Cloud is an architecture.
When you architect servers with virtualization, distributed computing and the ability to handle big data, so a single Web site or job can take power of the whole system when needed, that's a cloud.
Salesforce.com (CRM) (whatever you may think of it) is not a cloud. It is Software as a Service (SaaS), something that can result from a cloud architecture but does not require a cloud. The same can be said for Apple's (AAPL) iCloud. It too is SaaS, which could come from a cloud or from a standard enterprise set-up.
What we have learned about the cloud in the last week is that it will, without doubt, be based on open source and open standards. We know this from two news stories:
Open source has become a cloud requirement because clouds are complicated, requiring a lot of cooperation among vendors to develop, and because everyone who has been involved in enterprise computing is rushing to the architecture just as fast as they can. Time is of the essence, in other words, and proprietary advantage has to go out the window.
Over the next few years nearly all Web hosting, all online services and Software as a Service offerings will become cloud based. Even phone companies are moving to the cloud, or at least their suppliers are.
Which tells us something else important.
Cloud is going to quickly become a commodity, not a differentiator, practically a requirement. Customers will differentiate based on the quality of a cloud, or its software, or the people doing the hosting. Enormous price earning multiples based on having cloud, like those of Rackspace (RAX), VMWare (VMW) and Red Hat (RHT) are going to bleed away, replaced by lower and more nuanced valuations based on product and service leadership.
This does not mean that cloud leaders will have PE multiples similar to every other company. It just means they're going to have to earn that leadership in the market, through sales and earnings, rather than by hyping the cloud.