Brainwave Podcast: Kurt Sand, Broadcom – Cloud-Native Poster Series

• Listen to the podcast on your favorite player here: https://anchor.fm/intellyx/episodes/Kurt-Sand–Broadcom—Intellyx-Brainwave—Cloud-Native-Poster-Series-ehcb54

• Watch the video version on YouTube here: https://youtu.be/2eNf0m0CXrs

About Kurt Sand:

As the head of Automation at Broadcom, Kurt is responsible for delivering on-premise software and SaaS AI/ML solutions that solve today’s real-world problems and prepare global enterprises for the challenges of tomorrow.

In this role, Kurt was instrumental in developing Broadcom’s proprietary Automation.ai, the industry’s first AI-driven software intelligence platform that accelerates digital transformation at scale. Automation.ai powers BizOps from Broadcom, a new solution that leverages AI insights to drive decision-making across multiple business and technology domains transforming customer experience, employee productivity, operational efficiency, and helps speed innovation.

With a career spanning 25+ years, Kurt is known for his strategy development and product management capabilities, leading global teams that deliver strategic growth. He has held management positions at several leading software companies including IBM, NetApp, Telelogic and GE. Prior to joining Broadcom, Kurt was the GM of NetApp’s Insight Software where he led a business unit that delivers analytics software (On-Premise and SaaS) for managing cost and performance of hybrid storage and cloud infrastructure.

Kurt earned a B.S. in Computer Engineering from Lehigh University and an M.S. in Computer Engineering from Syracuse University and has conducted post-graduate studies in Engineering Management at George Washington University. Kurt lives in Boston, MA with his wife and two daughters. He enjoys skiing, hiking and travelling.

About Broadcom

Broadcom Inc. (NASDAQ: AVGO) is a global technology leader that designs, develops and supplies a broad range of semiconductor and infrastructure software solutions. Broadcom’s category-leading product portfolio serves critical markets including data center, networking, enterprise software, broadband, wireless, storage and industrial. Our solutions include data center networking and storage, enterprise, mainframe and cyber security software focused on automation, monitoring and security, smartphone components, telecoms and factory automation. For more information, go to www.broadcom.com.

• The BizOps.com Community Site: https://bizops.com 
• Get your own copy of the PDF of this poster now: https://intellyx.com/cncposter
• Listen/download the Podcast on your favorite channel or player here: Intellyx channel on Anchor.fm
• Watch the YouTube version here (video interview): https://youtu.be/2eNf0m0CXrs 

Transcript of the podcast:

Jason English: All right, welcome to the Intellyx podcast. Today’s guest is Kurt Sand. He’s the SVP of automation solutions for Broadcom. thanks for joining me, Kurt.

Kurt Sand: Thanks, Jason. Great to be here.

Jason English: Yeah. can we just have a brief introduction to, what you’re doing at Broadcom now.

And, how that fits into the overall cloud native landscape.

Kurt Sand: Sure. So maybe many of you already know by now, but Broadcom acquired CA Technologies. So that’s how we ended up part of the Broadcom family now, which has been great journey for us all. What we’re focusing on inside our enterprise software division, which is what we formed when we came into Broadcom is an end to end software solution that we call BizOps — that starts from the front end of figuring out what to build and doing portfolio analysis through agile execution, through the software development, you know, CICD pipeline all the way out to production and monitoring customer experience.

Jason English: Yeah, that’s, that’s a lot of coverage. So that basically spans the whole life cycle from a strategic planning point of view, all the way to kind of the execution and deployment, right?

Kurt Sand: Yeah, that’s right. We have an open data integration platform that we pulled it all together. And of course we embrace a broader ecosystem. We totally understand that people don’t necessarily use all the pieces from us, but we do our best to make sure that we pull it together in an integrated way.

Jason English:  So given that broad coverage, I guess that’s why they call it Broadcom. Right.

But given that broad coverage, what technologies would you say are kind of upstream or downstream from your solution or is it really more of a thread that kind of runs through various different kinds of software that are used in, in the overall life cycle?

Kurt Sand:  There’s definitely still stuff upstream of us, you know, if you think about like HR systems or CRM or financial systems, right? That would be data feeds into our platform to understand the finance, the sales, customer data, you know, the things that give a business context to their customers.

I think we’re pretty comprehensive then all the way through software delivery and software operations, but there are certainly other things downstream around us, like data center, infrastructure management. Boxes still live in buildings and buildings of heating and cooling and space and all that other complexity of the data centers itself.

Jason English:  What challenges do you find that your customers are facing when they embark on a journey like this?

Kurt Sand: Where we all live today, the first one is just incredible rapid change. You know, I think it’s something that we set off to help customers adapt the change factor, but none of us could have predicted change would come this quickly.

Um, and you tack that on top of a ever increasing complexity in the actual landscape, right? As we look at these Cloud-Native landscapes, you know, for the most part, people still have a hybrid world, potentially mainframes still on premise, traditional on-premise, private cloud, and then growing technologies in the cloud, lots of new ephemeral, container-based stuff.

It comes and goes so fast. Kind of adds that broad and broader landscape. But at the same time, you’re trying to deal with fast change. That seems to be where a lot of our conversations start these days.

Jason English: Hmm. so, you know, one of the big parts of your, kind of rollout is this is you’re announcing what’s called BizOps.

And so, what is BizOps and how does that tie into this story?

Kurt Sand: So we kind of think of a BizOps as having three key kinds of subdomains, if you will.

So the first being ValueOps. And that’s traditionally been breaking down those silos between strategy and agile execution. Then, you know, you look at DevOps where it’s really done an awesome job at breaking down the dev and test continuous integration, continuous test, and then third domain of AIOps, which is really about breaking down the ability to troubleshoot proactively between kind of software and infrastructure.

But then when we take the BizOps view, we realize that there’s really chronic misalignment across those domains, between the business domain, the dev domain and the ops domain. What we’re really releasing is something we called the automation.ai, which is a platform that brings in data from those different domains, integrates that, correlates it and help us get new decision making insights across that end to end view.

Jason English: Yeah, that’s a big part of the story we’ve seen so far. I think most of the focus in the early days of cloud computing were really just about provisioning the resources or running applications in the cloud. And now, as the intelligence level increases, you’re finding that the cloud resources are being used even more extensively for purposes like AI.

So, what’s your secret sauce as far as processing AI, using the cloud-native resources?

Kurt Sand: So we’ve got at this from a few different angles of trying to do squared-out AI projects where we can show value to end users. So if you think about in the DevTest domain, one of the really big challenges we went after was the constant bottleneck of test automation itself taking longer and longer to run.

And what we’ve realized, if we can reorder the tests, we can more likely run the tests that will fail earlier and fail faster. So one example, there’s a bunch of machine learning that we do to figure out as we observe tests run, and as we observe changes, come through the pipeline, how to reorder the test, to put the ones most likely to fail first.

That’s one AI project. Another example would be, observing the quality of a software project underway and analyzing different characteristics of it and comparing it to others. And with that, we assign a release risk score. And we’re trying to really elevate the conversation. That every time you release software, you always take risk.

Right. You’ll also, you’ll also hopefully get some reward, right. But we’re trying to help people kind of balance that risk/reward where you’re saying, should I hold onto it just 30 more days in my test lab and burn more risk or is it time to ship? And what’s the reward of doing it now versus waiting and helping people make the right timing decisions .

Then of course, downstream on the Ops side, there’s a lot to do with these very complex growing hybrid cloud environments to ingest that data and use machine learning, to really dig for those anomalies and trying to help people troubleshoot faster, find issues earlier across that full stack.

So we’re trying to bring the integrated data together. And one of the keys, if I could say too, is also bringing the data to the other personas. So for example, having the DevOps persona really see the operational data. And having the operational persona see the development data while it’s in flight, and making sure that they each get overlapping data that they need to make their job easier.

Jason English: Yeah. That’s an interesting angle because I think people, they think of testing and then at a higher level, observability is just about looking at the inner workings of the software, but there’s also this kind of process awareness that needs to happen right?

Between those teams, as people are pushing issues from operations back to development, I mean, how, how do they make sense of all of that without a lot of help along the way, wouldn’t you say?

Kurt Sand: Yeah, definitely. I think that’s why AI is making this possible because it’s just become too much data for humans to really digest fast enough. Or if even at all, you certainly can’t just throw more people at it and figure it out. You know?

So what we try to do is give very simple starting graphics like a score, a preproduction score, a postproduction score, you know, something like an ABC or a D. Now that gives you a quick view and then you can click on that and go down and down and down and see the evidence of all of the different characteristics we use to build that score, to figure out where to take a remedy, but you really want that kind of easy, simple first glance view.

And then you, you want the depth behind it to kind of one trust the algorithms, but then two, take action.

Jason English: Well, that’s a pretty significant change. Um, as far as that, the level of coverage you’re achieving with that. Do you have any interesting kind of stories about how customers might be leveraging this now?

Kurt Sand: Yeah, a few. So, you know, there’s a really large movie theater chain that we’ve worked with really closely. By them bringing both operational data, and actually all their transactional data together, they’ve been able to do some really interesting things to drive the business. For example, looking at correlation between styles of movies and what people buy at the concession stand, or movies that are selling out more than others and pushing it to more screens, um, even examples of.

Depending upon the season that you’re in, the number of humans sitting in seats. Then you actually turn up the HVAC or turn on the HVAC. So it’s kinda cool and unique around the integrated data that you get all the way from business transactions, through your operations.

Jason English: Interesting. So what would you, what would you see as a, as a pretty unique trend to watch over the next one or two years?

Kurt Sand: I think hybrid is here to stay, I guess is my one point of view. I think what happens in the cloud will keep evolving and what happens on premise will keep evolving. But I think most large enterprises that we work with will continue to have that mix. And that complexity I think is here with us to stay.

I also would say that, AI and ML are math that have been around for eons, you know, the real clever thing is how we apply it. To assist humans in making better decisions. and even in a survey that we recently conducted, it’s, 80% of the people definitely believe that AI is helping humans make better decisions.

So I think we’ve gotten past this, like “the robots are gonna take over the world” and more of how does AI actually just help us do what we do, but do it better. And I see that only growing over time.

Jason English: Yeah, I think it’s really being employed to solve some of the stickiest problems that we have ever had.

And, definitely in our coverage, we’re very, sympathetic to this idea that most companies can never leave fully leave that legacy behind it’s always going to be a hybrid world where you bring all of the systems forward. And then you really think about running applications and application workloads, where they make the most sense.

I am interested in how you perceive risk in this kind of environment. How do you basically apply an algorithm to understand how much risk there is in a release environment? And what kind of value is available on the back end from that?

Kurt Sand: Yeah, so the risk score, there’s a bunch of things that we can look at. We can look at an amount of changes that were made. Typically, the more changes you make, the more risk there is. We can also look at the tests that were executed, both quantity of tests, coverage of tests compared to previous runs.

Of course, also test failures. And then we can also look at feedback from production on how well have past releases that are very similar performed. Those are kind of three examples. And there’s more that we’re evolving in the algorithm as we’re bringing more and more data from more customers into it.

Jason English: Yeah.

Kurt Sand: And then on the production side, of course, a lot of what you’re trying to do, is really monitor backwards now, customer experience. So actually I’m sitting there at the edge of where users are using the application and seeing what journeys they’re taking, where they’re getting stuck, where they’re failing out, is a big part of the current scores of post-production, of course, with the traditional stuff like performance.

you know, another example we have is we work with a retail store that just pushed out tablets to all of its sales clerks. And then the dropout was really high, meaning that they just weren’t getting end-to-end to the purchase on the tablets. And although this customer was really advanced in their DevOps pipeline, they could push updates really fast, but after a gazillion updates, they hadn’t improved.

And it was only once we instrumented the app on the tablet and we actually could see where the people were getting lost. It just wasn’t clear to the designers that they had made some things hard to find in a few workflows that weren’t so obvious. So, it’s an interesting thing. You can have really great performance on the tablets.

You could have really fast delivery, but if you don’t know where the issues are, you still don’t solve it.

Jason English: Right. That does bring it all back home. Really it’s about the customer experience after all. So if you’re centering the change around that, that’s digital transformation.

Kurt Sand: Yeah. Otherwise you just keep trying and hoping and trying and hoping, right?  Instead of having some real feedback loop from the customer.

Jason English: Any final thoughts that you’d like to add?

Kurt Sand:  Yeah, just really thanks for the time. And this is definitely an exciting journey that we’re on adding a data integration layer and bringing AI/ML to real projects.

So we look forward to taking it out into the wild with customers, and maybe stepping back with you after we get some more feedback and tell you what we’ve learned.

Jason English: Excellent. Well, thanks Kurt for joining me here. And good luck with that. It’s a pretty exciting scope of a project you’re undertaking there, but, kind of cool to hear some of the results  so far.

Kurt Sand: Awesome. Thank you.

Announcer: Thank you for listening to the Intellyx podcast. If you have any questions or ideas for future episodes, feel free to drop us an email at PR@Intellyx.com. Until next time, keep on transforming.

 

©2020 Intellyx LLC. All dialogue in this program represents the expressed opinions of the hosts and guests, and are not necessarily the official position of Intellyx, or any company mentioned or included in this podcast audio or video. Intellyx publishes the Cloud-Native Computing Poster, and the biweekly Cortex and BrainCandy newsletters.