And don’t forget viruses and malware. They’re coming to a cloud near you.
Those are the worst kind
Yeah, I got an image of some CEO at Mary Kay going “Arrrgh! Not again!”, sweating over a pile of lawsuits.
The big “Cloud” commercials right now (“Yay cloud!”) are for Microsoft actually (i.e., your laptop can talk to your desktop at home, yay!). IBM seems to be on a “our employees are super-geniuses” campaign lately.
Excellent.
You forgot “turnkey” and “enterprise.”
Since I’m not there in person, please find someone to punch you in the face.
I could feel my paradigms being shifted as I read this.
In other words, it is a highly abstracted system on both the server and client side, i.e. you neither have a central mainframe or cluster, nor do you have dedicated terminals. The user can access applications or data from a common cross-platform interface, and where the data is stored or computation performed is determined automatically by “the cloud” without bothering the user about the details. This is vastly different than a dedicated client-server architecture like a an IBM mainframe and 3270 terminals in that you can seamlessly port your abstraction layer to a new device and then run clients “natively”; or load your abstraction layer on new “servers” and distribute data or computation to them without having to build a new server-side application from the hardware up.
This is obviously a very idealized marketing version, but as Shagnasty says, the difference between cloud computing and previous efforts at abstraction is that the technology on both sides, as well as the communication infrastructure and protocols that serve as the medium–are now approaching a degree of maturity such that the practical applications of the system are converging with the marketing hype. Unlike machine cognition–which we are nearly as far from today as we were in, say, 1968–highly distributed heterogeneous computing is a practical reality and is becoming more accessible and less expensive by leaps and bounds. The scalar supercomputer has died not because of physical or thermodynamic applications, but because it is cheaper and easier to distribute most types of problems across a network of commodity-grade machines.
“Bingo!”
Stranger