Today's leading minds talk Cloud with host David Linthicum
Named by Wired.com as one of the 10 most influential people in cloud computing, Bernard Golden has served as vice president of strategy for ActiveState Software, an independent provider of CloudFoundry. He is the author of four books on virtualization and cloud computing, his most recent book being Amazon Web Services for Dummies. he is currently the Vice President, Cloud Strategy at Capital One. Learn more about him at www.bernardgolden.com.
Dave Linthicum: Hey guys! Welcome to GigaOm Voices. It’s the one place where you'll get to hear from the industry thought leaders providing ‘no nonsense’ advice on how to succeed with cloud computing, IoT edge computing and cognitive computing. I'm Dave Linthicum bestselling author, speaker, executive, and joining me as our special guest and longtime friend, is Bernard Golden, a technology visionary. He's been able to accelerate the market success for innovative software companies, was named by Wired magazine's one of the top Most Influential People in the cloud computing space, author or co-author of five books on open source virtualization and cloud computing. He’s a computing advisor for CIO magazine blogging over a dozen Best Buy computer lists, highly regarded keynote speaker at conferences.
Bernard Golden: Well, your question was, “So what are you up to these days?” I joined Capital One about a year ago, as Vice President cloud strategy. And I had been acquainted with the company because it had been a customer of a software company I was at. And my perspective is that cloud computing is enabling but also forcing a fundamental change in the way that IT is done throughout really the industry and throughout our economy.
And I felt Capital One was really an early adopter of that perspective and you know it's ‘all in’ on cloud computing and working very hard to take advantage of that and when they contacted me and said, “You know, we're interested in having somebody… join us to help us with our cloud strategy.” I felt like it was a real opportunity to kind of get in the real avant garde – the vanguard of what's going on at the frontier of cloud computing. So I went ahead and joined.
I guess at Capital One, you’re probably one of 10 people in the whole world that can do the job, so they’re lucky to have you. So what do you do? ‘Day in the Life’ the Capital One… What does this strategic cloud person do?
Well I focus in four areas. And the first is advanced practices. How can we take advantage of what's going on with the most advanced cloud users in the world? How can we share practices with them, learn from them? Maybe teach some of the things we've learned, and I'll come back to that in just a second. Also advanced technologies. What are the frontier technologies that are emerging out of this new platform cloud computing, you know that that help organizations take advantage of it or push the envelope?
I mean, we have a very active machine learning… we have a Mac machine learning organization within Capital One. But there are other technologies dealing with the kinds of challenges that come with scale or the new kinds of architectures with microservices. We want to know, how do we take advantage of those advanced technologies. So I focus in that area.
I've been doing a lot of work around multi cloud—that's a third area that I focus on and, you know, how can we make sure that we have access to the most advanced technology being provided by any cloud provider.
And then fourth, I have a kind of an active… I call it brand evangelism. Or you might call it public communication, much like what I'm doing right here, sort of sharing a kind of ‘what I'm up to,’ my perspective, but also in a way that sort of talks about how Capital One is kind of unique in the industry.
Circling back to that advanced practices, one of the things that's quite interesting about Capital One is: from a technology user perspective, I've concluded on the basis of a lot of research, our technical peers are really
more like the cloud native companies, companies like Lyft or Netflix, Spotify, whatever—those kind of companies, you might want to think about.
But we're unique in that in addition to that, we're also a very large regulated financial entity. So we have a whole set of security and
cyber requirements, governance, risk management requirements as well as regulatory relationships that we have to adhere to. So it's a really fascinating place both from a technology perspective, but also applying it in a real world setting. So those are the four areas advanced practices, advanced technologies, multi cloud, and sort of brand evangelism public outreach.
Yeah, that's great. I think they're looking to have you. Yeah, and I think that’s no secret, based on the amount of time I've seen Capital One present at conferences, that they're probably leading the industry in utilization of cloud and new technology. And I think that's awesome because it kind of sets the path and sets the standard for other best practices and allows them to disrupt the space as a large competitor, which I think is something that's going to be admirable in the market coming forward.
So let's get into the topics. And I challenge you to look at a couple of topics that I wrote about. And of course, I agree with me, but I would love to get your perspective in as well. And the first one is three ways that cloud and data centers work well together. And the gist of this was, you know, someone kind of asked me [about] the fact that we need to think more about how legacy systems and traditional data centers work and play well with clouds.
You know, I think ultimately that's what many people consider hybrid cloud today, so hybrid cloud used to be a very private and public cloud. Now, now they talked to the notion of hybrid IT, which basically is anything that's a traditional legacy environment, could even be client server on the Internet, starting on the land kinds of things that interacts with the public cloud environment. And the thing is, while we're moving into the public cloud, and we're also progressing forward their legacy infrastructure, not a lot of thought in terms of how that technology is going to work and play well together has really been around. We have certain tools and techniques and we certainly have synergistic security systems and governance systems, monitoring management systems… but really no good big thinking in terms of strategic ways in which we approach it.
So I mentioned a few things in the blog, but I'd love to get your perspective on this in terms of how enterprises should be thinking about approaching hybrid IT and making their existing legacy systems or data centers work and play well with the public cloud providers.
Yeah, well, you know, let me preface my remarks by saying Capital One's perspective on this is: it's decided to go all in on public cloud, it sort of and this is before me. I don't want to claim credit that you know it's somehow I brought that decision to Capital One. It had made that decision, but it was based on the conclusion that if you really accept that every company is a software company is, you know, is often the kind of the cliche in the industry or IT is now how you run the business, Capital One concluded our core competency is going to be about building applications to help us do that, not about running data centers. So it's a bit different from this.
But I have lots of experience working with companies that you know maybe aspirationally would like to go to that extent, but, you know, still have to balance, you know, existing data centers, maybe financial situations that they can't step away from. And so I do have a lot of experience around that and I think you're really right on this—that there was a sort of a mania of, like, ‘I'm gonna build my own private cloud’ and the bloom seems to have gone off that rose and now people are saying “You know, I'm not going to invest a huge amount of money into making an all fancy data center. I want to get value from the data center. I don't want to invest a ton more incremental money in it. So how do I integrate that with this new world of applications?”
And so we definitely see that you identified several things, one of which was, ‘How do you connect data that's going to live in your public cloud environment with the data that exists in your existing environment?’ For many companies, the majority of data will be in their existing environments. You talked about sort of bringing up your data storage your databases into the same you know at the same level so they can communicate, which I think is a very insightful approach.
Just as important in in my experience is to make sure that you've got the right kind of bandwidth and connectivity in those environments so that you can cross connect that data. The hybrid world is one that's going to require shuttling a lot of data back and forth, and so you got to make sure you've got the bandwidth, the security and so forth of your network.
I can say that our network group lives in the same organization I am in. And so I interact with them and listen in to them a lot. And that's, you know, a huge amount of effort and investment Capital One's made to make sure we've got that kind of connectivity into our cloud, and so forth.
You know, you've talked about common security—a security approach and technology that can span both public cloud and on premise systems. A common directory… yeah, for sure you need to have that. I mean, there's no question. You've got to have common identity management or else you are in a world of hurt that you don't even know that you're in a world of.
Unfortunately I see ultra-frequently that companies do that. They've got their on prem identity management, then they use the cloud native one provided by the provider and then they wonder why things operate differently or people have different permissions or worst of all: somebody gets walked out the door and all the internal systems are shut off to them, but they still have the private key to get at systems running up in the cloud.
Well, that's so common security is extremely important. And you know, fortunately, the providers have made good strides in being able to have Federated Identity Management where you really use your corporate identity management as kind of the single version of truth, or whatever you want to call it—and use that as the basis for providing access and authentication in the cloud.
So that is what… You just hit upon the big topic. So going forward, I think enterprises kind of want a procedure and a process for taking a look at this and what would you recommend as far as step one, step two, and step three? You know: what they need to do the first year in the second year to make the most out of this because it's always like everyone tells me about a diet. I always say, “Well, what are you eating for breakfast, lunch and dinner?” This is more descriptive than someone just telling me what mix of carbos and things like that that they're eating. So what should be step one, step two, step three?
Well, I mean, I think a lot of times the right step one is, sort of: Get your feet wet, which is learn about cloud. But I think if I can use the kind of cliche, the journey of 1000 miles begins with a single step. Okay, that's the right first step, but you have to know that you're going to be on a journey.
And so you have to start planning for these kinds of things as part of that first step, you know. I'm going to start moving production systems into a cloud environment. What do I need around this identity management to make sure that I've got security in place? How do I ensure that same practices and requirements that I have for my on premise systems, I can extend those? And what tooling, what processes, what personnel and maybe what new tools am I going to have to use in a cloud environment, or how do I map my current practices into what's available in the cloud?
And that's where I see a lot of companies sort of fall short. They had the first flush of victory, and then it's kind of a land grab without necessarily recognizing that there needs to be more of a governance process and a structure about how the next wave of application deployments go that are surrounded by these kind of core capabilities that you've identified, that you've put into place over a period of 30 years in your own data centers.
And that's super important and it's incremental investment that it's critical that the companies plan for. It's like you're going to need to develop new practices in addition to what you've got. That means you're going to spend more money on those kinds of things. You're going to need new skills, probably going to have to train or hire in new talent. That's the kind of next step to go to production systems.
And then you know you've got lots of stuff like: So what are you going to do around your network? How do I make sure that I've got the right bandwidth and how do I make sure that the traffic is secure when transiting between my site and the cloud provider site, how do I make sure that's in place? So a whole range of kind of practical blocking and tackling kinds of things need to be put into place. That would be the breakfast, lunch and dinner recipes.
And we bring that back to the to the to the ask, so ultimately, and this is kind of the final question on this topic: Are we diluting the value of cloud by actually creating links to the infrastructure that should this be kind of an ‘all in’ approach where we kind of create a demarcation line between the legacy systems and cloud computing? And even though there is some cross integration, it's kind of not the core focus? Are we just putting off the inevitable, or is this something that's going to be reality as we move forward for the next 5 to 10 years?
Well I think a huge mistake that I've seen many companies make during my consulting days and also while I was working at software companies is they use a mental map and they use a process where they go ‘oh cloud is like my data center, but at the end of a wire. I'll leave everything just the way it is—the same tooling, the same practices, so forth.’ And I just went through a fairly extended description of like: you need to make incremental investment to adapt those practices for the cloud. And I think that as a cloud adopter, you've got to think, ‘How is this going to let me leverage all the goodness that those cloud providers offer that I can't get today?’
And so there has to be a demarcation point in the sense of this is the way I do things in this sort of static limited world that I've lived in for the last 30 years. This is the way I can do things now and then, you know, what are the right demarcation points? What is the right connection just talking about the network and so forth and so on.
But you have to have the aspiration and the ambition to say “I want to become like a cloud native company in terms of the way I use cloud. It's not going to be just my data center thing at the end of wire.” You know the people who really take advantage of cloud… and that's a huge impetus, because the people who do that first are going to gain tremendous business benefits and then bring competitive pressure to those that lag behind.
And you have to kind of decide, ‘Do I want to be an adoptor or a laggard?’ I truly believe that this technology capability is going to be such that it will provide business benefit that will exhibit itself as competitive advantage and you don't want to be on the wrong side of that competitive advantage/disadvantage divide.
So changing gears a bit. Let's talk about edge computing. Ultimately, the ability to kind of leverage those computing as the ability to kind of put a central point of processing between centralized cloud system and some sort of a place where the data is actually gathered. So to do immediate responses and not necessarily have to deal with the network latency is becoming adopted by many enterprises out there.
So ultimately, what will your guidance be in terms of integration? Integration of this links back to our previous topic as well—integration of edge computing with public cloud providers. What are the core ‘do's and don'ts’ and what should be the tradeoffs considered? What kind of application should live at the edge and just some general guidance.
Yeah, you know, it's, it's an interesting thing, because I, you know, I've heard people opine that edge computing means ‘the cloud is dead.’ And I don't think that's accurate. And in fact, I think it's a dangerous sort of mental model to use. I mean, I think about edge computing as where digital meets analog, right? It's bits meets atoms. The use case for edge is you want to be close to something happening in the environment, the physical environment and every use case is different in a way that running one program versus running another program is rather similar typically, but the difference between ‘I'm going to check the temperature out in an oat field’ is really different than ‘I'm going to have a self driving car.’ They're both edge computing. They're both at the at the interface of
bits and atoms, but they're vastly different kinds of use cases.
And so the edge computing what you need to put out in place at the edge—is very different, and you kind of made mention [of it] in your article which is: a lot of what's put out at the edge—the computing power out there is something like a Raspberry Pi. And so if you’re a user and that's the amount of functionality you need, you need to ferry data back. Whereas if you're doing something worth a lot of real time very heavy processing, like what you want in a self driving car, you're going to need a lot of processing locally, and so you've got to really think through ‘What is my use case?’ And for what it's worth, my opinion is, there's going to be a lot more: small amount of processing locally and large amount of processing being ferried back to the to a central place like the cloud, than is currently sort of posited by the ‘edge is going to make the cloud die’ school of thought.
You have to really think through what the specific use case of your edge is and then figure out what's the right computing model, data model, local processing capability, data transfer ability that you need support that particular use case.
Yeah, I agree. It's use case dependent. I mean, it's funny, as I'm building some edge-based stuff now and as I'm finding that the biggest issue is keeping things off the edge-based computer because you can saturate that fairly quickly and making sure it's on the back end cloud environment. Also the ability to kind of see you’re dealing with something which is really around tiering.
So in other words, the first tier being the edge or first tier can be the client, which is doing some rudimentary processes and anything that needs a response instantaneously, which is usually less than 10 - 20% of the application. And all the big data processing and big operational process occurs in the back end and making sure that lives in the cloud. So that's the best place for it to live on, number one because it's more secure, typically.
And you're able to do a lot of heavyweight process in which you can't do in the edge device, but in many instances, you know, people kind of violate that they put as much as they can, on the edge and ends up being saturated and they have to retrofit in the back end. So in one minute or less, what advice would you have for them in terms of how these things they do exist on a particular tier?
Well, I mean, really, again, the first place to start is: ‘What is my use case? What am I going to accomplish?’ And then figure out how much of that has to be processed and executed at the site and how much doesn't. And I sort of agree with the implication of your statement, which is: anything that doesn't have to be local should be ferried back and done centrally because that's a place where you can have more control, you can have more security, you get greater economies of scale, you have a better opportunity to administer and so forth.
That's one of the challenging things about edge is. It's this concept edge computing. It sounds like it's all one thing. It's a bazillion different use cases—each of which has to be examined by the particular person or group actually implementing it to figure out ‘How do I partition the app. Where do I keep the storage? What gets transferred what stays locally?’ So you start with the end in mind and look at the use case to you're trying to implement.
I can’t argue with that. That's great advice in terms of edge computing. So anyway, please pick up a copy of my book Cloud Computing & SOA Convergence available on Amazon and other places books are sold. Make sure to follow me on Twitter, @DavidLinthicum, as well as LinkedIn where I have several cloud computing courses on LinkedIn Learning. Where can we find you on the web?
Well, of course, for my personal stuff, you can find me on Twitter, which is @BernardGolden, and you can find me at my personal website BernardGolden.com I also do a lot for Capital One, a lot of activity, publish articles on the Capital One Medium site. So you can just do a search on Capital One around that.
You may be able to see me… I speak at conferences quite a bit, and I'm starting a new podcast series called the Cloud Strategy Podcast, which will get launched pretty quickly. And you should keep an eye out for that because I'm gonna be talking to interesting folks, such as yourself, and you know, take a look for that. That'll be, I hope, a pretty interesting place, looking at lots of different interesting topics, talking with really interesting people looking at interesting technologies.
Keep an eye on Bernard. He's the best in the industry and I’m sure he's one of the better voices out there that understands what he's doing. Like I said there's very few people in the universe who know that. So until next time, and we’ll talk again in about seven days. You guys take care. Bye!