Karl Matthias and Sean P. Kane co-wrote the recent book Docker: Up & Running (O’Reilly Media, 2015), a practical guide to packaging applications with the popular open-source containerized application deployment tool. The book covers much of what the pair has learned through using Docker at New Relic, which also helps form the basis for the panel they’ll be participating in at FutureStack15, our annual user conference in San Francisco, Nov. 11-13.
Karl and Sean both hold the title of Lead Site Reliability Engineer at New Relic. Karl has been in the field for nearly two decades; before joining New Relic, he worked as the head of operations and architecture at insurance telematics provider MyDrive Solutions in the U.K. and as senior systems administrator at cross-platform mobile social gaming network Scoreloop in Germany.
Sean came to New Relic after stints as platform operations engineer with collaboration solutions provider Kavi Corporation and as lead systems engineer at TV and movie measurement service Rentrak.
We talked with Sean and Karl about their experiences, and about their expectations and hopes for the future development of the technology stack.
New Relic: New Relic is fond of the term “data nerds” for its community. What’s your personal relationship with data?
Karl Matthias: I’ve spent a lot of time working in data-focused companies, and my very first professional gig was at one of the country’s largest insurance companies. Our largest database took up two entire aisles in the data center. These days that doesn’t seem like that much, but at the time it was enormous.
Sean Kane: Being a lifetime operations engineer, data has always been incredibly valuable to me. When you’re trying to resolve mysterious issues in a complex production environment, data can often be the key. Single graphs are sometimes of limited use, but when you can look at multiple graphs together, you often get a picture that reveals a pattern or timeline that points back to an incident’s root cause.
New Relic: Why were you interested in speaking at FutureStack 15?
Karl: I first started using New Relic when I worked at a startup that was an early beta customer. It was a game-changer for us, and after that I helped bring the tools into other companies where I worked because being without them was like flying blind. I’ve heard such great things about this conference, and with the interest in Docker and in our book it seemed like a great fit.
Sean: Technology can be incredibly complex, and running an always-on Internet-based Software-as-a-Service company is challenging for even the best companies. It is important that we share what we’ve learned about successfully running at modern Internet scales, so that the whole industry can benefit and provide better service to our customers.
New Relic: What are the topics/themes you’re planning to address in your panel at FS15?
Karl: We’re going to be part of a panel discussing Docker in production with Google’s Mark Mandel, Docker’s Banjot Chanana, and Joyent’s Tim Gross, moderated by our own Stevan Arychuk (Thursday, Nov. 12, 1:35 – 2:35 p.m.).
Sean: New Relic started working with Docker very early on, and over time we used it to help improve our processes and address a number of issues that the company was facing. At the same time we also found ourselves stumbling over some concerns that often come with adopting new technologies.
Karl: We’ve been running Docker in production long enough that we’re already building out our second-generation platform. That’s a rare position to be in, and we’ve been trying to share our experiences with other people, both to help them out and to help show what ops is like at New Relic.
New Relic: What’s your interest in or connection to the “future stack” architecture underpinning modern software?
Karl: The biggest driver of change in the stack has been how fast people are expected to ship software to production and how rapidly it must scale. Everything we’ve been building in this space for the past 15 or more years has been aimed at that goal. Containerization is one part of the most recent wave, and it’s going to drive change across all the cloud providers and every data center.
Sean: I live in the stack, with a constant eye on everything from the electricity powering our servers up through the applications that make up our consumer products. Technologies like Docker, Mesos, Kubernetes, and CoreOS are rapidly changing the way we interact with servers and the vital computing resources that they provide. Only five years ago, configuration management and Infrastructure-as-Code were the future, but it has quickly become clear that many infrastructures are simply too large to manage thoroughly from a single code base.
New Relic: How do you see the stack evolving, and what do you think are the most important issues we face in that evolution?
Sean: People’s expectations are constantly changing, and I believe the industry must continue to learn at a pace that mirrors the pace we deploy new code at. Never before has the consumer’s demand for always-on service, available from anywhere in the world, been so great, and we are the ones challenged with making that a reality.
Karl: I’ve got an almost 20-year career in systems and ops, and I’ve watched the stack change a lot in that time. We’re now working to help drive the next level of that change at New Relic—we’re pushing boundaries both in containerization of our entire production infrastructure and in how you monitor and measure your applications and the health of your business. New Relic is not just talking about the future stack, it’s our mission to help build it.
New Relic: What are your personal goals or wishes for the future?
Sean: Personally, I want to see individual servers and cloud instances become truly disposable and replaceable, by building resilient systems that are designed for the unpredictable environment that is the Internet. This raises the challenge of discovering and monitoring dynamic services that might move around the data center and cloud in a completely automated fashion, responding to changes in load and failures as they happen. We are building some of the tools to make this a reality, but there is always room for better, generalized open source solutions that can be used by everyone in a variety of environments.
New Relic: What are you most looking forward to at FutureStack15?
Karl: I hear the speakers are great!
Want to join Sean and Karl, The Woz, and hundreds of your nerdiest friends and colleagues at FutureStack15 at San Francisco’s beautiful and iconic Fairmont Hotel, taking place November 11-13? You’ll also get to attend the kickoff event at the historic Fillmore Auditorium, headlined by comedy music legend “Weird Al” Yankovic.
Note: Event dates, speakers, and schedules are subject to change without notice.