Haderle became a major influencer on Srinivasan when he started Aerospike, a real-time data platform. To this day, Haderle is an advisor to Aerospike.


"He was the first one I went back to for advice as to how to succeed," Srinivasan said in the most recent episode of The New Stack Maker series, "The Tech Founder Odyssey."


A young, ambitious engineer, Srinivasan left IBM to join a startup. Impatient with the pace he considered slow, Srinivasan met with Haderle, who told him to go, challenge himself, and try new things that might be uncomfortable.


Today, Srinivasan seeks a balance between research and product development, similar to the approach at IBM that he learned -- the balance between what is very hard and what's impossible.


Technical startup founders find themselves with complex technical problems all the time. Srinivasan talked about inspiration to solve those problems, but what does inspiration mean at all?


Inspiration is a complex topic to parse. It can be thought of as almost trivial or superficial to discuss. Srinivasan said inspiration becomes relevant when it is part of the work and how one honestly faces that work. Inspiration is honesty.


"Because once one is honest, you're able to get the trust of the people you're working with," Srinivasan said. "So honesty leads to trust. Once you have trust, I think there can be a collaboration because now people don't have to worry about watching their back. You can make mistakes, and then you know that it's a trusted group of people. And they will, you know, watch your back. And then, with a team like that, you can now set goals that seem impossible. But with the combination of honesty and trust and collaboration, you can lead the team to essentially solve those hard problems. And in some cases, you have to be honest enough to realize that you don't have all the skills required to solve the problem, and you should be willing to go out and get somebody new to help you with that."


Srinivasan uses the principles of honesty in Aerospike's software development. How does that manifest in the work Aerospike does? It leads to all kinds of insights about Unix, Linux, systems technologies, and everything built on top of the infrastructure. And that's the work Srinivasan enjoys so much – building foundational technology that may take years to build but over time, establishes the work that's important, scalable, and has great performance.

Episode Transcription

Alex Williams  0:00  

Hey everyone, Alex here. Hey, listen, we have a new series starting on the new stack. Baker's is called the tech founder Odyssey. It's about technical founders. So as people who are engineers who are starting their own companies, we've had the best feedback from people already on Twitter. The questions are awesome. And it's really going to be guiding us. As we develop this series. We want to know about these people why they got started, what motivated them? What were the lessons that they learned along the way? What is it that they've learned about themselves, about their backgrounds, about the people who they've hired to work with them? How have they grown this venture into something that will last it's a big deal for these people. And I think it's a big deal for a lot of us who thought about doing these kinds of things, and are part of the companies that these engineers have started, we hope you enjoy it. Again, it's the tech founder Odyssey on the new stack bankers.


You're listening to the new stack makers, a podcast made for people who develop, deploy and manage at scale software. For more conversations and articles, go to the new stack dot I O. All right now on with the show.


Hey, everyone, welcome to another episode of the new stack makers. today. My guest is 3d Surrey vasan, CTO and founder at Aerospike Srini. It's very nice to see you today.


Srini Srinivasan  1:35  

Thank you, Alex, the pleasure to be here. Hi,


Alex Williams  1:39  

first would just like to know what you do at Aerospike. I know you're a founder there, and but you're also the CTO, maybe you could tell us about Aerospike a little bit, then we'll get started in just talking a little bit more about your experiences.


Srini Srinivasan  1:52  

Right? I have a background in database technology from a long time ago, after I worked in some large companies initially building database technology, I worked for almost a couple of decades, in systems that used databases, mostly in interactive TV and mobile applications. You know, think of the kinds of things that you see today on your iPhone, we were building those in the early 2000s, before they became mainstream. During that process, I discovered that traditional databases couldn't handle these high scale, real time applications for the internet and mobile users. And that's what started this journey to build Aerospike as a high performance, and low latency database that is able to service, you know, hundreds of millions to billions of users. And that's kind of how it was like came about?


Alex Williams  2:52  

Well, I would love to talk about the architecture at some point. Maybe we'll get to a little bit of this conversation. But more so I'm really curious about those years where you're building those components. And what were those experiences? Who are the people that you met that you remember now, who helped you discover some of these insights, for instance, in thinking through low latency and scale,


Srini Srinivasan  3:20  

if I look at my some of my earliest experiences was at IBM, just after I completed my PhD from the University of Wisconsin, I joined IBM, in an organization run by an IBM fellow named Don had a lead, who was basically the father of IBM's db to relational database. So this organization was in between research and product development, called the database Technology Institute. And what I learned in that initial set of products that I worked on, continues to this day, you know, it's it's pretty much shaped a lot of my career, we build systems before they became mainstream. And then we pushed it into the product. So that's where it was in between research and product development. So we could take the latest technologies and push it into the product development space. And that's essentially started off my career. Over the years, I have met amazing people when I moved to startups, people who essentially, you know, are so good that even a group of about five to six people can build an important product, which is used daily by millions of users. That's kind of an experience I had in various startups. And especially one of the startups I was in had a founder, who basically believed that anything was possible. While that might sound so what I would call impossible in some sense, but what it gave me was many of the things that I used to think were impossible. I know And my experience was just very, very hard. There are always a few things that are impossible. And the trick to building the kinds of companies and the kinds of products that I have been working in over the years, is to isolate the impossible, from the really, really hard problems. And the really, really hard problems can be solved by the kinds of people I've met with in my career, you know, starting with handily, but they've been a number of people over the years, who have influenced me, the interesting thing why brought down heavily up was when we started Aerospike, he was the first one, I went back to, for advice as to how to make this company succeed. And he's been an advisor to us for the entire life of this company.


Alex Williams  5:43  

How did you get to know him? How did that come about?


Srini Srinivasan  5:46  

Well, I don't didn't know him because I joined IBM, and then I quit. So until then, I observed him in IBM. And essentially, I had a very interesting conversation done before I left IBM to go to a startup, where I was actually, you know, much younger than, and a lot more impatient, I would say, than I am today. So I was essentially very frustrated, speed of pace of what was happening, and so on. And I still remember the conversation at Davos dollars really smiling at me and laughing and going like, Don't worry, see me you've taken an interesting step forward. And I think you will figure out what's actually the right thing to do. And what they told me was two things. One is, there's always a time, when an engineer like myself, who wanted to challenge myself, I would have to go out into areas which are uncomfortable, that is trying new things. But also, what Don and team had built in IBM is, interestingly, what I'm trying to build at Aerospike today. So there is this cycle of things, people have to first go learn how to do things. And then they can do something, hopefully, really consequential again, but the people I talked about, you know, it's all just people do this by themselves, not at least in the kinds of work that I do, we really need to work with extremely smart and competent people and capable people who come up with new ideas and be able to work and collaborate together to build a better system than what each of us do by ourselves. Right. So that is kind of the journey, which I really enjoyed, thankfully. So the years. What did


Alex Williams  7:18  

people like Don, you know, in other CMAT, teach you about being part of a team, and in essence, now leading a team,


Srini Srinivasan  7:27  

to some extent, when I started off, I never really wanted to lead a team, I wanted to be an architect. And I wanted to build leading technologies obviously work with people collaborate. But leading a team was the last thing in my mind. In fact, I would make it very clear that I wouldn't want to manage, I wouldn't want to do these things. But very quickly, I realized, because of the kinds of work that I was doing, essentially trying to bring to mainstream technologies, or applications, or systems that are at the state of the art, so to speak, nobody else is going to actually manage a team to make it happen once we invented it. So you would have to roll up your sleeves, and get into the nitty gritty of managing people figuring out how to inspire them how to solve problems, not just for our team, but also deal with customers, with partners and so on. So there is a level of management that just creeps in. And I kept going from company to company, and every time I would start as an individual contributor, a few years later, I would learn to everybody that I would have to manage. So eventually, that's kind of how we came to it, it is almost a necessity. But there is something also even more important, because what you cannot do as a single person, you can do it as a team of people, right, you can solve bigger problems, essentially, world class problems can be solved by a small team of world class people. And in the case of Aerospike, it's not just Aerospike, which is wonderful as we stand on the shoulders of giants, as they say, the work that has happened billions of dollars of investment from storage companies like Intel and micron and Metis. Or Samsung, for example, also processors that have been invented, and researchers to the years who have invented technologies and distributed systems we have built on top of that. We've also invented our fair share of algorithms, which have been patented, over a dozen patents. So we basically all of this comes together to building a system which can handle the kinds of requirements of real time applications. And that's kind of how the team works is you're not only worked with the team that you're working with right now, you're leveraging working with teams outside the area you're in on which you depend on and all of that pretty much boosts the technology industry today.


Alex Williams  9:45  

You mentioned two things in particular that I picked up on. You talked about being a contributor to a team and eventually becoming experienced enough and working on that project to become a lead. I would expect you all also talked about the need to inspire. And clearly Don is a person who inspired you, what have you learned about inspiration, and what's your thoughts on it and your approach to keeping people motivated and excited about their work?


Srini Srinivasan  10:21  

The first thing, what inspiration is honesty, is to actually face the situation that you're in front of, without fear, obviously, there's only some fear. Not having any fear is also not healthy. Definitely complete honesty is really important. Because once one is honest, you're able to get the trust of the people you're working with. So honesty leads to trust. Now, once you have trust, I think there can be collaboration, because now people don't have to worry about watching their back and so on, right, you can make mistakes. And then you know that it's a trusted group of people. And they will actually watch your back all of those bases. And then with a team like that, you can now set goals, which seem impossible to start with, like I was saying earlier, but these are hard goals. But with the combination of Honesty and trust and collaboration, you can lead the team to essentially solve those hard problems. And the interesting thing about inspiration is, you inspire everybody to give their best to solve the problem, you're actually focused. And in some cases, you have to be honest enough to realize that you don't have all the skills required to solve the problem, and you shouldn't be willing to go out and get somebody new to help you with that. And all of that is part of the inspiration, you know, setting the goals, and then creating this trusted environment with brutal honesty, which enables people to collaborate and thus, get to the basically the goals that we set, which is kind of the inspiration part is eventually you look back and you go like, why did you actually do that? And that's the level at which you can lead teams. I've learned that over time. And I've seen don do that in my short stint at IBM, but then I hope I haven't been able to do that earlier. In the last 12 years.


Alex Williams  12:20  

How do you set that up for teams at Aerospike? I don't know how many engineers you have. But it comes down to in many ways, your own vision, the approach to new features and capabilities. I was listening to an interview the other day where they talked about engineering management and talking about crazy goals have been crazy goals, you know, are often thought of as impossible but at least one manager in this discussion said, Yeah, but you can turn it around and say, well, let's put our minds that actually how we would do it. What would you need to get that done? Right? So I'm just trying to get to the idea of like, how do you even approach developing this honest culture, so to speak? And what are you doing building so when people walk in that door to interview, you're almost like making them part of this process to see if they're going to actually work out? All right.


Srini Srinivasan  13:15  

I think the fundamental thing about this as I work in software, which means it's really 100%, about the people. So when you're trying to actually solve a hard problem, you better make sure you have a set of people who are capable of doing that. That is the number one thing. Now, in terms of the mechanics of how you go about doing it. I'll give you an example of what we went through maybe four or five years ago, we had various I would say inflection points, in terms of how our product evolved. The first one was a hybrid memory architecture, which myself and my co founder, Brian Bukowski worked on the first couple of versions. And we basically built out the system with a very small team of people who joined a couple of years after we started the company. The important thing is the two of us solve problems because it was easy. They were just two people and the problems we can solve is about what can be solved by a few people. And then we started setting up more ambitious goals. You know, we said we're going to become a database. You know, if you've listened to Mike Stonebreaker, who is one of the Great's of database systems area, he's a Turing Award winner, and he's built like six or seven days complete. I don't even remember how many of those. But the point is, it takes about a decade to build database technology from his during our talk, you mentioned about how it takes about 10 years for technology to build infrastructure like this. So I mean, we knew when we started, I actually had a feeling about it, that it takes a while. But then I heard Mike stock once when he was and remember where exactly it was, but he mentioned it and I was about at the five year mark For a company that was really good timing. So went back to my team and said, Look, we know it's gonna take us a while to get this done, we only about halfway. So we're going to build a feature called strong consistency into a little bit, we were initially focused on performance at scale, which we wanted to maintain, as we added strong consistency, think of money transfer, 10, transactions and so on, when you cannot even use a single bit, people be very unhappy, they will do $20 that the banks thought they would do $200. I mean, that's resulted in them removing gasoline, moving the bank account, right. So strong consistency, by itself, it has been the problem has been solved a long time ago. But we set ourselves the task that we have this system, which is really fast, can we maintain the speed? What are the trade offs that we have to make? Is it even possible to do that? We postulated it was possible because we had to do that. Because if we didn't become a database, our five years of work is not going to be useful anymore, right. So we set ourselves a challenge. And, frankly, we didn't know where the breakthroughs were going to come. But they came because we have the right people for basically sweated it. And some of them, one of them actually especially came up with a very clever solution, which enabled us not only to add strong consistency, but also maximize availability, it says system doesn't go down a lot when you do rolling upgrades, and so on, it stays up all the time. And it was also very, very fast, we maintained it in a set of 70. So it was actually quite a bit of hard work, as you pointed out some amount of inspiration and goal setting based on the work we had done already. And then we said in order to go to the next level, it has to be really substantial, incremental changes aren't enough, we set ourselves the goal and worked hard at it. And then there were people who are capable of coming up with these inventions, which we did. And that's how we solved the problem. So there's a lot of hard work and invention involved in solving these. And that's why we have to have the right people, the


Alex Williams  17:12  

people you find don't always turn out to be the right people, right?


Srini Srinivasan  17:16  

It definitely happens from time to time. How do you handle that? This is something which I have learned how to handle over time, there are a few principles I have followed over the years. And please get to ended up with what's the following. Number one, always respect the person, which I think everybody I've worked with, I can say this, honestly. And no one has actually actively tried to sabotage anything at work, okay. They've always been trying to work on good faith, it seems that sometimes they're the wrong place, they are not able to step up to the plate, or they don't have the capabilities at the level that is required in some of these hard technologies we're working on at the time. So some principles I follow are very, very simple. I basically give people enough chances. So not in a hurry, there is no like, oh, this person doesn't work out. We don't go out and immediately start to kind of badmouth the person and try to kind of do strange things. We just respect the person. So which means we will keep everything confidential. First of all, that's automatically leads to respect. That's what respect means. But also have pretty simple conversations with them seeing what's actually going on trying to understand not to tell people is not working. Find out? Well, one of the important things I've learned is, when you have a problem and a bunch of people, conciliate disaster, I've seen many disasters in my life, and also great successes. And you don't go in and saying who caused the disaster. You go and say, what is happening? What's actually broken? Forget about who did it, right. That's kind of the approach we take. And what happens over time is you give people chances, you have an environment where they start realizing, you know what this is may not be the right place for me. So if I look at the last, I don't know, four or five, even 10 years, it's been mostly people realizing they don't fit and leaving, because we gave them the chance and the right feedback. They're obviously cases where you can't allow that to happen if there's an economic downturn, and you have to essentially let go a bunch of people, because the business isn't any in those cases, we make it very sure that those are actually simpler cases, in some sense, because in a case where a person isn't actually working in a well functioning team, that's a harder situation, right? In some sense, because you will reduce the productivity of the team if you keep a person who's slowing down the productivity around for too long. So you have to be honest and get them to see the point. In many cases, I'm friends with In fact, most all cases, I'm still in touch with people who we've let go or move along because they realize that there is no personality involved here. It's just a situational issue. I've not always like this. Okay. So you know, in the early parts of my career, I made a lot of mistakes in these situations. But over time, I felt that, given the situations, we have to be respecting the individual and give people a chance to make it right, and give them a chance to actually move on to. And that actually works better. If you can afford it. Unless you have a case where economically you have a car, then you do it in a fair way.


Alex Williams  20:23  

What are you learning right now, in there's a macro economy in play, and it's gone through some chefs, sourdough Ella market still continues to grow. But what are you learning now, as an engineering founder about your company, your place in the world, your individual perspectives,


Srini Srinivasan  20:43  

it varies depending on the company you are right. But for somebody like Aerospike, which is a really small company, obviously, we have to be more careful about how we spend our money, and so on, have your growth will be a little slower. But as a small company, we have both, I would say certain challenges, but also certain opportunities. What we can do as a small company, because our burn rate is not that much. We can offer to move investments around so we can prepare for when gold starts again. And this is something which is interesting, right? I've noticed in the valley is during economic downturns, it's the best time to start a new company. Why? Because you don't get buffeted by a lot of noise, you get a chance to try out your technology tested or figured out if it works. And by the time growth returns, you are better positioned to take advantage of it. And to some extent, that dynamic always happens in Silicon Valley, many of the startups have joined joined them during downturns, frankly. And then we build a product and then the exits happened quite well, when the economy turned around. So for now, what I would do is I would be careful, given that the future is always something which is unknown. But also the internet isn't going anywhere. Okay, the mobile users going anywhere people's expectation online isn't going anywhere we are in the real time market. You know, what I've seen is if the internet always doubles, the usage always doubles in a certain period, it might be 18 months now instead of 12 months, and it's going well, but it's happening, okay. And that's, I think, something we can think about and figure out how to take advantage of so I'd basically encourage people to kind of hang in there and figure things out, based on what the future will be. And to some extent, right, it's about, I think the famous statement about skating where the puck is going to be versus where the puck is, I think that's we need to do more of that in downturns than and then things are going well, because there's like anything actually works.


Alex Williams  22:44  

One of the learnings I've had from talking with people like you is that you bring a perspective as an engineer to the company that you're founding. And I'm curious on how you feel about the engineer who comes into starting a company? What advice would you give people who are engineers who want to become founders.


Srini Srinivasan  23:10  

To some extent, our company is particularly hardcore in terms of engineering. So if you want to find an engineering driven company, two things I would say upfront, number one, make sure that whatever you're building works really, really well. And hopefully is the best of the the second one I would say is don't build anything you can sell. The interesting thing for me to say as an engineer is selling it, because you're starting a new company, you're not just doing a project, the research project can do something which is interesting, if you are building a company, build something well, which works, hopefully the best in the world, and to make sure it can sell so find that there is a market and make sure that it is possible to sell it. And that's been the principle we followed Aerospike from day one, even before we got our first round of funding, Brian and I basically we bootstrap the company, and we had two or three customers sign paying us real dollars, before we got the first funding money into the bank. And that actually, culture is really important to be successful, long term.


Alex Williams  24:19  

I always like to ask the founders, so I interview about that first computer that they had, and if they took the programming quickly with that first computer.


Srini Srinivasan  24:30  

It's a computer but I programmed in assembler, and then 88 Intel 8080. So obviously, you know a little board on which you know, that's what I started. And the thing I remember is, this was an story. 18 years old, I was in college at two in the computer science thing and a professor took the challenge to write the shortest program on that. And I remember winning that contest. I didn't know I couldn't even program and I've never seen a computer in my life. And then I got this thing and then he started writing I said look The price was nothing, I surprised myself that I could write assembler in very few instructions to do something. I don't remember the sorting some, something very simple. But that was number one. And then that led to programming, believe it or not, using punch cards on an IBM mainframe, followed by a prime computer, which actually was gets us closer to the modern world, so to speak, civilized, civilized world. For me, the most interesting thing about the computer is the system side. The first ever production code I worked on as a grad student was on the adding a new file system to Unix. And that page of scheduler code I looked at, I still can't forget, because it was written so well. And there's no way any of the code I've written in my life even gets close to that even today. But that's been the inspiration why I've been working in systems programming, you know, database technology. That's what kind of inspired me to some extent.


Alex Williams  25:55  

So does that still inspire you today and your work and trying to reduce latency even more and be able to scale higher? Yeah,


Srini Srinivasan  26:01  

if you can build systems, which really work well, like UNIX did, and like hopefully, Aerospike is doing today, that can supercharge a lot of interesting things. You can say, oh, Unix is this technology. But you know what? Linux came after that. But the point is the ideas of Unix. And what they did was the supercharge all of the online, the whole internet runs on the servers and data centers, right? And what are they running next day running apps on top of this thing, and their databases to foundational? And that's why all kinds of new databases got invented. Why? Because you couldn't build applications. Why are all these high throughput applications running on Aerospike? Because you know what, you'd have to build the database yourself if you had to write those applications. So that's kind of what we do. Right? So you build on top of these infrastructure pieces. So yes, it's definitely inspires me see, writing good code at the system's level drives a lot of innovation on top of it. It's a foundational thing, if you didn't have these infrastructure pieces, you couldn't build your applications, you couldn't add new user experiences every day, which people do and deploy them at scale. And it's not just us doing it, right. It's a lot of people working on it. And we are a part of it, which is what I really enjoyed. We are the bottom layer. It's hard. It takes like 12 years to build companies, you know, but the point is, that's what I


Alex Williams  27:25  

thank you so much for taking the time to talk. I chatted with some of this morning about API's and data. And your last commentary wanted me to ask you a lot more questions about that. But we'll have to find another time for that conversation. But I just want to thank you for your time and sharing your experiences with us in the new stack audience. Really appreciate it.


Srini Srinivasan  27:51  

Thank you, Alex. It's been a pleasure.


Alex Williams  27:54  

Thanks for listening. If you liked the show, please rate and review us on Apple podcast, Spotify, or wherever you get your podcasts. That's one of the best ways you can help us grow this community, and we really appreciate your feedback. You can find the full video version of this episode on YouTube. Search for the new stack and don't forget to subscribe so you never miss any new videos. Thanks for joining us and see you soon.


Transcribed by