This is my seventh AlwaysOn Stanford Summit. I thought it looked a little smaller, but Tony Perkins told me at lunch that they had 850 people on the register, and that they had sold out of CEO Showcase spots. I can see that, because the showcases seem packed. There are many startups in the world.
Here’s my friend Rafe Needleman, CNET Editor whose writing I’ve been following for ten years through various incarnations as a tech journalist, moderating the next panel on Rainmakers in Cloud Computing. The participants are Kenn Comee, CEO of Cast Iron Systems, Monica Lam, Co-founder of MokaFive, Juan Carlos Soto, VP, Sun Microsystems, Chip Hazard, General Partrner, Flybridge Capital Partners, Michael Stein, CEO of Darkstrand, Michael Peachey, Chief Architect & Dir. Cloud Computing, TIBCO.
Over lunch, several people talked to me about cloud computing. Most didn’t know what it meant. Some language: “outsourcing,” “the mainframe is back and it’s sexy,” “hosted applications,” “pay as you go,” “software subscriptions.”
There has been a technological shift in the last five years that has changed “utility computing” to cloud computing and in ten years the software business and the cloud will be synonymous, some of these panelists say. Desktops have been virtualized, and almost all desktop apps can be put up in the cloud.
We can now “pay by the drink.” It’s the mainframe all over again, shared computing power with a “thin client” that looks like a browser. But will enterprises want to shift all their digital data to the cloud? Not quickly, and not to the public cloud. But private clouds will run in ways that look like the public cloud.
What are the big opportunities for making money on cloud computing? Infrastructure is actually one of them, and new kinds of apps enabled by the cloud will help the business user who doesn’t know how to deal with IT.
What’s the next new Salesforce.com? The landscape of horizontal software-as-a-service is pretty built out. But there will be needs for new infrastructure that will help IT departments run apps in a private cloud. The side benefit is that the cloud is an enabler of application development, as Dave Winer can already tell us. Disaster recovery and backup will cross the boundaries from the enterprise to the cloud.
Issues? Quality.Rafe says all these web apps can be built for next to nothing, because they’re built on Amazon Web Services and running in the cloud. But many of them aren’t making any money. So there will have to be a rationalization to optimize how clouds work (Ken Comee). Then many things running in the cloud will die, but the best ideas will percolate up.
Infrastructure. Michael says that GE is trying to deploy a SmartGrid for utilities, but it can’t figure out how to run that in the cloud because the backbone of utility companies has data rates that will bog down the traditional internet. The cloud will get too crowded, and there won’t be “room” to put everything on the internets. Moving a terabyte of data on Amazon Web Services is no fun. So Sun has tried to bring many of the concepts of running your own datacenter into how you would run your own cloud so you are not on the public cloud.
So the networking side of cloud computing has a lot of opportunity.
Security. How far can people in the enterprise go with an app before the iron hand of IT comes down on them? That depends on the application. People will then develop private clouds for niches like HIPAA compliance. There will be many clouds ooptimized for different market segments. That would be happening today but for interoperability issues — which will be solved. Security is the number one reason enterprise apps are afraid to go to the cloud. 75% of enterprise people wouldn’t move their apps because of it.
To what extent do security issues like the break-in at Twitter undermine the premises of cloud computing? Most of the apps Twitter used were in the Cloud, like Google docs, and the hacker came right in and got all their data. But the panelists agreed that security policies at Twitter have been lax, and that even if the docs had been on internal servers, a hacker could have gotten them given Twitter’s security policies.
There are no HIPAA compliant clouds out there today, although there are SAAS-based EHRs. As a user, I’m not worried about someone taking my medical information, but I am very worried about Google mining my data, or my insurance company. This is a reason for a private cloud and locked down security.
Return on Investment. From a VC standpoint, it takes a long time to run a subscription business past the Valley of Death. Time to revenue for a cloud business is very long, because revenue comes in a stream, not a lump. Driving sales has to be very swift and very focused.