In this DevOps Chat we speak with Steve Boone, product owner and development manager for Urban Code Deploy. Now that DevOps is in practice at more organizations, measuring DevOps success is a question many are asking. Steve has helped hundreds of companies adopt, implement and scale their agile and DevOps practices as part of their digital transformation. Steve gives us some insights into how can you measure your DevOps success.
As usual the streaming audio of our chat is below. Following that is a transcript of our conversation to follow along.
Audio
Transcript
Steve Boone is a DevOps SME at IBM. He has helped hundreds of companies adopt, implement, and scale their agile best practices including CI, CD, Environment Provisioning, and Application Release Automation. Steve is currently the Product Owner & Development Manager for IBM UrbanCode Deploy.
Alan Shimel: Hey, everyone, it’s Alan Shimel, DevOps.com, here for another DevOps Chat. I’m happy to be joined today by Steve Boone. Steve is the Product Owner and Development Manager for IBM UrbanCode Deploy, one of the leading CI/CD solutions in the market. Steve, welcome to DevOps Chat—how are you?
Steve Boone: Doing great, Alan. Thank you so much for having me today.
Shimel: Great. Steve, I wanted to talk to you—well, about a couple of things today. But let’s focus in first on, I’m just back from another major conference, and once again I’m—not amazed, because by this point, it’s expected. But still, you know, going strong, the biggest thing we get, or the biggest feedback that I hear from organizations looking to do DevOps transformation is—hey, this all sounds great, it sounds really good, but how do we know, you know, first of all, how do we get started with our DevOps transformation, and then once we’re doing it, how do we know it’s working? Right? How do we measure the ROI, or even if we can’t get to a real ROI, how do we measure progress? What does success look like?
And I would imagine, you know, the guys at UrbanCode, the team at UrbanCode has been dealing with this for some time. What do you tell people?
Boone: Yeah, that’s a great question, and it’s—you know, I think it’s a little different for everyone. And when we start looking at measuring success and first finding out, you know, where do we start, a lot of times we’ll work with customers to identify, you know, what are your problematic areas? There’s a lot of aspects to DevOps, it’s not all CI/CD, there could be aspects of agile planning, there could be release management. There’s a number of different things that come into play.
So, we usually sit down and try to focus on, you know, where do you think your bottlenecks exist today, and then start to address those. But when we went to measure—and you know, you’re right, the allure of DevOps comes with all of this ROI promise, right, where if we do these things and we change our people and our process and our tools, we’re gonna be better off for it, the business is gonna be better off for it.
So, how do we then make sure that’s actually happening, right? That the time and energy that I’m putting into changing my processes are actually showing me that ROI.
So, what we’ve seen a shift in the industry towards is really starting to gather the different sets of metrics that are coming from, really, the end to end products that we’re dealing with in DevOps whether it’s your build tools, your testing tools to understand what I’ll all our key metrics. So, you know, if in fact we are capable of speeding up our delivery to our pipeline, we should be getting to market faster, right? So, we look at statistics like cycle time—how long does it take for an application or a release of application to go from dev to prod? And then we look at that data over time. Is it, in fact, improving? And if not, we want to identify other bottlenecks that are kinda holding that progress up.
So, if I look at a set of components that make up an application and I see that the majority of those components have a really good cycle time, you know, we’re only as fast as our weakest link. So, if the database component, for example, is routinely slower to get through QA than the rest of the component, that’s going to slow down our time to market and therefore gonna slow down our release.
And the beauty is, in turning on the flashlight, so to speak, so that you can see that data and actually do something with it, a lot of times we look at it and say, “Well, we know we have automation and therefore we know we’re going faster.” But that just might mean we’re going to QA faster or we’re going to dev faster. We want to make sure that we’re actually getting that ROI, so if we can set baselines with teams and say, “Hey, prior to doing this automation, it took you four months to get from dev to production out the door, right? This past release cycle we did it in three and a half months—great. What is that time savings? What does that time savings mean in cost?”
And, we’re starting to see a big trend within the industry of many different companies starting to put together products to start to track these. So, I think it’s a whole new area for us to focus on, and in hopes—what we’re trying to share with customers is that, look, not only are the changes in people and process that you’re making working, but the tooling is working, and hey, here are some areas where you can improve those processes.
Shimel: Yeah. So, Steve, obviously, I want to—you said a mouthful, there. [Laughter] But let me focus in on one thing you said that I want to re-emphasize to our audience, and that is, you know, DevOps is so much more than CI and CD, right? There’s config management, there’s APM—there’s a lot that goes into this umbrella term that we call DevOps, and it seems to be growing every day.
That being said, though—you know, I once heard someone describe DevOps is the symptom and CD is the cure. So, you know, I don’t want to minimize the importance of CI/CD, but you know, automation—there’s so much that goes into this equation or this term that we call DevOps. And I think that’s—you know, Steve, it’s almost like it’s its greatest asset but it’s also its greatest liability, you know what I’m saying?
Boone: Yeah, there absolutely is. I think that a lot of times, we definitely see, when people are talking DevOps, their focus tends to be on CI/CD. And I think that’s because, you know, out of the gate, especially if you’re just starting off, it’s kind of your biggest bang for your buck, right? You can get a lot of fast return on investment just by improving your processes and automating them and making them consistent.
As we see teams mature, and as they transition into a more Enterprise wide DevOps solution, then we start to see—well, where are the real boundaries in my organization, right? Is it, we see a growing trend where app dev teams are focusing on automation, but there’s a whole different silo in infrastructure, you know? So, an app dev team wants to request an environment to do some testing in—they have automation, and the tools that they have for automation could easily provision that environment. But the business dictates that that’s the infrastructure’s just, and they have a completely different set of tools that they use for that, right?
And so now we get into this challenge of really continuing to try to break down these barriers to make sure that these assets and tools that are out there can work together, and that there’s clear cut guidelines on that.
So, yeah, I think it is a blessing and a curse, right? Because when we start thinking about DevOps, everybody wants that—you know, we’re a culture of having that instant gratification. So, it’s nice that we can get that with deployment automation and build automation and test automation. But we need to be looking ahead to the future, which is gonna be—hey, I’m gonna provision these things on demand and make these environments available to my teams, and today we use Amazon for that. What happens if, six months from now, we have to change and we need to support Azure or Google or SoftLayer or Bluemix, right? We need portability, we need full stack provisioning and config management on that.
And that’s—I think that’s something that, a lot of times, people just starting out, they don’t look too far ahead. And my suggestion to folks is—hey, it’s okay to take that first step. You should be cautious about the first step, but keep in mind where you wanna be two years from now, three years from now, five years from now. What’s your DevOps end goal?
Shimel: Yeah. You know what? So, Steve, I think that might be the first question we need to—especially for consultants out there who may be listening to this. You know, whether you work for an IBM or a small consultancy, I think that has to be the first question. You know, people come and say, “I want to do DevOps.” You gotta ask them, what’s the end goal, here? What’s the end game? Why? And that’s a huge piece of it.
Steve, if I can, I’d like to turn to a little bit more about UrbanCode and, you know, UrbanCode’s kind of a unique animal, if you will, in that, you know, it was one of the earliest, earliest kinda startups playing in the CICD space, right? And so it had that whole startup culture, if you will, the startup thing going on. And then, of course, you know, acquired by IBM and really become a linchpin of, you know, IBM’s DevOps, and not only DevOps, but you know, in IBM, as you know, DevOps is tied into cloud and hybrid and Bluemix and SoftLayer and all the things you mentioned. You know, it’s such a critical piece of that puzzle.
What is happening lately with UrbanCode? I saw Gartner give you guys some accolades recently.
Boone: Gartner gave us some accolades. We scored very well, as well, on the Forrester Wave that just came out.
Shimel: The Forrester—yes.
Boone: We were named as a leader in the—yep, in the Forrester Wave. So, that was exciting. It was a great score for us. We’re about as far right in that quadrant as you can be.
And there’s a lot of really good things coming out. I think the reason we scored so well is that they’re looking at the strategy within UrbanCode and seeing that, you know—as you mentioned, this isn’t our first trip around the block. We’ve been helping customers do this for over 10 years. And for us, you know, we work with companies that are small startups to everything that are enterprise financial institutions and healthcare and government agencies.
So, there are, at any given time, different, I’ll say, business focuses that are in a different spot of transition. So, I can go to an enterprise and talk to 30 different groups and organizations in that enterprise—they’re all in a different maturity, they’re all in a different level and area of that DevOps transformation. But what we’re able to do is, we have a solution that can help everybody regardless of where they’re at, and as they mature together over time, ultimately help them reach the goal where the company wants them to be.
You know, some of the new things that we’re focusing on—you know, last year, part of our vision, we had announced at Interconnect our DevOps Insights product as well as our continuous release product that are focused on helping customers one, transition into DevOps, do a lot of that deeper reporting and ROI analysis on where they are in their transformation, but as well as help companies manage to the cloud and for the cloud.
What I mean by that is, you know, today, we talk about cloud like it’s ours, but for large enterprises, it’s still very much a challenge. You know, it’s hard for a team to just go request access and say, “Hey, I’m gonna push my data to the private cloud.” There’s still a lot of—or even public cloud, whatever it might be. There’s still a lot of hurdles to that.
So, what we’re trying to do is, as teams embrace this and they’ve got deployments going on-prem and off-prem, how do we organize that? How do we have visibility and traceability and understand as you’re taking large applications and breaking them down, let’s say, into microservices? You know, some of those applications are easy to re-architect, some are not. And sometimes you’re left with applications where you’ve only got maybe three quarters of that application re-architected, or you’ve got databases that are always gonna be on-prem, or you’re using mainframe services.
And that’s really where the UrbanCode products excel, is allowing you to really deploy anything, anywhere, but also have a cohesive look at, what are the interrelated appendices of those parts? If something on my backend mainframe changes, how is that affecting the things that I’ve deployed into my mobile app? And if something needs to fall out of just release, you know, what’s the impact, even? How is that gonna affect business? And I think those are some of the main reasons why we scored well within the Forrester Wave and why they marked our strategy so high.
Shimel: Absolutely. You know, Steve, we’re coming up on the end of our allotted time, unfortunately, but I wanted to mention one thing if we can talk about it, and that is that you spoke about Interconnect, which has been the big IBM DevOps related show, as well as other technologies for the last couple years. This year, IBM is kinda consolidating a lot of the shows they do into a pretty big mega-event called Think 2018, and of course, that is gonna be once again at the Mandalay Bay in Vegas, and it’s happening March 19th to the 22nd. I believe the call for speakers has just opened, and I’ll put some links in the show notes so that people can, if they wanna speak or if they want to attend the event would be great.
But in any event, what’s UrbanCode’s plans around Think 2018?
Boone: Yeah, thanks. I hope folks do and are able to register and come out. It’s going to be a fantastic show, and it is very much a mega show for us. So, we’ve taken—I believe there’s 11 different topics that really, I’ll call them focal points or even think of them as investments within IBM. Everything from cloud to Watson to mobile to security, internet of things—there’s a number of different topics going on at this.
UrbanCode will be under the cloud focus, and we’ll—underneath the cloud topic, there’s a section specifically for DevOps. So, if that’s what you’re interested in, come out and check us out—we’ll be there. We’ll be talking about our latest release that we’ll have done at that time, and we’ve got a lot of really cool features that were actually in development and testing right now that we plan to announce at the Think conference, so without be demo-ing those, and we’ll have—as always, have a bunch of UrbanCode customers there sharing their ideas, sharing their experiences, and really helping educate how they’ve transformed their business.
And I think what a lot of folks will find appealing is that it’s never a one tool solution, right? And that’s one thing I always encourage folks is, when you’re trying to be in a DevOps transformation, a single product, a single methodology—it’s not gonna do it. You really need to embrace the various tools that are in the market, you need to embrace existing tools that are already in your company, and these are gonna be working examples and explanations of how some of the top companies in the world are doing just that using UrbanCode and using a variety of tools that we support.
Shimel: Absolutely, absolutely. Alright, Steve, we’re actually way over time, I apologize, but I think we got some good information here that people will be interested in. I want to thank you for being our guest this week on DevOps Chat and invite you back soon. We want to hear more about Think 2018 as we get closer as well, but good luck with UrbanCode, Steve, and we’ll speak to you soon.
Boone: Hey, Alan, thank you. As always, appreciate you having me today.
Shimel: Alright. Steve Boone, Product Manager, UrbanCode Deploy—or Product Owner, excuse me, UrbanCode Deploy. Thanks for being our guest on this DevOps Chat. This is Alan Shimel, and we hope to see you soon on another DevOps Chat.