Recently we had the real pleasure to sit down for an extended conversation with David Linthicum, CTO and Founder of Blue Mountain Labs and one of GatheringClouds.com’s favorite cloud thought leaders. David writes for a number of online publications, including GigaOM, TechTarget, and Infoworld, among others. Our discussion covered a range of topics including enterprise cloud adoption, cloud as commodity, what companies are “doing it right,” and more.
Gathering Clouds: The notion of do-it-yourself (DIY) versus the as-a-service model, in the long term, are you saying that, as-a-service, is going to win? Or is there still a DIY component that is necessary to function properly for a business?
David Linthicum: I think it’s always going to be a certain amount of DIY that’s going to be part of the deal. And I think the larger the organization, probably the more DIY is going to occur. But I think, eventually, the utilitarian nature of the cloud and the ability to do things with less is going to win out. And DIY is going to start moving in that direction.
We’ve seen shifts like this occur with the move into the internet, the way to gather content. We’re kind of redoing that in this decade. It’s just the ability to see our own data. And instead of seeing our own data and our systems that we maintain and host, we realize that other people can provide those services.
I remember that I spent all of my time and energy building systems so people could see a freaking stock quote on their computer screen. Well, when the internet came along, we could do it on Yahoo. And so that became an antiquated way of doing that. I think cloud computing is going to follow the same evolutionary gap.
GC: You recently published a piece that was a nice companion piece to a Barb Darrow article on AWS usage at the enterprise level. What’s the way in for the cloud at the enterprise level? What’s the evolution of cloud usage at that scale?
DL: A lot of the business cases that I see are utterly just science fiction. But understanding realistically what you can do in a certain amount of time and what the low-hanging fruits are, you can get to start migrating in that direction.
You have got to create a cultural change. I called this problem out, in a recent blog, as the so-called folded arm gang who just hate clouds. They’ll fold their arms and they’ll tell you how much they love cloud computing. They think it’s the future. But they don’t want to have anything to do with it because they think it’s going to change their lives. It’s getting those guys out of the organization who are not able to evolve their thinking. But I think a lot of these guys won’t change no matter what. They need to be gone. And by the way, that’s happening. People are now gone, based on the fact that they’re not thinking ahead about evolving the organization and they think they’re going to keep on doing the same thing and get different results sometimes. That’s the definition of insanity.
Once you create the vision and you change the culture, you do some proof of concepts (POC) and get into it, in a baby step kind of a way. I think that it’s okay to spend a year monkeying around with the technology before you start deploying large scale systems. And that way, you gather the data points. You do your own proof points. You also take a lot of the risk out of it.
And then, it’s a matter of putting together a plan that is going to operate within your budgetary constraints, fiscal pressures and your political pressures within the organization. People will hate this. They write a three-to-five-year plan as to where you need to be. And I think there’s no way in hell that that three-to-five-year plan isn’t going to change by the time you’re done with it. But I think it’s a good idea that everybody understands where we’re heading.
One of the things that drives me nuts is when someone says you need to move your data to a public cloud service when they have no understanding of what information they’re storing and how they’re using that information. Well look at the information. Identify it. Catalogue it. Make sure you know what it is. Make sure you figure out how to migrate it into the public cloud effectively. There has to be a plan of how you’re going to integrate that information back to the existing internal systems.
How are you going to do that? What are the 20 steps to make that happen from the inception, getting the team together, to the testing and acceptance of it, even the operational aspects of it? The devil’s in the details. Enterprises seem to be missing the fact that they end up seeking immediate gratification and want cloud computing to save them from all their problems. And I don’t think it can save them from all problems. What I think it can do is eventually get them in a much more effective, efficient and strategically effective state. But it’s going to take a lot of planning and work to get there.
And then, it’s important to publish your success. I always say pick the low-hanging fruit, the ones with the least risk, the ones you’re going to be successful at because nothing drives more success in the cloud like success in the cloud. So you pick the projects you think are going to be successful within a year. And then you make sure those victories are communicated back to the stakeholders and the executive leader. Cloud is truly a game changer and we need to start moving in that direction.
It’s also important to not pick the wrong battles. People love taking their mission critical systems and throwing it out in the cloud in some kind of a dopey way without any kind of redesign or redevelopment of app architecture. They just change the platform. And you’re not going to get success from that. You have to re-architect, redo, re-understand your systems or start moving it in that direction. Or else, you’re just going to get in trouble. So, take the baby steps first, the low-hanging fruit.
One big thing to do is make sure the culture changes. In many instances, I’ll go in and talk to a client. They’re willing to pay my fee for a couple of days to come out and talk to their leadership. And the leadership has no interest in doing it. And there’s nothing I can do to change their minds. I can show them a million presentations and write a million papers. They may even like me as a person and respect me as a thought leader in the cloud computing space, but they’re not going to change their minds. And like I say to them, unless you guys are willing to open your minds to new options, I can’t help you. If you’re going to fight me every step of the way, in the passive aggressive way, it’s just not going to work because I’ve been there before. I’ve tried it in the early days and it just doesn’t compute. So, there’s got to be a willingness to change. People need to understand that they need to reinvent the way that they look at technology, be more innovative and creative in how they look at technology.
And that’s something enterprises typically don’t do because they’re not compensated to do that. They’re compensated to maintain the status quo, to not rock the boat and keep the business happy, only increasing the budget by 10 % every year. That’s kind of the way they’re wired. And cloud computing kind of throws a monkey wrench into those sorts of plans.
GC: I want to propose a concept to you to get your response: One of the benefits that cloud presents is the commoditization of IT infrastructure. But internally, here, we view cloud as a combination of the service component of that technology, as well as the commodity technology itself. And you can’t really commoditize the people/service aspect of it, except for those very automated, repeatable tasks. But what parts of cloud do you view as commoditized and what don’t you see that way? And is that sort of a moving target? Is that going to evolve?
DL: It changes rather quickly. storage-as-a-service, I would say, is commoditizing fast. Where, two years ago, it was just a complete miracle, as far as I was concerned. We were able to store stuff out on the Amazon Simple Storage Service (S3) cloud and retrieve it, just like it was down the hall. It was on an internet-based system. There’s going to be a lot of cloud providers out there able to provide similar services. And it’s going to be very difficult to differentiate your ability to do object and block storage versus some Chinese company that can do it for a fraction of the cost, in some cases, a penny a GB. I’m hearing those sorts of things being bantered about now.
People are going to use the lowest cost, most compliant, reliable service, ultimately. I don’t think it’s been commoditized yet, but definitely they’re racing to the bottom. Google just dropped their prices. Amazon keeps dropping their prices. And there are a lot of storage and service companies that are out there trying to race to the bottom as well. So, I think that’s something that we’re going to see commoditized pretty quickly.
Platform-as-a-service (PaaS) and the higher level technologies in the stack will never be as commoditized. I think they’re just so different. The value propositions are so different from product to product. It will change over time. And the fact that there’s a ton more stuff going on to create in that area, especially, around the identity and management space that I don’t see it being commoditized for another 10 or 15 years, believe it or not. I just think that everybody’s going to be duking it out for a while before initial winners come to the forefront or there’s some consolidation that occurs which kind of takes a lot of the players out.
I think the players in that part of the cloud space are going to be new players. It’s not going to be the Oracles and the IBMs. It’s going to be the Amazons. Even the smaller guys now, as they start to grow up will begin to take the lead. And many of them are going to go public and become independent companies. It’s an explosive market that is creating a very interesting space. That’s why I love this space. It’s always going to be changing. I think the players are going to be changing. And I think the biggest losers out of all of that are going to be the IBMs and the Oracles and the HPs to some degree. Unless they’re willing to step up and buy everybody, which is going to be very expensive, they’re not going to have too much of a dog in the fight. Their offerings are what are going to commoditize within the enterprise.
GC: The very easy, high-level view of cloud paradigms that we see a lot of are SaaS, PaaS and IaaS. Are these always going to be the main buckets or is there going to be sort of an evolution within them? Is one of these going to trump the others, in terms of how people consume cloud most?
DL: Yeah, I think they have to evolve, just because they’re too simplistic. I mean, those labels were simplistically created to define cloud computing. They got those labels from crowd sourcing within emerging patterns at the time.
In the PaaS realm, there’s design-as-a-service. There’s the ability to do database-as-a-service stuff. And you can break it out further. Software application-as-a-service, utilities-as-a-service, connected and disconnected modes, PaaS design process, testing, design-as-a-service, the ability to design apps and deploy them on these systems. In essence, kind of a “next generation” cloud case. I mean, database as a service, management-as-a-service, middleware-as-a-service, integration-as-a-service – and I could probably name ten more. They are, in essence, kind of emerging categories under themselves.
And so, I think cloud computing broadly is still going to use buzzwords. I’ll call them buzzwords, SaaS, PaaS and IaaS. But I think it’s bigger than that. I think that they have to evolve. They’re too limiting for what cloud computing is.
So, in a couple of years, instead of having a PaaS conference or a cloud conference, we are going to have a design-as-a-service or testing-as-a service conference. And it’s going to be big. And database-as-a-service for sure. And that’s going to be big. And I think we have to break them out into these sub-disciplines, these subcategories of cloud computing to make better sense of the market.
GC: For the cloud technology providers, and the cloud service providers, how do you see those two parallel spaces developing? How are those two paradigms, private versus the big public clouds going to coalesce? How are they going to compete with each other? How is that space going to change?
DL: Yeah, that’s another good question. Private cloud providers are going to find that they eventually become gateways into public clouds. They will become the way in which you can point resources in public cloud providers. In essence, private cloud providers will become hybrid cloud providers. So, working and playing well with Amazon, like for example, Eucalyptus is a smart move.
I see public clouds as certainly a legitimate architectural option that people can leverage, if there’s a need to do it. But I think that space is just going to combine with the public cloud providers. And they’re going to be either joined at the hip with a standard such as OpenStack, CloudStack or others. I don’t think there’s going to be a major private cloud provider that everybody’s going to go to. The stuff that IBM and Oracle are doing is crazy. I mean, Oracle with their cloud solution being sold to the enterprise for a million dollars – that’s nuts. They don’t get it. They don’t understand how this stuff works.
And that’s how that world is going to evolve because I think people are going to kind of come to their senses and go, “Why am I building on a cloud that I have to maintain? Shouldn’t I be building somebody else’s infrastructure?”
The value is basically to have a cloud infrastructure that mimics the public cloud. This will enable businesses to leverage a public cloud as a resource for a private cloud that will be able to share new processes between the different cloud platforms. A few companies are trying to accomplish that these days. Well, that’s something because that means I can take processes, move them on-premise or I can move them off-premise. I can move my data off or on-premise. And I can balance as to where my storage, my computer resources and my process resources exist. The public cloud providers are going to be more than willing to go along with those business requirements to make that happen. And certainly, that’s the vision for a lot of them these days. And I think they’re going to start moving in that direction. We’re just at a point where people are just standing on private clouds. And sometimes, if they’re actual private clouds, you can do provisioning out of them.
Public clouds just need to become the path of least resistance. I’ve always felt dealing with the public cloud providers was like dealing with porcupines. They were something you respect. And you may want to leverage one every once in a while. But the minute that you have a problem and you call or email somebody, you just never get the right answers. And it is very difficult, unless you have a good consultant with you to get your infrastructure correctly set up and maintained. Public cloud providers need to change, and I think they are beginning to.
Rackspace is doing a much better job on customer service than Amazon is, and that’s becoming their main differentiator. Amazon is beginning to get their customer service right. They’re adding different customer consulting organizations, and developing the necessary solutions around their cloud versus just providing good cloud services . And so they’re going to evolve in the fact that they’re becoming kind of a mass market where they’re going to have different levels in which they sell. They’re going to have the individual user, the retail use, and the larger business user. They’re going to have a small business user, which is 80 % of their business these days. And they’re going to sell differently to that constituency.
And then, they’re going to have the enterprise users. The enterprise users are not used to linking into Amazon and giving their credit card. They’re used to calling up and negotiating some kind of a range of SLAs and all that kind of stuff. And they’re going to sell $50 million with Amazon services in a chunk. For enterprise clients, that’s the way they’re going to need to sell it going forward. And that’s going to be the prevailing model. That’s how they’re going to get into enterprises.
And then, you’re going to see, perhaps, a few of the major cloud providers consolidate. And I think the government’s going to be interested in that. Cloud providers have become kind of the new phone companies, in a way. Therefore, their business strategy is going to be managed a little differently and eventually, we’re going to have problems with too much power being put into one particular provider. And maybe that’s Amazon or somebody else. No single company can dominate the market with what they have right now. But certainly, the trajectory is to achieve that level of dominance. Amazon, in five or six years, could be a 100 billion dollar company. Suddenly, everybody will need to use them because they have everything and anything you would need.
And then, everybody will be beholden to them. Then they will raise their rates or lower their rates and do nasty things to their clients very much like IBM and Microsoft did at certain times in the past.. That’s going to be the catch-22 in the cloud. I think it behooves us, as users, to make sure that doesn’t happen. We need to make sure that not all the power and resources go under one particular company. But the market does that. I have no control over it. Those who buy or reject Amazon’s cloud services have no control over it.
That’s why we think the private cloud providers, like I said, are going to be gateways, for a lack of a better term.
GC: Both on the technology development side and also, the service provider side, who do you see as doing it right and doing it wrong? And also, as sort of a corollary to that, what is missing in the cloud space for you?
DL: Technology providers, like HP and SAP are prime examples of companies that really don’t get “cloud” and aren’t moving correctly in the right direction. I’d add Oracle to that as well, only to pick on three. So, they need to take a very different approach than what they are taking right now in order to take advantage of cloud computing. And I’m not sure they’re willing to move in that direction.
The upheaval, the cultural changes, and the fact they have a lot of sales guys that are running around with white belts trying to sell enterprise software and selling cloud services is a very different model from an enterprise model. So, they’re going to undergo some very uncomfortable times, I think, before they change.
In the public cloud world, my number one for who’s doing it right would be Salesforce.com. We don’t talk about them a lot because they’re a SaaS company. But they kind of created the cloud space and proved it out before everybody else. I think they’re moving their technology in the right direction; the purchase of Heroku was the right thing to do. And they’re able to move in different directions now. I think that they’re getting more customer-focused right now. They understand that there’s a limit to what they can do. They can’t grow their market forever. And they have to diversify, become more creative and innovative.
Amazon, just from the point of view of the technology, is doing it right. Just walk the halls of that place and you can feel the intelligence.. They’re committed and dedicated almost to a cultish degree. They know what they need to do to capture the market. And they just focus on building the best thing in the world, in terms of a public cloud computing service. You can’t argue with that.
But they need to get their customer service right. They also need to diversify their sales models. But I think, eventually, they’ll figure it out. They’re not making many mistakes. Let’s put it that way. And then, I think that Rackspace is doing it right. I wasn’t a real big proponent of theirs and when the OpenStack stuff came around, it was a little gimmicky. However, I think that they found a market and they were able to diversify themselves from the existing public cloud computing space by, in essence, working to create an ecosystem, which is the OpenStack platform. They, in essence, drove it, and now they’re planning a new market because people are seeing opportunities in the public cloud space, in the private cloud space, or on the OpenStack platform to build a cloud without starting from scratch. I think that’s a smart way to do it. And I’m very impressed with Rackspace in not only coming up with that strategy but executing on the strategy.
GC: What’s missing in the space for you?
DL: What is missing right now are management solutions that are scalable and really keep up with cloud. We’re very good at building these things. But as far as maintaining reliability and maintaining the service levels, there are not a lot of management and governance solutions within a cloud computing space. They’re emerging and there’s a few potential companies out there that are starting to make an impact. But when you build clouds for big companies, they’re going to need governance solutions in there with compliance and service monitoring. And there aren’t a lot of good ones out there that do it. There’s RightScale, who operates at a much more primitive level. But for someone to do traditional service governance, data governance, things like that, it’s a huge hole in the space.
Security solutions would be another area that’s lacking. Sophisticated identity management-based systems that have native hooks into the existing public cloud providers don’t seem to be around. Existing public cloud providers don’t put a lot of emphasis on security. They’re looking for somebody else to solve it. But I think that’s the big hole that somebody can step in to solve the space. Most of the companies working in this space haven’t penetrated the market very much. They’re still at the early stages. I think someone needs to come along and ultimately, pick that up.
The final thing would be performance management. It almost goes back to the governance area but the ability to basically manage performance and move meta-processes in between public and private clouds, in and between public clouds. In essence, brokering the least cost, the most performance, the shortness of hops from Point A to Point B, the ability to understand this from an infrastructure point of view and dynamically make these changes is something people are screaming for, in order to get their infrastructure up and running.
Right now with cloud computing, if they’re in production and they’re dealing with multiple cloud providers or even a single public provider, you have one guy kind of looking at a screen at a couple of consoles trying to move things around and relocate processes as he needs to support the performance expectations that his business has. It’s a very difficult thing to do. I think if someone steps up and fills that space, you’re going to have something.
By Jake Gardner