>>from around the globe. It's >>the Cube with digital coverage of Dev >>Ops Virtual Forum Brought to You by Broadcom. Welcome to Broadcom, Step Ups, Virtual Forum I and Lisa Martin and I'm joined by another Martin very socially. Distance from me all the way. Coming from Birmingham, England, is Glynn Martin, head of Q. A transformation at BT Glenn. It's great to have you on the program. >>Thank you, Lisa. I'm looking forward, Toa. >>As we said before, we went live to Martin's for the price of one in one segment. So this is gonna be an interesting segment, Guesses. What we're gonna do is Glen's gonna give us a really kind of deep inside out view of Dev ops. From an evolution perspective, Soglo's Let's start transformation is at the heart of what you dio. It's obviously been a very transformative year. How have the events of this year affected the transformation that you are so responsible for driving? >>Yeah. Thank you, Leigh. So I mean, yeah, it has been a difficult year Bond, although working for BT, which is ah, global telecommunications company. Relatively resilient, I suppose, as an industry through covert, it obviously still has been affected and has got its challenges on bond. If anything is actually caused us to accelerate of our transformation journey, you know, we had to do some great things during this time around. You know, in the UK for our emergency and health workers give them unlimited data and for vulnerable people to support them and that spent that we've had to deliver changes quickly. Um, but what? We want to be able to do it, deliver those kind of changes quickly, but sustainably for everything that we do, not just because there's an emergency eso we were already on the kind of journey to by John, but ever so ever more important now that we are what we're able to do, those that kind of work, do it more quickly on. But it works because the implications of it not working is could be terrible in terms of, you know, we've been supporting testing centers, new hospitals to treat covert patients, so we need to get it right and therefore the coverage of what we do, the quality of what we do and how quickly we do. It really has taken on a new scowling what was already a very competitive market within the telco industry within the UK. Um, you know, what I would say is that you know, we are under pressure to deliver more value, but we have small cost challenges. We have to obviously deal with the fact that you know, Cove in 19 has hit most industries kind of revenues and profits. So we've got this kind of paradox between having less cost, but they're having to deliver more value quicker on bond, you know, to higher quality. So, yeah, certainly the finances is on our minds. And that's why we need flexible models, cost models that allow us to kind of do growth. But we get that growth by showing that we're delivering value, especially in, you know, these times when there are financial challenges on companies. >>So one of the things that I want to ask you about again looking at, develops from the inside out on the evolution that you've seen you talked about the speed of things really accelerating in this last nine months or so. When we think Dev ops, we think speed. But one of the things I love to get your perspective on we've talked about in a number of the segments that we've done for this event is cultural change. What are some of the things that scene there as as needing to get, as you said, get things right but done so quickly to support essential businesses, essential workers? How have you seen that cultural shift? >>Yeah, I think you know, before, you know, test test team saw themselves of this part of the software delivery cycle. Andi, actually, now, really, our customers were expecting their quality and to deliver for our customers what they want. Quality has to be ingrained throughout the life cycle. Obviously that you know, there's lots of buzzwords like shift left. How do you do? Shift left testing. But for me, that's really instilling quality and given capabilities shared capabilities throughout the life cycle. That Dr you know, Dr Automation drive improvements. I always say that you know, you're only as good as your lowest common denominator on one thing that we're finding on our Dev Ops Journey Waas that we were you know, we would be trying thio do certain things quicker and had automated build automated tests. But if we were taking weeks to create test scripts or we were taking weeks to manly craft data, and even then when we had taken so long to do it that the coverage was quite poor and that led to lots of defects later in the lifecycle or even in in our production environment, we just couldn't afford to do that. And actually, you know, focusing on continuous testing over the last 9 to 12 months has really given us the ability Thio delivered quickly across the the whole life cycle and therefore actually go from doing a kind of semi agile kind of thing where we did you use the stories we did a few of the kind of, you know, as our ceremonies. But we weren't really deploying any quicker into production because, you know, our stakeholders were scared that we didn't have the same control that we had when we had more water for releases. And, you know, when way didn't think ourselves. So we've done a lot of work on every aspect, especially from a testing point of view, every aspect of every activity, rather than just looking at automated test, you know, whether it is actually creating the test in the first place, Whether it's doing security testing earlier in the light and performance testing. Learn the life cycle, etcetera. So, yeah, it Z It's been a riel key thing that for for C T for us to drive, develops, >>talk to me a little bit about your team. What are some of the shifts in terms of expectations that you're experiencing and how your team interacts with the internal folks from pipeline through life cycle? >>Yeah, we've done a lot of work on this, you know, there's a thing. I think people were pretty quiet. Customer experience. Gap. It reminds me of a cart, a Gilbert cartoon where, you know, we start with the requirements here on Do you know, we almost like a Chinese whisper effects and what we deliver eyes completely, completely different. So we think the testing team or the the delivery team, you know, you know, you think they've done a great job. This is what it said in the acceptance criteria, but then our customers the same Well, actually, that's not working. This isn't working, you know, on there's this kind of gap Way had a great launched this year of actual Requirement Society, one of the board common tools Onda that for the first time in in since I remember actually working within B. T, I had customers saying to may, Wow, you know, we want more of this. We want more projects, um, to have a actual requirements design on it because it allowed us to actually work with the business collaboratively. I mean, we talk about collaboration, but how do you actually, you know, do that have something that both the business on technical people can understand? And we've actually been working with the business using at our requirement. Designer Thio, you know, really look about what the requirements are. Tease out requirements to the hadn't even thought off and making sure that we've got high levels of test coverage. And so what we actually deliver at the end of it, not only have you been able Thio generate test more quickly, but we've got much higher test coverage and also can more smartly, you're using the kind of AI within the tour and with some of the other kind of pipeline tools actually deliver to choose the right tests on the bar, still actually doing a risk based testing approach. So that's been a great launched this year, but just the start of many kind of things that we're >>doing. But what I hear in that Glenn is a lot of positives that have come out of a very challenging situation. Uh, talk to me about it and I like that perspective. This is a very challenging time for everybody in the world, but it sounds like from a collaboration, perspective is you're right. We talk about that a lot critical with Dev Ops. But those challenges there you guys were able to overcome those pretty quickly. What other challenges did you face and figure out quickly enough to be able to pit it so fast? >>I mean, you talked about culture. I mean, you know, Bt is like most come countries companies. So, um, is very siloed. You know, we're still trying to work to become closer as a company. So I think there's a lot of challenges around. How do you integrate with other tools? How do you integrate with you know, the various different technologies and bt we have 58 different whitey stacks? That's not systems that stacks all of those stacks of can have, you know, hundreds of systems on we're trying to. We're gonna drive at the moment a simplified program where we're trying Thio, you know, reduce that number 2 14 stacks. And even then they'll be complexity behind the scenes that that we will be challenged. Maurin Mawr As we go forward, how do you actually hired that to our users on as an I T organization? How do we make ourselves Lena so that even when we you know, we've still got some of that legacy and we'll never fully get rid of it on that's the kind of trade off that we have to make. How do we actually deal with that and and hide that for my users a say and and and drive those programs so we can actually accelerate change. So we take, you know, reduce that kind of waste, and that kind of legacy costs out of our business. You know, the other thing is, well, beating. And I'm sure you know telecoms probably no difference to insurance or finance we've got You know, when you take the number of products that we do and then you combine them, the permutations are tens and hundreds of thousands of products. So we as a business to trying to simplify. We are trying Thio do that in a natural way and haven't trying to do agile in the proper way, you know, and really actually work it paste really deliver value. So I think what we're looking Maura, Maura, at the moment is actually, um is more value focus? Before we used to deliver changes, sometimes into production, someone had a great idea or it was a great idea nine months ago or 12 months ago. But actually, then we end up deploying it. And then we look at the the the users, you know, the usage of that product of that application or whatever it is on. It's not being used for six months, so we're getting much we haven't got, you know, because of the last 12 months, we certainly haven't got room for that kind of waste and you know, the for not really understanding the value of changes that we we are doing. So I think that's the most important thing at the moment is really taken that waste out. You know, there's lots of focus on things like flow management. What bits of the our process are actually taking too long, and we've We've started on that journey, but we've got a hell of a long way to go, you know, But that that involves looking every aspect off the kind of software delivery cycle. >>What are some? Because that that going from, what, 58 i t stocks down to 14 or whatever it's going to be go simplifying is sounds magical. Took everybody. It's a big challenge. What are some of the core technology capabilities that you see really as kind of essential for enabling that with this new way that you're working? >>Yeah. I mean, I think we've started on a continuous testing journey, and I think that's just the start. I mean, that's really, as I say, looking at every aspect off, you know, from a Q, a point of view. It's every aspect of what we dio. But it's also looking at, you know, we're starting to branch into more like a AI ops and, you know, really, the full life cycle on. But, you know, that's just a stepping stone onto, you know, I think oughta Nomics is the way forward, right? You know all of this kind of stuff that happens um, you know, monitoring, you know, monitoring systems, what's happening in production had to be feed that back. How do you get to a point where actually we think about a change on then suddenly it's in production safely. Or if it's not going to safety, it's automatically backing out. So, you know, it's a very, very long journey. But if we want Thio, you know, in a world where the pace is ever increasing the demands of the team and you know, with the pressures on at the moment where with we're being asked to do things, you know more efficiently Ondas leaving as possible. We need to be, you know, thinking about every part of the process. And how do we put the kind of stepping stones in players to lead us to a more automated kind of, you know, their future? >>Do you feel that that plant outcomes are starting to align with what's delivered? Given this massive shift that you're experiencing, >>I think it's starting to, and I think you know, Azzawi. Look at more of a value based approach on. Do you know a Zeiss? A princess was a kind of flight management. I think that's that will become ever evermore important. So I think it's starting to people. Certainly realized that, you know, people teams need to work together. You know, the kind of the cousin between business and ICT, especially as we go Teoh Mawr kind of sad space solutions, low cold solutions. You know there's not such a gap anymore. Actually, some of our business partners expects to be much more tech savvy. Eso I think you know, this is what we have to kind of appreciate. What is I ts role? How do we give the capabilities become more for centers of excellence rather than actually doing Mount amount of work And for May and from a testing point of view, you know, amount, amount of testing, actually, how do we automate that? How do we actually generate that instead of created? I think that's the kind of challenge going forward. >>What are some? As we look forward, what are some of the things that you would like to see implemented or deployed in the next say, 6 to 12 months as we hopefully round a corner with this pandemic? >>Yeah, I think you know, certainly for for where we are as a company from a Q A perspective. We are. Yeah, there's certain bits that we do Well, you know, we've started creating continuous delivery. A day evokes pipelines. Um, there's still manual aspects of that. So, you know, certainly for May I I've challenged my team with saying, How do we do an automated journey? So if I, you know, I put a requirement injera or value whoever it is, that's why. Then click a button on bond, you know, with either zero touch of one touch, then put that into production and have confidence that that has been done safely on that it works. And what happens if it doesn't work? So you know, that's that's the next in the next few months, that's what our concentration is about. But it's also about decision making, you know, how do we actually understand those value judgements? And I think there's lots of the things Dev ops, ai ops, kind of always that aspects of business operations. I think it's about having the information in one place to make those kind of decisions. How does it all tied together, as I say, even still with kind of Dev ops, we've still got elements within my company where we've got lots of different organizations doing some doing similar kind of things but the walking of working in silos Still. So I think, having a eye ops Aziz becomes more and more to the fore as we go to the cloud. And that's what we need to. You know, we're still very early on in our cloud journey, you know. So we need to make sure the technologies work with Cloud as well as you kind of legacy systems. But it's about bringing that all together and having a full visible pipeline. Everybody can see and make decisions against >>you said the word confidence, which jumped out at me right away. Because absolutely, you've gotta have be able to have confidence in what your team is delivering and how it's impacting the business and those customers. Last question for you is how would you advise your peers in a similar situation to leverage technology automation, for example, dev ops to be able to gain the confidence that they're making the right decisions for their business? >>Yeah, I mean, I think the the approach that we've taken actually is not started with technology we've actually taken human centered design a za core principle of what we dio within the i t part of BT. So by using humans tend to design. That means we talked to our customers. We understand their pain points, we map out their current processes on. But when we mapped out, those processes also understand their aspirations as well, you know, Where do they want to be in six months? You know, Do they want to be more agile and you know, or do they want Teoh? Is this apart their business that they want thio run better? We have to Then look at why that's not running well and then see what solutions are out there. We've been lucky that, you know, with our partnership with Broadcom within the P l. A. A lot of the tortures and the P l. A have directly answered some of the businesses problems. But I think by having those conversations and actually engaging with the business, um, you know, especially if the business hold the purse strings, which is you know, in some companies, including as they do there is that kind of, you know, almost by understanding their their pain points and then saying This is how we can solve your problem We've tended to be much more successful than trying Thio impose something and say We're here to technology that they don't quite understand doesn't really understand how it could have resonate with their problems. So I think that's the heart of it is really about, you know, getting looking at the data, looking at the processes, looking at where the kind of waste is on. Then actually then looking at the right solutions. And as I say, continuous testing is a massive for us. We've also got a good relationship with capitals looking at visual ai on. Actually, there's a common theme through that, and I mean, AI is becoming more and more prevalent, and I know yeah, sometimes what is A I and people have kind of the semantics of it. Is it true, ai or not? But yes, certainly, you know, AI and machine learning is becoming more and more prevalent in the way that we work, and it's allowing us to be much more effective, the quicker and what we do on being more accurate. You know, whether it's finding defects, running the right tests or, you know, being able to anticipate problems before they're happening in a production environment. >>Welcome. Thank you so much for giving us this sort of insight. Outlook at Dev Ops, sharing the successes that you're having taking those challenges, converting them toe opportunities and forgiving folks who might be in your shoes or maybe slightly behind advice. I'm sure they appreciate it. We appreciate your time. >>It's been an absolute pleasure, Really. Thank you for inviting me of Extremely enjoyed it. So thank you ever so much. >>Excellent. Me too. I've learned a lot for Glynn Martin and Lisa Martin. You're watching the Cube?

from the silicon angle media office in Boston Massachusetts it's the queue now here's your host David on tape hi everybody welcome to this cube conversation I'm Dave Volante and you're cloud and cloud migrations are a major challenge for customers they move things into the cloud and variably they've got things that they want to maintain on Prem they've got to figure out what to move how to move it how to maintain performance and how to maintain the experience from on Prem into the cloud rate Krueger's here is a solution architect at net scout rate thanks for coming on nice to be here thank you so tell us a little bit about nesco yeah I mean net sky I mean primarily it helps you provide the visibility required to protect your digital digital business transformation we give you availability information performance information and security insights into what's going on in your environment we do this for 90% of the fortune 500 we do this for 95% of service provider so we're kind of Carrio class service provider and enterprise sophistication and we basically give you that visibility without borders and the visibility without borders is all about saying wherever you deploy your application whether it's being on Prem in your private data center software-defined data center or West en or whatever it might be or whether you migrate some or all of that into the public cloud AWS or as your we give you that same visibility same metrics wherever you host your application even in this hybrid world or this multi cloud world of today okay so top level one of those discussions like you heard my sort of intro and some of the challenges but what a customer is telling you about their cloud migrations well ok that's interesting so so that's kind of been around for eight years we're in like as I said thousands of customers and and these guys have been tasked they've been tasked with going to the cloud for business agility reasons and the idea of business agility is can you sort of create new services quicker new business initiatives new projects new application new ways that customers can we communicate with the business and they it's all about wrapping this and delivering these applications very quickly so the guys that we're talking to us said are being our task to move it to the cloud for various reasons it's not necessarily cost reasons as well it's LT and the the view is of the businesses the cloud will give them that agility maybe easier to manage maybe it's quicker to deploy applications quickly and all that sort of thing so they mean tasks to do that and that's a challenge because you know providing that visibility on premon in the cloud has been historically true well the other thing about the cloud is it's it's easy to test you know you test things you experiment you fail fast try the next one and it's relatively inexpensive to do that versus you know buying infrastructure but now so you see that but so talk a little bit more about some of the the real challenges that customers are facing you know when they start that migration as I said before they've got on-prem they've got workloads in the cloud they want a consistent experience but what are some of the problems yeah I mean yeah yeah yeah that's that shadow IT if you thought it has been a big problem but that's business utility isn't it okay because it's taken so long to deploy stuff on Prem ok to take four days before I have a new host ready for you to do that application so no wonder they've done that shadow whitey right but but anyway okay so on task to migrate this application so okay so I got to understand what that application looks like what are the components what it's what is it talking to because if I miss something right if I don't migrate all the components and don't forget these application it's not just one server or one component of the application it's maybe ten components might be whatever it is I need to know what that is and I can't just go to the documentation team to actually see all the protocols it's talking to all the dependencies whether it's one app tier talking to a database tier or whatever it might be the documentation just doesn't exist and the developers who developed that application no longer are part of the company they've long gone if ever they wrote any documentation so to understand right what you need to migrate is one of the biggest challenges and as it happens it's one of the challenges that we can help in netscape well this is a huge problem because you mentioned dependencies so if as you say an application talking to a database and maybe an ancillary application downstream those are going to affect business processes and unless you understand those dependencies if you effect one it's going to have a ripple effect on others and it could affect the business process so so that is a critical problem okay well so how do you nets go solve that problem I mean I have a question how does the industry generally solve it and I want to understand how you're different yeah okay so there's a couple of problems there is what one is understanding the components the dependencies and then one is understanding the performance so you can migrate successfully and all that sort of thing yeah so the industry typically will actually try and use some rudimentary network data to try and take a look at one application communicating to another and trying to get that from some devices various devices around the network because what they'll try to do to do that looking for connections is ok looking for connections and how they're doing that and in terms of performance they're they're resorting to looking at the different logs or the different infrastructure information like CPU utilization or those sort of things or developers are looking at instrumenting code into the applications which give them that performance information trouble with those they only see what the developers put into them rather than the whole picture of all those dependencies so while a bespoke data a lot of bespoke data trying to bring that together and come up with a conclusion that they this is all the components and this is how it's performing it is it's tricky ok so how do you guys do so yeah ok so as you know we use the network the wire data in order to understand what's going on so think about it if an application if I'm talking to my CRM application I might have a web browser it's talking to a web server talking to an app server to talking to a micro survey database or whatever it might be but all of those are interactions in a network different protocols HTTP HWS database Active Directory DNS so because we look at the network we can see it all so we can see all the traffic on the network we can see how things are communicating in reality so you don't necessarily need the documentation because we're documenting what's going on right now and that's kind of where we really score big in terms of understanding those dependencies and it's the it's the secret sauce that we've always known about the that that net Scout has your ability to to probe the network your your layer that analyzes that data the architecture that you've created right that's your IT yeah that's our secret sauce so we translate why data trauma is why data there's a lot of it and it's hard to interpret so that's one thing so we we've cured that problem by creating a patented technology called ASI adaptive service intelligent which translates that wire data into meaningful key performance metrics so you name the application it's all the applications going on your network translate them into performance metrics let's say application performance metrics and then differentiating that's a application latency from Network latency so we can see whether it was a network problem slowing things down or the application server slowing things down but also errors we can see all of that in that that wire data so that's that next layer up and then we have the analytics platform which we call ingenious one which actually takes that metadata and then allows us to display okay it's service dependency map so this is how your application is communicating all the nooks and cranny's the things that you didn't expect and not only does it do the dependency it does the performance as well the metadata oh it always comes back to the metadata one of the challenges that customers tell us they have is just creating the experience between on Prem and cloud you know the so called hybrid a lot of times it's it's different and they want to take that cloud experience and bring it to wherever they are cloud a cloud be on Prem are you able to maintain that experience in in this hybrid model yeah so to multi cloud or or not to multi class yeah no that's the beauty of number one why data and what we do why data is everywhere ok so if your applications communicate communicating in the cloud it's still communicating over IP and so we can actually instrument into the cloud collecting that wire data and then doing the same analytics asi in the same taking the same meta data and actually bring together a view of now the dependencies across the multi-cloud so whatever the cloud were able to get at that wire data and translate it into a si all uniform it's the same metrics okay so let's say we're out in a bar and you meet me and I'm an IT guy and I start chatting and I say hey I got this I'm doing this big project I'm really you know get this important it's got visibility at the board level and we're moving to the cloud and it gets your attention say whoo that's interesting and you start asking me to what advice would you would you give me I'm open to that okay obviously it's a talk to Nets character but the important thing is is this is that the question is that I've got a migrate this to the cloud and all that something and it's like sort of quite scary because I don't necessarily understand the cloud I don't realize that it's either the same or it's it's it's it's different or how its performing and it's I'm losing that visibility so you want to give that guy confidence you also want to give that guy the ability to say okay I understand the cloud and when things aren't the cloud I can continue to monitor it because that's after all the important thing so we've given them that confidence by saying hey we can instrument that application when it goes to the cloud and we can instrument beforehand so it goes it goes in the view understand what you're going to migrate all the components because you don't want to miss something migrate it and still have that visibility when it goes into the cloud we can give you that we give you this is interesting we give you access to that wire data when there are no wires that's to say the magic of nets carrots because we can instrument inside the workloads and get access to the traffic that's going in and out of those virtual machines those ec2 instances those virtual machines in in different clouds get access to that wire data and translate it into those key performance metrics and that's unique to Nets code like how do you do that well okay so the ASI is unique and the our agent technology is also unique to us to actually translate in the virtual machine in the cloud that wire data into metrics and then doing that all on the workload itself is very powerful if we can't instrument in the workload then there's another solution as well to get access to that wire data and that's what recently people like Amazon web services and as you I have announced the ability to tap in to that traffic so as you offer V tap which allows you to copy packets from VM to a destination which would be one of our probe technologies in the cloud Amazon have V PC traffic mirroring to actually get access to that data as well and we do the same thing the point is whether their workloads in the cloud workloads in the private cloud or the data center it's the same metrics and we get that visibility end-to-end visibility is the key ray thanks so much for coming on the cube and explaining so that your approach to a cloud and multi clouds great have you thank you very much you're welcome Eric thanks for watching everybody this is Dave Volante thanks for watching this cubed conversation [Music]