Cloud Computing Myths, Falsehoods and Outright Lies
 
 
 
 
 
 
Cloud computing advocates note that it's one of the great advances in tech in a generation. But is cloud truly without its challenges?

See transcript of video discussion below:

Cloud computing, though it enables great promise, also offers significant challenges. For businesses seeking to get the most from their cloud deployment, it's important to understand the myths and falsehoods that surround cloud computing. These contradictions center around concepts like vendor lock-in, the future of the private cloud, cost issues, analysts vs. customers, and role of the datacenter in the years ahead. Complicating matters, what's true and what might be true is hotly contested. To discuss these issues, I'll spoke with three top cloud computing thought leaders.

Greg Knieriemen, Technology Evangelist, Hitachi Data Systems

Andi Mann, Chief Technology Advocate, Splunk

Scott Sanchez, Director, OpenStack, Cisco

Transcript of video discussion: 

J. Maguire:

[kidding] So, cloud computing. It’s pretty amazing that it’s so amazingly easy. You know, we just log on with our app to your remote location. There’s no more IT department needed . No more data center. It’s all public. It’s all handled for us. And it’s cheap. It’s gotten cheaper than ever before, so I mean, it’s pretty amazing. I mean, why do we even need cloud experts because things have gotten so easy? Andy, is there any more use for expertise in the cloud due to how easy it’s gotten?

A. Mann:

Oh James, where do I start? Talk about myth busting. Let’s take that one. It’s not easy, right. Even just selecting a cloud is not easy. Let alone starting and using one, getting the skills. You, I know, pointed out some articles on the skills that one should use. There’s… don’t get me wrong. I’ve always been a big supporter of cloud computing, of course. But yeah, it’s not as easy as people make it out to be. The simplicity of it is a huge challenge. And then we’re getting the things like the fragmentation - what I call a cloud of clouds, what I know some people call the multi-cloud, I think. It’s a lot of things, but it’s not easy

J. Maguire:

What is the fragmentation of the cloud? I bring it up because I just saw a research report recently that said, on average, companies have like six different cloud instances? What’s going on with that?

A. Mann:

Yeah, this is something I’ve been working on quite extensively recently. The fragmentation of IT in the cloud is just generally a part of that. There’s no one cloud to everything. There’s no one cloud to rule them all. To start with, there’s the different types. There’s the SaaS, PaaS, and the infrastructure, and the community - public or private. But different use cases, different types of cloud, especially in larger organizations where you’ve got empowered developers and all of a sudden they’re putting down credit cards. Then add to that business owners and business people in sales sick of the IT people not coming to the party in time so they put their credit card down for a Salesforce.com or a Marketo or all sorts of cloud based services. And all of a sudden this proliferation happens and you’ve got just fragmented cloud. You lose track of it. You can’t find performance. You can’t find cost. You end of paying money for services you’re just not even actually using anymore because someone forgot to shut them down. So yeah, the fragmentation is an issue in itself.

J. Maguire:

Greg, the easiness of the cloud. Are you enjoying the free and easy ride?

G. Knieriemen:

Yeah, I don’t know why were having this conversation. What’s the problem? You know, James, you talk about some of the myths and lies when it comes to cloud. The one thing that keeps coming up… it’s funny, I was looking back at some of the debates that Adrian Cockcroft, Christian Riley, Chris Hoff had back five years ago debating cloud types, and still today we’re arguing quite frequently… not arguing. I would say there’s a fair amount of discussion around… it boils down to one central point. You’ve got a core set of analysts and some in the media that think there’s only one way to do cloud, especially Infrastructure as a Service. I think that’s the biggest myth that’s out there. I think when you look at some of the recent articles about Bank of America, Goldman Sachs and other folks, they’re doing cloud differently. It’s not a ‘right or wrong’ judgement, but clearly it shows that there’s more than one way to do cloud.

J. Maguire:

Well, I hope they’re not doing private cloud, because for the private cloud it’s a waste. The only people who do private cloud are those who don’t get it. So, you’re not saying BofA or Goldman Sachs are doing private cloud, are they?

G. Knieriemen:

No, I was reading back, again, I’ll refer back to Adrian Cockcroft five years ago saying, ‘There’s no technical reason to build a private cloud’, and I think he’s probably right when you talk about technical reasons. But, the technical aspects for why you deploy a cloud are one of many aspects including business, including culture, that really kind of factor into how companies are going to leverage their IT resources and what that looks like.

J. Maguire:

And compliance, of course, for those two giants. Scott, the easiness of cloud? Are you riding on that wave? Or what’s your taking on that?

S. Sanchez:

Well, it is easy if you are developer. If you’re an application developer, like a true, legit application developer, and you want to build an app, yeah, it’s easy these days. You go out to a public cloud, or hopefully you’ve got an IT team that has seen the light and has implemented an actual private cloud solution that feels cloudy, so, from a developer’ perspective, things are great. Things are better than they've ever been.

J. Maguire:

That was the AWS story early on. All of those developers getting on there.

S. Sanchez:

Right, and I think when you looked at private clouds five or six years ago, really what people thought was a private cloud was really just fancy virtualization, and maybe it had… I’ll just say it, some crappy dashboard that people could go to and click some buttons. Well, the line of business end developers that have gone out and had an amazing public cloud experience and came back and were like, ‘Yeah… I don’t know what you think this is but it’s not really a cloud’. So, I think that’s part of the problem, still. So, it’s very easy today, if I’m a developer, to go get what I need done. Internally, if I have a good IT team. And externally with any public cloud provider.

It’s very hard if you’re IT today, though. Cisco has a service called Cloud Consumption Service, which kind of scours the network and looks for all of the traffic, and very often the results that come back they ask the CIO, ‘How many approved cloud services do you have?’ and they’re like, ‘Oh! We have 35.’ And like 800 come back in the result. So, it’s staggering and startling and everything else and then they start to ask, ‘Okay, how do we get a handle on this?’ I think it’s very difficult. Just the proliferation of a lot of these public services. But also, anybody can go out and download some of the best database technology in the world or the best whatever in the world and put that onto their network without even asking IT anymore. So, it’s getting harder internally, not just externally.

J. Maguire:

Which is the legacy of shadow IT, which is a bad thing. It seems like it’s gotten more accepted because people couldn’t wait around for the IT departments so they just used the company credit card. It seems almost like shadow IT has been absorbed into the company culture in many companies.

My favorite myth, my personal favorite, is that cloud allows you to avoid vendor lock-in. Like, if you just.. in the bad old days if you bought from ACME company. You had their vendor gear and their server gear you had to stay with them no matter how they raised their prices. These days you can switch around so easily because it’s just a cloud. Greg, is there vendor lock-in with a cloud? Is that a reality or is that a myth?

G. Knieriemen:

It’s funny, we talked about virtualization technology. Storage virtualization, server virtualization technologies, and the myth there was now that you abstracted your servers or abstracted your storage resources you could avoid vendor lock in. The reality is vendor lock in just moved up the stack. And I think we can look at cloud as the same way. For argument’s sake, talking about AWS for a second. The cloud doesn’t remove vendor lock-in. In fact, I would argue that is probably solidifies vendor lock-in. I think, of course you know I’m a big advocate for OpenStack, but if you look at OpenStack for technologies, there’s your best opportunity to avoid vendor lock-in. I don’t think public cloud does that.

J. Maguire:

Andi, is public cloud necessarily vendor lock-in, or is that a myth? And how do companies work around it it if is real?

A. Mann:

I’ve always had an issue with the whole concept of vendor lock-in. Every software tool does something specific in its own specific way, unless you’re talking about total copycats, which, you know, those actually do exist and have their place. But on most software you have not just the software itself, but you build up skill sets, you build up processes around it. Some of it is more harder than others to move away, but the whole concept of vendor lock-in I find a bit of a furfy. Every cloud is different, and sometimes that’s a good thing because individual clouds will do things that are specific and useful. So, you want to choose your cloud based on what you want to try to achieve. Now, if you want to try to move from one platform provider to another that’s often going to be difficult. The challenge is there why do you want to do that? Are you moving for a specific feature? You’re going to lose some and you’re going to pick some up. This whole idea of vendor lock-in has always, I think, been a bit of a furfy. Cloud certainly doesn’t make it any easier, I don’t think.

J. Maguire:

Can you define that term ‘furfy’ so if I’ve not been to Australia… I want to use the word ‘furfy’ in a sentence.

A. Mann:

[laughs] That might be a little big of a colloquialism, I’m not sure. A Furfing is a false trail.

J. Maguire:

So, in other words, you’re saying that vendor lock-in is almost part of the beast no matter what. It’s not something to be so worried about because it’s just part of using something all the time anyway?

A. Mann:

Yeah, exactly. Some lock-in is hard. If your data is involved and structures and that sort of thing then it’s obviously a lot harder. I’ll scale up to architecture a little bit. Code can be moved more easily and I think that cloud does actually make it easier in some ways for that, especially when you start to think about containerization and micro-services. That starts to become a little bit more portable. But frankly, I’m not sure that it is ultimately that portable, but I don’t know the math.

S. Sanchez:

But then you’re locked into... [Greg’s comment]. You’re locked into the next stack up the ladder. Oh great, I’m not locked into my physical hardware anymore because I have this operating system. Well, now I’m locked into the operating system. I’m not locked into the operating system because now I have virtualization. Now I’m locked in there. Now I have OpenStack or ‘pick your orchestration layer’ on top of it. Now I’m going to fix that. I’m going to get a DevOps Automation scripting platform. Well, now I’ll write all of my scripts in someone else’s format. Oh, I”m going to skip that and I’m going to go with a cloud management platform. Well, now I’m locked in there. I don’t think… and containers are just another form of that. I don’t think public or private cloud or anything changes that. I think this is about making bets on technology that are right for your business at any given point in time. And I think what cloud and the shift to easy access to resources, I think what that has done for us is in the industry is really to give you the ability to not be locked in so broadly across your portfolio of applications. So, maybe you’re locked into a particular app or a particular business unit or a particular set of projects, but you’re no longer really locked in across the board as a company the way you might have been had you had to make a 300 million dollar investment and bet on something. Now you can make a $300 bet on something.

G. Knieriemen:

That’s right. James, I just want to piggyback on what Scott just said there. If you look at the white board behind me, this goes back to the cloud arguments from five years ago. Christopher Hoff said, ‘Use the right tool for the right job at the right time and the right cost.’ Five years ago. It still applies today.

S. Sanchez:

There you have it.

J. Maguire:

Yeah, it’s the universal truth. Scott, what is one of your favorite myths or falsehoods about the cloud? When people talk about what leaves you saying, hey, that can’t be true?

S. Sanchez:

I think it’s that you have to… and this sort of touches on what we were just talking about, that you have to pick one and that’s your strategy. That’s a very kind of old school mentality to me. I’m not making a company-wide ‘we’re going to forklift and burn down the old and move everything over here’ you know, even some of the folks that were onstage at the last Amazon event. They very carefully used their words to say ‘We’re not going to be in the data center business anymore. We partnered with Amazon.’ They didn’t say they’re getting rid of all of their servers. They didn’t say they’re not going to have any applications that run on stuff that they might manage. They just said that they don’t want to light up a building that’s a data center. So, I think that’s a big distinction. People look at it as I have to make a decision on either cloud or I’m not. That’s, to me, one of the most frustrating myths. How do I get my 3,000 application portfolio over to cloud? Well, you don’t. You may transports 50 of them this year. You may build a hundred new ones. But the other 2,950 are probably just going to keep doing what they’re doing. I think that’s a myth we have both collectively put into people’s minds and now have to remove.

J. Maguire:

So all cloud is hybrid cloud, or it’s a hybrid scenario one way or another?

S. Sanchez:

It is to me. I’ve been drawing up this tic-tac-toe board on the white board and kind of explaining, hey, if I’m a bank, or example, when I’m building the next cool mobile app for my bank, you can fill in one of the boxes on a tic-tac-toe board with a piece of my app. There might be an Oracle database up here, or there might be something bare metal. Or there might be a piece that runs in VMware, OpenStack, public cloud, some SaaS service, micro services, whatever. Fill all of that in together and that’s what really makes up my mobile app these days if I’m a well established organization. It’s not, ‘I’m either in the cloud or I’m not.’ It’s not even ‘I’m virtualized or I’m not.’ There’s this real mishmash of stuff that makes up what people think of as an app and I think that’s a really good mindset for people to start adopting.

J. Maguire:

What about… and please just educate me on this because I’m honestly confused. Thinking of PaaS. Sometimes I think it’s charging upwards. Sometimes I’m not sure it’s real. Andy, are you a believer in platform as a service?

A. Mann:

Oh geez, I go back and forth on this, James. I’ll probably get into trouble. I think, to a large degree, PaaS is about standardization and we can all agree that cloud is about standardization, at least until a certain level. So PaaS, I think, is a great thing because any way I can get up and running, get content running faster, easily lift and shift into a test, a Q/A, a staging, a product using the same platform, the same environment, the same stack, that’s a great thing. But I’m not sure I’ve ever met two developers using the same stack, let alone entire teams or companies. So, I don’t know. I see great success in certain companies that I work with using platform as a service and doing really well with is. So, I do see success. I do see positive aspects in it. But I also wonder if it’s everything it’s cracked up to be, because again, pick your use case, find the business value, yes it’s going to be… but, I’m on the fence on this one, James. I’m still waiting to be convinced, I guess.

J. Maguire:

Is there a leading platform or a couple of leading platforms where someone could say, you know, so-and-so is PaaS, Azure is PaaS, that you think are really the dominant ones if it’s going to really happen?

A. Mann:

Obviously, the CloudFoundry is doing really well. The Pivotal guys, they’ve done some great work. I’ve talked to a bunch of their customers who are super happy. I know in the past five years RedShift. The RedHat. I’m sorry, OpenShift. RedHat PaaS. Those are fantastic. We’ve got a whole SaaS infrastructure using the…

J. Maguire:

Did we lose Mr. Mann? I think he kind of froze. You’re back. Okay, you froze for a second, Andy, after the RedHat piece you froze a little bit. We have you now.

A. Mann:

Good, yeah. Occasionally I get lost. Occupational hazard. Yeah, so I the OpenShift, the Pivotal. There’s a bunch of them around, as well, that I don’t know very well but I know have very good people. So yeah, I think it’s still nascent. I think there’s still opportunity for it to be really successful. I do think they have to make their case a lot more succinctly and clearly.

J. Maguire:

Greg, do you have a sense of the whole PaaS world? Is it happening in your eyes?

G. Knieriemen:

Absolutely, it is. I think the challenge there is it’s not just about establishing a platform as a service. Also, we talked before that the movement to cloud is not just about the technology but about the culture and business cases, and I think that platform as a service gives you the ability to help build that culture and help get everything moving in the same direction. So, I think philosophically, platform as a service adds an incredible amount of value for most organizations. But, I think it is still very early in the market and I think that whether you look at RedShift or Cloud Foundry or Pivotal, there’s lots of opportunities there and I think this is a good opportunity for companies to explore their options with that because I think we’re very early on with it. I would just go back to emphasizing that there’s just no one way to do cloud.

J. Maguire:

Scott, are you a believer in platform as a service? Do you think it’s going forward?

S. Sanchez:

I would say that PaaS is very prevalent with our cloud customers. I don’t want to quote a statistic, but I’ll just say that an overwhelming number of our private cloud implementations include a PaaS component. I see a ton of value in platform as a service when… I forget, it might have been Andy that touched on this a minute ago, but when it uses a tool for the development teams to move faster, then I really like it and I think it’s fantastic. When it’s used as a choke point or a control point for IT because they look at it and they say, ahh, these people are using all of these tools, they’re using all of these different platforms, everything is configured differently, we can’t manage this. We’re going to stick a platform as a service in there and that will be our lockdown control point. If you can do it, here’s your guardrails, have a nice day. I hate that model. I hate the idea that 20 years of IT scar tissue policy are going to define how we can use some of these new tools. And so, if it’s a business driven implementation of a platform as a service, which is most of what we see, I really like it because, yes, you will be able to write code, push code, run code. It’s all about going fast. If it’s IT pushing it I’m much less of a fan.

J. Maguire:

Speaking of code, since we have Andy here with us. Is there a DevOps/Cloud myth? I think you practically wrote the book. Is there a myth we need to know about when it comes to dev ops and cloud?

A. Mann:

Oh look, I wouldn’t say I wrote the book. I read the book, though. It was pretty good!

J. Maguire:

It’s important to have read the book if you haven’t actually written it.

A. Mann:

Yeah, I definitely read the book. There is definitely a component to it. I have written a lot about the intersection of cloud and devops. You know, when you talk about devops as a cultural change and a way to write better and to move faster, clearly the ability to have a common cloud for a team to use, taking the operations barriers away from developers to be able to use their services, to be able to jump on and try to things, to be able to iterate really quickly. That’s one of the things that cloud is really good at. You can stand up a service very quickly, shut it down, try again, fix, go. Iteration is great, and devops is a lot about constant iteration, small pieces, feedback loops and that is great for that. There’s a big intersection between devops and cloud. Really positive synergy there.

J. Maguire:

So, the myth might be that devops is to be avoided, or is there a falsehood in there or do we just basically believe in devops?

A. Mann:

I don’t know. I’m a true believer, let me tell you.

J. Maguire:

You are. And Greg, you and I were talking a little about the politics of chargebacks, like, charging for cloud use within the company. That’s kind of interesting, the way it gets charged across divisions. Talk a little bit about that.

G. Knieriemen:

Sure. The reason why this came up is there is a little bit of a debate, I think, within the analyst community, I don’t think end users debate this, but certainly in the analyst community, about how do you define what a public cloud is. How do you define what a private cloud is. And one of those nits for me has been when analysts try to say that charge back or show back is requirement for you to have a private cloud. A lot of organizations and a lot of large enterprises, the whole concept of charge back or show back is extremely controversial and extremely political because they’ve built their IT infrastructure not based on business lines or departments. It’s a shared service across the organization. So now when you try to apply a strict set of rules, which is what charge back exposes you to, as far as who’s consuming how much and when, that can create some animosities and some political challenges within an organization because you may have an unprofitable side of your business, maybe a development team, that is really doing great work but they’re not able to show profit back to the business. So to have that aligned against their cost creates all kinds of political controversy inside of a company.

I’m going to say - and I’m going to be curious to hear what Andy and Scott say - I’m going to guess that probably great than 90% of Fortune 500 enterprises have pushed back against the concept of charge backs, at least from what I’ve seen. That number may be a little bit lower, but I think it’s safe to say that the majority of enterprises are not keen on charge backs. And I think to create that as a requirement… you know, I get the analysts’ need to put some definition around what a cloud is. Their job is to measure and report back to their constituents and their clients these numbers about cloud adoption, but I think that it’s very very difficult, especially in private cloud environments, to pin that down and put charge backs squarely on a definition for a private cloud.

J. Maguire:

What about, in the time we have remaining, looking into the future. What is cloud computing going to look like in 2-5 or 3-5 years from now? Maybe some of these messes will get cleared up. Maybe some will remain with us. What is going to be happening in cloud in the next few years? Scott, what’s your take on sort of the immediate and near term future of cloud?

S. Sanchez:

Oh, that’s an easy question.

J. Maguire:

Oh, good. It all goes to Amazon Web Services… It’s so broad, I guess it’s my way of saying, Scott, what do you want to say? Say whatever you want to say.

S. Sanchez:

What do I want to say?

J. Maguire:

Yeah, or we can just go out to your pool if you want. If you have a glass of wine handy you can just go out to your pool.

S. Sanchez:

Yeah, let’s absolutely do that. But, I think it just continues to… the idea of IT continues to move further up. Today we talk about PaaS and containers and dev ops and there’s not a lot of people that come in and say, ‘I’d like to have a serious conversation…'

J. Maguire:

Where there might have been in 2010 or something, you mean.

S. Sanchez:

Well, yeah. You know, people were like, ‘Well, cloud is just fancy virtualization.’ And now I think people kind of realize that it’s not. And even devops is not about the scripting language you write things in. It’s the mindset that brings you to a wholly different way of thinking about IT. I’ve seen, I’d say in the last 12-18 months, I’ve really started to see an awakening with some of the larger enterprises and realized it is very much about the apps. It is very much about the data. And, you know, IT and the IT teams and data centers are all still very important. All the hardware? Very important. But the conversations are finally moving to the point where they’re about ‘what do I do with all of it?’. And a good business outcome, no secret I work for Cisco, the idea of business outcome has shifted from something like, hey, IT guy or gal, we’re going to shave off one millisecond from your ping times. And that was like a business outcome five years ago. Now, the business outcome is we’re going to help you figure out how to place that trade one millisecond faster than your competitor. So, that’s the conversation that we go into. So, when you ask me where does cloud go? Cloud goes further and further away from things that fit in a 19-inch rack. In terms of a conversation, those things are still there, they still show up and they’re still super important. But the conversation around cloud keeps moving, truly, into what you do with it. And I’ve only seemed to see us approach the tipping point of that probably in the last year.

J. Maguire:

You mean, cloud is being talked about in the C-suite in a larger and strategic sense more? Is that what you mean?

S. Sanchez:

No, I mean we’ve almost stopped talking about cloud as, you know, for five years you can’t say cloud without thinking about OpenStack, right, in my world. And now, you know, OpenStack is near and dear to my heart. The next summit is next month here in Austin after six years. The reality is cloud isn’t OpenStack. OpenStack is an ingredient into a whole organizational change that leads you to do great things with technology. And I think that’s finally where we, if you want to call us experts, that’s your word, not mine, but if you want to call us experts on cloud, those are the conversations that we need to be having. And two or three or five years from now I don’t even want to be using the world ‘cloud’. I just want to be talking about great things we’re doing with technology. Cloud is one of those things that we’ve stopped talking about because it’s more of an idea as opposed to a piece of technology.

J. Maguire:

Yeah. Greg, looking into your crystal ball for the future - what do you see inside the year 2020 whereabouts?

G. Knieriemen:

I think Scott nailed it. I think we’re going to start talking less about cloud and start talking about more applications and services and value that IT is driving. I think it’s more tangible to get into services and applications. That’s going to be the biggest difference. I think the tone of the conversation is going to change. it’s going to be less about what’s behind the screen and more about delivering and how we’re performing against what the expectations of the business are. It really is entirely about business outcome.

J. Maguire:

Andy, your take on the future? And would you please use the word ‘containers’ in your answer? Just as a side change, if I can share this - It’s so interesting the way VMware was so disruptive in it’s day, and then containers came up and VMware seemed like it was slowly threatened by containers. But VMware sort of put it’s arm around containers and now VMware and containers are together so it’s interesting the way the disruptor can get disrupted and then it all.. whatever. But anyways. The future - where are we going?

A. Mann:

I think Scott and Greg had some really good points in terms of business value. I continue to see my customers asking to connect and correlate the business value impact with the activity that’s happening. So, it becomes less and less about solo performance, network performance and any kind of performance or response times. It becomes more about how much revenue did this new program or application generate. How many user sign ups did I get from this new change that I put in last week? These sorts of things.

The other area, and I will use the word container, again, I think that cloud becomes ever more fragmented and disintegrated as we start to shrink down the unit of work, and I think to Greg’s point, we start to choose multiple clouds for different purposes and put together that whole stack of cloud. I think it becomes more disintegrated. More fragmented. More functional. More repeatable because you don’t have to re-write code. You just get to reuse it from this cloud or that cloud or wherever it comes from. That will create a lot of challenges in terms of correlating that business impact because you’re going to have one ‘service’ using fifty different clouds, with just a micro service from each of them, so it’s going to be challenging but I think it is going to generate much better results. We’re going to be able to get to those business outcomes faster, easier and probably cheaper, although that might be another myth we didn’t touch. Part of that is the micro services and containers, James. So, we’ll see how that all works out.

J. Maguire:

What do you mean about shrinking down the unit of work? Just so I get you on that one.

A. Mann:

So, this is the whole concept of creating micro services and an API driven development where you create a single small unit of work, it might be a customer look up or a database update or it might be a call to all partners service where it’s a very specific and small unit of work and you connect these all together in a complex system and you may not even have them all in one place. You probably won’t have them all in one cloud. These individual, tiny micro services will be distributed to the four corners of the earth.

J. Maguire:

Good, we got micro services in there as well. I feel complete now.

G. Knieriemen:

I think I filled out my cloud bingo card here. I’m good.

J. Maguire:

Micro services for 39! Yes!

Alright gentlemen, thank you. I’ll send you the link. We’ll get a transcript of this and also an audio version, believe it or not. It’ll be words and podcasts and video. We’ll send it out all over the word and we can tweet about it. Thank you very much. It was great.

All:

Thank you, James.


Latest Videos



Upcoming Hangouts
See All Videos