The panel, which consisted of six industry leaders whose respective companies are immersed in various parts of the cloud, agreed that though cloud computing is hard to define, it's the reality of how it affects us on a daily basis -- rather than its definition -- that is important.
The panel of experts was as follows:
- JA: Jonathan Appavoo, research scientist, IBM T.J. Watson
- AA: Dr. Anthony Arrott, special assistant to the CTO, Trend Micro
- MC: Mark Canepa, president and CEO, Extreme Networks
- VC: Vance Checkletts, COO, Decho, an EMC company (of Mozy backup software fame)
- BL: Barry Lynn, chairman and CEO, 3tera Inc. (of App Logic fame)
- PK: Prasanna Krishnan, associate, Draper Fisher Jurvetson
- KH: Kartik Hosanagar, professor of information and operations management, Wharton (moderator)
Everything from the very definition of cloud computing to the challenges it poses to the business model upheaval it is causing was discussed. Here's what happened.
KH: So what is cloud computing? How would you define it?
JA: A way of delivering value and monetization efficiency.
KH: Larry Ellison once said in an interview, “What the hell is cloud computing?” I would say that cloud computing is "the notion of data and applications and hardware sources being accessed remotely."
AA: In a recent Economist survey, 20 years ago your typical knowledge worker got 80 percent of the info needed to do their job came from inside the company. Today, it's completely flipped. Cloud computing is the technical response to this reality.
JA: Providing value with computational devices. Maybe we can trade off where those devices live…but it doesn’t change the fact that our goal is to provide value to users, independent of where it’s located.
BL: It’s more than remote access and software as a service. I agree with Larry Ellison, I think cloud computing is a lot of hype…it’s amazing to me who puts ‘cloud computing’ on their website. To me, the future of IT is completely extracting services from servers and networks.
PK: I would say cloud computing is getting all the advantages of computing [without one]. Step 0 was individual PCs. Step 1 was personal PCs and users using Flickr and services like that. Step 2 was enterprises moving into that space, like Salesforce.com Step 3 is the real cloud. Cloud computing is all about making it easier for people who create applications to provide them without the headaches of hardware.
VC: Cloud is the worst buzzword I’ve ever heard. Vapor, something you can’t touch. I think it’s a lot of hype, but then again, here I am offering that service. For Mozy, it was “how can we protect users’ data?” Look at Netbooks – we need to make sure we can protect that data. A lot of it will live in the cloud. Everyone needs backup. We have 18 petabytes backed up across a million users. Cloud to me? Pay attention, but don’t pay homage to the cloud.
MC: The journey began with IBM 360 in the 1960s. It was probably a pretty good cloud platform.
AA: We’re not so concerned with the hype or what you call it – for us, what it amounts to is real. The best example for us are the cybercriminals – they’re the ones who are the most effective users of the cloud today, with botnets and tens of thousands of zombie computers. That kind of power. That’s what we’re dealing with.
JA: What are we putting in place to make sure we have a stable system? There’s only so much electricity, silicon, storage. It’s systemic in our society, but not everyone understands. People need to start demanding, what do we need from our technology?
BL: So who’s confused?
KH: Let's circle back. What is driving the trend?
BL: I think cloud computing – extracting applications from hardware and networks – is inevitable. The question is whether it will be fast enough for people like me to make money out of it. Everybody wants IT services, and if you need to go to a systems administrator to configure everything for you, we’re probably 10 years away. It’s almost inevitable – self-serving technology on demand, as you need it (like a self-service telephone/operator). It’s inevitable, otherwise the IT world will come to a halt.
VC: Thirty percent of this audience has 1TB or more at home. That's incredible. Nobody really wants to manage all that data and store it. The cloud is the solution to what you have in your basement. That’s how we make money – people call us after they’ve had a fire or theft.
JA: It’s centralized, so how do we avoid those problems? Allow competition, allow for diversity. We can’t let “clouds” let us lose the fact that all layers – storage, antivirus, etc. – are free for competition, going all the way down to the bottom of the system. Those are questions we are faced with when we try to articulate what we want from the cloud.
VC: It wasn’t that long ago that you used to have to generate your own power if you wanted to run a business. You probably know your provider, but you don’t know the secondary and tertiary sources of that power. The same thing applies to your financial relationships. I think the same thing is going to happen in cloud computing – services behind the scenes. It’s not going to be one cloud – there will be multiple service providers.
PK: The cloud as a utility. As more and more people need it, it makes sense to provide it as a utility.
AA: If you can imagine it, you can implement it, and that’s true in financial services.
JA: Virtualization is the promise to compute, but it is not actually computing. You really start to think of hardware as a commodity.
VC: You just don’t want to admit that hardware is obsolete. To some degree, we’re talking about capacity…at some point, hardware becomes obsolete.
JA: Hardware should become even more of a transparent commodity.
BL: The value is in the software and the value is in the data. There’s still value into knowing things about the infrastructure. All of that [software] stuff needs to be deployable so we have a network.
KH: Let's shift gears. What are some interesting entrepreneurial things that are emerging? What is the industry value chain look like? What are the firms out there contributing to this cloud ecosystem (types of players)? What are some interesting startups emerging and the kinds of things they’re doing?
PK: There are many interesting applications in the cloud. Nesting environments in the cloud, for one. As large enterprises start using the cloud, it won’t be a one-step function – there will be private clouds. There will be opportunities for startups to provide the horizontal layers to bridge public and private clouds. VMware has talked about vCloud. Doing down the stack to virtualization, even desktops, there’s a big opportunity. The cost of managing it is at least as much as the cost of buying it.
VC: Two companies that are cool: Animoto (a service in which you load photos and music and it will create a slideshow or music video). Simple service offering straightforward value. It should be very, very portable. The service they provide should be able to sit on top of any cloud. Businesses like that – almost cloud-neutral – that I think is exciting. Also EarthClassMail (which scans your physical mail and delivers it to you) – that to me is another interesting cloud opportunity. I do think the Netbook phenomenon is kind of an interesting one that’s trying to ride this cloud away.
MC: One area to go into is to help build the cloud. The other – where 90 percent of companies will be – finding a way to do stuff with that cloud. I think the choices are just going to be limited by your imagination and by what users demand out there. What would you like to see as a set of services delivered to you?
BL: They’re not cloud companies – they’re great services that lend themselves to operating in the cloud. Companies that run in clouds – that’s where the new stuff is going. Not every application lends itself to, say, Google App Engine. The hosting industry is going to play a big role. Everybody is kind of defining their cloud platforms now. With bandwidth and processing speed what it is, the network has become the computer.
AA: For us, the cloud isn’t a business opportunity. For us, it’s the answer to a business problem. The bad guys have figured out an inherent weakness in the antivirus industry. We have to move all the information we have about them to every computer and network gateway. If the information isn’t there, we’re not there to protect that computer. They’ve figured out that if they produce enough variants at a high enough rate and jump from URL and URL, the system is overwhelmed by this data.
MC: They just figured out what the common cold figured out a billion years ago.
AA: The answer to them is what ants have figured out: social immunity. In order not to infect the colony, ants spread the immune information physically. For us, even if we could keep up with the bad guys, there’s no way we could deploy that information to the protection points. We’ve moved to the protection point has to query the reputation service in the cloud, so that when you come on something, you check in with the cloud to see if it’s OK.
JA: Let me say it strongly: startups are the killer app for the cloud. You can be the driver of all of this. Computation is so accessible now.
KH: Applications enabled at the cloud level, or cloud applications? Google Docs is an enabled app. Firms that are using the public cloud tend to be smaller startups. Larger firms tend to use private clouds. Is that how it’s going to be? Are there exceptions?
VC: Why would you trust your data to somebody who’s mining it and selling ads against it? Google’s got a fundamental trust problem.
BL: Five percent of my customers are enterprise, but they provide 30 percent of my revenue. If you try to cookie-cut what the enterprise is going to buy, you’re going to lose. I do believe that it will evolve to almost all external cloud. If you’re in the healthcare business, why are you spending billions on IT infrastructure? The answer is because you have no choice. Eventually, the data center operators are the ones going to run computing and the healthcare services are going to be running healthcare and the financial services companies are going to run financial services and so on.
PK: Enterprises are going to move to using the cloud as a utility.
BL: I believe cloud computing -- done right and done well -- greatly enhances private and public security. Right now, the applications are sitting ducks. Cloud computing makes applications moving targets – never on the same machine, or in the same country even.
VC: Trust and security is a huge issue, but it’s not a technology issue. It’s a business issue.
MC: The practicality today is security and privacy. As a computer scientist, I’m personally very excited about the prospect of cloud computing. As a CEO, I think of it in different terms. The ability to great the right networks and infrastructure will be important.
AA: There isn’t anything fundamentally different or new here. It’s just another evolution. Value chains have a way of sorting themselves out. That applies to this one as well. We all trust Mastercard and Visa, and 100 years ago people thought that unimaginable. The winners who figure out from a business prospective where the future is.
JA: Cost can really change the equation. Technology evolves, innovation solves problems. Don’t let all the hype get in the way that you want certain requirements to be met.
Question from the audience: is consolidation on the horizon?
MC: The industry’s been in a period of consolidation for a number of years. There’s still an awful lot of innovation. There will be companies of various sizes that will nibble and take chunks of the market.
BL: The big guys in the server-only business - like the Dells of the world, or HP – are really going to have to think of ways to add value.
Note: this is not a full transcript of the event. It has been edited for brevity.