You could build an infrastructure across multiple clouds. When AWS went down your software would automatically fire up more instances on GAE or Rackspace Cloud to replace it. A complex layer of software that knitted everything together would result in a large increase in reliability.
It could be 'serverside' rather than 'clientside'.
As in, sure you can connect the AIM and MSN chat networks, but you yourself have to run a server to connect the two.
Currently one also has to have migration functionality written by themselves. Instead, amazon could provide it, just like AIM could provide an MSN gateway themselves.
Presumably, Cerf is talking about being able to use different cloud service providers as "parts" in assembling a larger whole.
Cerf's comments suggest an opportunity for someone to assemble a "meta cloud" cloud service which provides this type of interconnection and interoperability.
Such a service would have a limited lifespan, as eventually one of the major cloud operators would acquire it, or build the same functionality.
I'm not so sure. Even instant messaging has not done much interop. Users on AOL and MSN cannot IM each other. This is a very popular, end-user scenario with network effects. Wouldn't that be the "same pressures" as email?
Why would "clouds", which are mostly a concern of IT and software implementation folk, get pressure to interop like email?
He's obviously smarter and has better perspective than me, so what does he mean? Perhaps API compatibility because everyone will adopt AWS and OpenStack APIs?
So what is the cloud protocol? I'm not sure this guy even understands what a cloud is. Email messages all look the same (headers, then message) and they were mostly transferred on SMTP. How would anyone possibly interconnect AWS and Rackspace? Dropbox and Let's Crate? etc.
I think his point was mostly that pressure would eventually cause different cloud services to allow intermingling of functionality. For instance, editing a document on a Google storage service though Microsoft's webapp, or vise versa, without having to manually move stuff over.
I'm skeptical though. Companies always find some excuse to prevent the customer from using other companies products.
That is why I mentioned common format (headers + message) and 'mostly' SMTP.
But for early email, allowing users to email someone at a different computer wasn't an issue of profit.
'Cloud' companies don't want you to use another company's products. Google doesn't want your docs stored on Dropbox. That issue didn't exist with email, and if you look at something like Facebook messages, you can't send a message from a Facebook account to a Myspace (or whatever else) account.
That, more than anything else, is going to inhibit 'cloud' growth.
