Robert Scoble leads a debate about whether today's cloud environment is a revolution or a bubble--and notes a gulf between marketing and engineering teams on cloud.
Is cloud computing a revolution or merely the froth of a new technology bubble?
In a panel discussion at Web 2.0 Expo, a UBM TechWeb/O'Reilly Media event in New York, it was almost a foregone conclusion that the verdict would favor revolution--and yet the panelists acknowledged that it may be a bubble, too.
Robert Scoble, the technology blogger and super-connector who works for Rackspace Hosting, moderated the discussion, with Rackspace CTO John Engates as one of the panelists. Other participants included Allan Leinwand, CTO for infrastructure at Zynga, Brad Hargreaves, a founding partner of General Assembly (a New York City technology incubator and "educational campus"), and Ari Levy, who covers tech startups and venture capital for Bloomberg News.
"There is no good revolution without a bubble," Levy said. Even though the cloud is creating real business opportunities, the amount of money flowing into cloud startups is probably excessive, he said. "There may be money lost--I think there likely will be money lost along the way."
"I do see companies raising money using the keyword 'cloud,' which gives them a valuation increase of 30 to 40%," Hargreaves said. That's partly a phenomenon of hype, he said, even though "the reasoning about why cloud brings higher valuation is ultimately based in reality," he said.
Scoble noted that some of the confusion about the cloud is that marketers talk about the cloud as a philosophical concept about how to deliver services over the Internet on a subscription basis, while engineers have a different concept of the technical underpinnings that make the cloud possible.
"A lot of people jump to the conclusion that the cloud is just virtualization," Rackspace's Engates said. That leaves out the layers of automated provisioning and orchestration of services needed to build a flexible, scalable service, he said.
Zynga has made aggressive use of cloud infrastructure, which Leinwand said is valuable because it allows him to scale up and scale down smoothly. That is important because a new game can attract millions of users "in what seems like seconds," he said.
In addition, cloud infrastructure provides some geographic independence as opposed to sweating the details of how many servers to locate in a data center cage in Virginia. Cloud services also need to provide reliability so when users go to play a Zynga game "that service has got to be there right away," he said.
Still, Leinwand cautioned against as thinking of cloud infrastructure as being magical.
"Clouds aren't water vapor. Ultimately, there is silicon and hardware behind it," Leinwand said. "I think that's kind of the trap of cloud computing. You do need to understand a little bit about what's under the hood," particularly if you're staking your business on it, he said.
Those infrastructure details still matter, even though a good cloud service turns them into "something that's consumed easily and scaled easily," Leinwand said.
Zynga runs parts of its infrastructure on the Amazon Web Services public cloud, but it also maintains its own data centers that are architected according to crowd principles, Leinwand said. "Having both is an important part of our approach to infrastructure."
Engates said one compromise model he sees taking root in more established organizations is "owning the base and renting the spikes," where the organization maintains some core essential services on its own infrastructure but uses the cloud to manage the overflow and the experimental new ventures.
Startup companies today tend to start by looking for software as a service (SaaS) to meet their operational needs, Engates said, and they only go further if they need more customization than SaaS can provide. "Platform as a service (PaaS) is sort of a fallback strategy," he said, and companies that don't get their needs met through PaaS can go a level lower to infrastructure as a service (IaaS) or, if necessary, dedicated hosting of some services.
One of the ways Rackspace has tried to open up new possibilities for the cloud is by promoting an open source architecture called OpenStack.
"One of the things we felt was holding people back from a move to the cloud was that there weren't any standards," Engates said. "We wanted to give people the cloud on their own terms." RackSpace worked with NASA to define a suite of open source standards for cloud style computing, which is now freely available for other Web hosts to implement and for enterprises to implement internally.
Just Wednesday, eBay announced it is adopting OpenStack as a component of the X.commerce platform it is creating as an integration point for services from eBay, PayPal, and Magento e-commerce products.
Cloud architectures also require different tools for managing and measuring service performance, since traditional enterprise network management products tend not to work well in the cloud, Leinwand said. Companies like RightScale, New Relic, and Keynote Systems are filling in the gaps, he said.
"I wouldn't call these new ideas, but there's a new format and a new deployment methodology that they support," Leinwand said.
"That's what makes it possible for new franchises to be born," Levy said, because the existing tools and incumbent vendors aren't keeping up with the new requirements. Bubble or not, a lot of real opportunities are bubbling up, he said.
Not every application is ready for the cloud, but two case studies featured in the new, all-digital issue of InformationWeek Healthcare offer some insights into what does work. Also in this issue: Keeping patient data secure isn't all that hard. But proposed new regulations could make it a lot harder. Download it now. (Free with registration.)
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.