DevOps Chats: Automic + CA + Broadcom, a Release Automation Journey

Mitch Ashley: Hi, everyone. This is Mitch Ashley with DevOps.com and you’re listening to another DevOps Chats podcast. Today, I’m joined by Scott Willson, he’s product marketing manager, Release Automation at CA Technologies at Broadcom—that’s a mouthful. I think we’re gonna explore a little bit of that.

So, our topic today is really catching up with what’s been happening with the Automic product and CA’s, this part of the acquisition back into Broadcom and just kinda catch up with Scott. So, he’s been on the podcast before. Scott, welcome to DevOps Chats.

Scott Willson: Thank you, Mitch. Glad to be here.

Ashley: Appreciate you being here. Would you start just by introducing yourself for any of our audience that doesn’t know you, a little bit about what you do, and tell us a little bit about CA Technologies at Broadcom.

Willson: Yes, yes, so my name is Scott Willson—two Ls in Willson. I think many listeners may know me. I have been, in the past, an author with some of Gene Kim’s DevOps Forum papers, and I’ve been at the DevOps Enterprise Summit several times. I’ve been fairly active on blogs and all that stuff.

And where my kinda claim to fame was, I came from a company called Automic. That’s spelled Auto-mic for the phonetic spelling of it. My background is, just to throw this in here, too, my journey, as I always tell people, is I started from a development standpoint writing the code, migrated my career kinda doing the DevOps thing before anyone called it DevOps, like many others.

Ashley: Mm-hmm.

Willson: And yadda yadda yadda, here I am now with a software company speaking all things DevOps and trying to help people get better at releasing their software.

Ashley: You’ve certainly been a part of the community for some time now, so we appreciate all your contributions. Well, you’ve been through quite a journey. I know being acquired by CA and then CA being acquired by Broadcom—why don’t you kinda catch us up? Maybe tell us a little bit about that journey. I think you described it as an 18 month history going through that.

Willson: [Laughter]

Ashley: You know, we don’t have too long on the podcast, [Laughter] but if you can kinda give us a little bit of that path that you’ve been on.

Willson: Yeah, yeah. You know, I kinda looked at it like a, it was almost like a barracuda eating a great white who was eaten by a whale, right?

Ashley: [Laughter]

Willson: It was a really interesting, very quick journey. It was a couple years ago when I was with Automic Software. We were acquired by CA Technologies, and about a year or 18 months after that acquisition or so, Broadcom bought CA Technologies to the—not just our surprise, but of course, to the surprise of the entire market.

Ashley: Mm-hmm, yeah, it was.

Willson: There was a lot of press about it.

Ashley: It was a surprise.

Willson: And so, yeah, it was a very, very quick and interesting turnaround. We are now known as CA Technologies, a Broadcom company. As far as Automic goes, we are retaining the Automic brand, so you’ll see a lot of our old software and all that will have the Automic brand. And not just Automic, but our current leadership has really felt to really keep and embrace the brands that we’ve had. So, you know, Blaze—CT is still gonna be Blaze, and Rally is gonna be Rally, those type of things.

Ashley: Mm-hmm.

Willson: So, the brands are remaining, which I think is useful, because even though CA had remained Rally—I know I, anyways, always refer to it as Rally, because—

Ashley: I do, too, yep. [Laughter]

Willson: – right, because that’s what I always knew it as. [Laughter]

Ashley: Especially knowing them from the beginning, it’s hard not to call them Rally, but yes.

Willson: Right, exactly. So, Automic is still going to be Automic from a product brand positioning. And then from a company standpoint—yeah, we’re still CA, just a Broadcom company. Within Broadcom, our group is known as the Enterprise Software Division, so we are a distinct division within Broadcom.

Ashley: Mm-hmm.

Willson: And then, to add to that, to the little journey, what has been interesting to me is—I remember it was a year ago, the last Forum meeting I was going to or what many of us in the inside called those meetings, the Gene Kim’s Great Slumber Party.

Ashley: [Laughter]

Willson: Right? We kinda talked about being lean, and some of the people, I remember having discussions with and running papers, we just really got to talking about what it means to run lean and the lean manufacturing and the whole thing, right, that DevOps is based on and that The DevOps Handbook espouses and all these kinda things.

And after that, right, three months after that meeting it was announced Broadcom was gonna buy us and then they actually did, and here I am all this time later. And I bring that up because what was interesting to me, Mitch, is that Broadcom embodies this. They run lean.

Ashley: Mm-hmm.

Willson: And it’s been an interesting journey to go from the short time at CA, which ran big, to now being with Broadcom, which runs lean. And they run even leaner than what Automic did, which was in ISV, right? You know, Independent Software Venture.

Ashley: I was curious about that, because I know Broadcom, you know, I had worked with them as a chip supplier with companies like Broadcom ad Qualcomm and, you know, manufacturers of chips in telecommunications think about the world, they’re much more, you know, operational efficient. It’s all about fractions of pennies and the cost of their goods to maintain profitability, but that’s super high volume. So, it’s a different mentality than a software company.

Willson: It is, and there are some significant differences, I suppose, from the business or sales aspects. But from a production aspect, there really wasn’t a whole lot of difference. When you think about taking a product to market, right? Now, we were doing this with software, they’ve done it with chips, but Broadcom is an engineering excellence company. They really pride themselves on having the best of fill in the blank of engineering, whether it’s engineering processes, engineering whatever. In fact, what has been interesting post acquisition is that our engineering staff for the bulk of our products have actually increased. Broadcom is very keen on ensuring that the products we deliver are best in class.

Ashley: Mm-hmm.

Willson: And so, they realized a way to do that in the market is to make sure they’ve got the top minds developing and producing and shipping these products, right?

Ashley: That has to be refreshing to you—you and the team.

Willson: Oh, yeah. It was very interesting to see that that was the case. It’s been a unique acquisition as far as that goes, and to see that—no, they wanted to increase the head count of R&D. [Laughter] It’s important for them.

Ashley: Mm-hmm. What? [Laughter]

Willson: Exactly. Like—wait, what? They want the products to improve, to get better, and to continue to be viable and market leading. And so, they embody the lean processes to do all of this stuff. And so, it’s been—there’s a little bit of a learning curve, I think, on both sides to adapt to this new way, but here we are. You name the things that are automated—I mean, Broadcom automates just about anything and everything. The very stuff you read about, Gene Kim talks about in his book, saying—and even in Nicole Forsgren’s book, Accelerate, you know, automation is at the key of a lot of these things, and Broadcom embodies that.

Ashley: It is. It’s a much different kind of company. If you’re not a hardware person or haven’t worked at a chip company, what I said, it’s—fractions of a penny make a difference in cost, and that cost can be in materials, but it also can be in process. And when you multiply that by millions upon millions of chips that they may ship a particular line, that’s a lot of money. So, it makes a huge difference in the profitability of the company—

Willson: That’s right.

Ashley: – to be able to reinvest and give you more engineers to be part of your team.

Willson: That’s right. And so, software is a little different, right? It’s not the pennies like that. Obviously, with software, the margins are different, because you’re not dealing with a physical thing, a commodity, right? But the principles to be efficient like that are absolutely espoused by DevOps. And one of the things that this transformation internally has occurred—because we’re transforming ourselves from a software company, right? I think we’ve made some press releases talking about our BizOps platform, a holistic platform—

Ashley: Right.

Willson: – that we’re offering the market. No longer should you be thinking of us like the CA of old where you had a bunch of different siloed tools. Now, they’re all coming together in a holistic way to provide real business value and create a true, interesting platform for everybody.

Ashley: Mm-hmm.

Willson: But just this way of doing things lean like that and embracing the agile way of turning things around has been a cool thing to behold.

Ashley: That has to be really good, because I haven’t talked with somebody that’s had that experience—not that others haven’t, but to have that common lexicon frame of reference of thinking about producing software like a software factory just like a hardware based factory. You have a lot of synergies and commonalities there. Yeah, there’s differences, too, but instead of software being this mystery to hardware guys or vice versa, right, you can talk and work together and understand the efficiencies, improvements, speed, impact, quality—all the things that drives the manufacturing company also drives software.

Willson: That’s right. And what’s been exciting, too, with this opportunity to change—really, the Broadcom offer has provided us an opportunity to change things and to really force these integrations and synergies between what used to be separate divisions of CA but are now different teams and the interoperability that’s happening between them to deliver this BizOps platform is really exciting. It’s cool to see that now it’s like—look, we have all these things, but what’s the point, unless we’re focusing on business outcomes, right?

Ashley: Right.

Willson: We can actually drive a business outcome if there’s harmony and cohesiveness between all these things, and that’s very exciting. In fact, one of the other things along these same lines that was announced is the free tier program—I don’t know if you see the PR a couple weeks ago, Mitch—

Ashley: I did.

Willson: – but now, a lot of our products are offered for free.

Ashley: Yeah.

Willson: So, for example, a continuous delivery director, or Automic Continuous Delivery Director, if you go to CDDirector.io, you can just use it for free, up to 10 active releases. And Rally has some free tier as well. So, basically, we’re now offering the freemium versions of what we had, and—

Ashley: Have the products changed substantially to be able to do that? Are they less functional or are they time based or just volume, kinda getting folks to up to date?

Willson: It’s really just volume, right? Yeah. They haven’t really—well, a lot of them didn’t really need the change. A lot of the products that are offering this are the ones that are already SaaS based already. There will be more that are coming, of course, as things are being migrated to be SaaS based, and that effort is under way.

But an example of Automic Continuous Delivery Director, there’s not really a time limit. It’s 10—10 active releases. That’s a lot of releases.

Ashley: Mm-hmm.

Willson: And you can just use it, and use it at will for that. You wanna get more than that or grow it across the enterprise or make it the standard or something, same with the Rally and Blaze CT. Oh, well, yeah, then you can buy the licenses that let you use it more broadly, right?

Ashley: Mm-hmm. That’s a great model. I mean, I think we’ve sort of disproven the crippleware approach. Nobody likes that.

Willson: [Cross talk], yeah.

Ashley: Exactly, and being able to use it on enough of software production; you talked about 10. You’ve gotta be able to get a pretty good feel for what the product does that’s gonna do what you want it to do and start to integrate it into your tool chain workflow platform.

Willson: Right, and these are 10 active releases. In other words, 10 releases in parallel, right?

Ashley: Mm-hmm.

Willson: Well, that’s—like you said, you should really be able to get a feel for it.

Ashley: Yeah.

Willson: Being able to run that many releases in parallel to each other.

Ashley: Well, that’s awesome. That’s great to hear. I’m excited to hear about more of that coming from you, too.

Do you have a new release coming up, here?

Willson: Yeah. So, speaking along those lines, we have a new release for Automic Continuous Delivery Director. We’re pretty excited about this release. It is our first release that includes some ML capabilities and heuristics that are involved. Basically, what we were looking to do with this is, amongst other improvements and advances we made in the product, right, to help with continuous delivery, one of the things we looked at is that there’s significant bottleneck in continuous delivery, the methodology, which is QA or testing.

Ashley: Mm-hmm.

Willson: Now, as you probably know, Mitch, continuous testing is technically a subset or a component of continuous delivery.

Ashley: Right.

Willson: And I might argue that unless you’re doing continuous delivery, you are not doing continuous delivery.

Ashley: Right.

Willson: You need to be doing—

Ashley: You need to think of continuous delivery is plan, dev, test, deploy, operate, right?

Willson: That’s right, and you’re testing continuously, right? But the problem is, QA wasn’t really built to be agile themselves. So, they kinda became the bottleneck, not just from the way that they run, but you, as a QA person, right, testers, you build out all these tests. And so, you release this small change per DevOps per agile per continuous delivery, right? And it goes into QA and you run, you know, 1,000, 10,000 tests—and I’m not making these numbers up. Literally, a lot of our large enterprises have thousands and 10,000 tests.

Ashley: Oh, yeah. Totally.

Willson: Alright, so then, we’re in the waiting game, right? And maybe it takes however long for the test results to come back and then you get feedback to development. So, we found with our, a lot of our large customers, you know, the Global 2000, that this was a significant bottleneck. So, now what we do, what occurs, is that when you do that release and you’re a pipeline, right, you’re a release pipeline and it gets pushed to QA, one of the things that we do is, we look at several heuristics.

So, we come around and say, “Alright, well, what are all the tests? What is the historical responses or outputs of the tests?” Right? How many of them failed? Which tests are flaky? How many tests are new tests? We want to actually run these tests first. So, a flaky test, for example, it ought to run at the front, and if it fails—pfft, we’ll send that feedback immediately to development or to QA so that we’re identifying these things earlier, quicker, faster, rather than having to wait on the back end for this, right?

Ashley: Try to get to that 80/20, right? Get the things—

Willson: That’s right.

Ashley: – upfront that are gonna cause the most issues, find the most bugs.

Willson: Right. So, we’re looking at all the data and intelligently figuring out the tests for you. The other thing that we’ll do is that we’ll look at the code changes you’ve made and we will map those code changes. At least in the case of Java apps, we’ll be able to map the code change to specific test suites.

Ashley: Mm-hmm.

Willson: So, now, when you submit that one component change, you are actually getting identified just the test runs that affect or exercise that component. So, altogether, when you put it all together, Mitch, what occurs then—now, when a change gets pushed into your continuous delivery pipeline, when it hits QA, you’re in a position to not have to run all 10,000 tests. You’re in a position now to run the high risk tests, the tests that you have marked are critical and important, right? So, those are included, that’s taken into account, too. And the test that we know actually exercised the change that you put in.

So, you’re able to run what you have to run to ensure you have a high quality and also order them in such a way where there’s high likelihood of failure that Development is fed back that feedback earlier and faster. And so, we’re very excited about this release.

Ashley: That’s awesome.

Willson: Yeah.

Ashley: Boy, gone are the days, I can remember when I would judge the quality of the release by how severe the bugs are and if we haven’t found really big problems yet, we’re not through. There’s more to come. [Laughter] You know, that was the heuristic. [Laughter]

Willson: That’s right. Exactly. That was the heuristic.

Ashley: That’s right. We’re far from those days, aren’t we? [Laughter]

Willson: Well, you know, because one of the other constraints, if you think about agile and continuous delivery, right—small change and you make more frequent small changes and the “State of DevOps” report just came out, right, a week ago or this week, and it again is showing how this way, this approach is actually reducing the bugs and reducing ________ quality.

All those things are great. But that now means one of your biggest resource constraints is time. And what we’ve found with our customers, it was time in QA. So, by intelligently assigning the tests on a run with a particular change, we’re able to shorten that down. You’re not having to wait for all 10,000 tests, you’re just waiting for the five that matter to run against that little small change and—pfft, get it shipped out the door.

Ashley: Exactly, right. Be smart about testing, yeah.

Willson: That’s right—be smart about testing. Exactly right.

Ashley: Now, do you have a date when this is going public? Are you in beta now? Where are you at?

Willson: Yep. So, if you were an on prem user of the product, it is actually GA now. It actually was up on the download site yesterday and—

Ashley: Oh, nice.

Willson: – it can be downloaded and you can go through the upgrade process. If you’re on our SaaS platform, you know, October 2nd, it’ll just be online working and available for you in the beauty of SaaS, right?

Ashley: Nice. Well, great. That’s fantastic. Awesome. Well, congrats on the great release. I hope it works out very well and also the free tier brings some new customers, new people onto your platform and expand that as well.

We have a few minutes left. I know you’re gonna be at the DevOps Enterprise Summit, correct, in Las Vegas?

Willson: Yeah, exactly. So, we’ll have a big booth there at the DevOps Enterprise Summit, so certainly looking forward to seeing everybody wearing our Broadcom wear. We’re gonna—not to give away too much, but I think we’re gonna have some pretty cool giveaways at our booth at various times, so, you know, definitely come check us out.

Ashley: Come early, come often for the swag.

Willson: What—get the swag? [Laughter] That’s half of what you go for. Especially if you have kids, you know? I see some of these people, they’ve got bags full of this stuff, right?

Ashley: Mm-hmm.

Willson: And I’m like—yeah, you’ve got kids at home. [Laughter]

Ashley: You don’t wear a Small—you don’t wear a Small. What are you—what are you doing, there?

Willson: [Laughter] Exactly right. Exactly right. But we’ll have some things, a couple of the giveaways we’ll have I don’t think you wanna give to the kids, let’s just say that.

Ashley: Okay.

Willson: And if you’re going to the Gartner Symposium, we’re gonna be at the three upcoming Gartner Symposiums in Orlando, São Paulo, and Barcelona as well.

Ashley: Fantastic.

Willson: So, yeah.

Ashley: Well, good. We’re gonna have to check in every six months or so just so we know what swag wear, logo wear to look for you in.

Willson: [Laughter]

Ashley: This time it’s in Broadcom—I’m kidding you, of course. But almost—it’s almost true.

Willson: It’s almost true. Feels true. [Laughter]

Ashley: Scott, it’s been great having you on the podcast. Appreciate you being here.

Willson: Thank you so much, Mitch. Appreciate it and look forward to seeing you and Alan at the DevOps Enterprise Summit.

Ashley: Absolutely. Look forward to it as well, and great to catch up with you about everything happening. Well, I’d like to thank you. I’d like to thank Scott Willson, Product Marketing Manager—or sorry, Product Marketing Director of Release Automation at CA Technologies, a Broadcom company—I think I got that right—

Willson: You got it right.

Ashley: – for joining us on the podcast. [Laughter] And thank you—you, our listeners, for being here as well. This is Mitch Ashley with DevOps.com. You’ve listened to another DevOps Chats podcast. Be careful out there.

— Mitchell Ashley