Anyone familiar with network diagrams knows that the cloud symbol is used to refer to the things outside of the control of the network owner. In the old days it meant our network connects to the Internet here, or connects to the telephone network here.
Wait, that is still what it means! By this definition we have had cloud computing since the 50s. What is the big deal about all of this “Cloud Computing” then?
True to the definition, we are shifting more computing from inside our networks to the part of the diagram depicted by the cloud – the part out of our control.
Web email (gMail, Hotmail…) was the first mainstream application of this, but network administrators know that the migration to the cloud started well before that with security services, enhanced phone services, distributed computing grids. And everyone else is watching as we are now getting cool cloud apps like Dropbox, Evernote, Google Docs, and Office 365.
So are we just back to timesharing the VAX? Well, no.
Yes MS Azure, AWS, Google App Engine, OpenStack, and the dozens of other offerings do look a lot like mainframe timesharing with one big exception – the new cloud services talk to things inside your network, and talk to each other.
All of this talking is done with Application Programming Interfaces (“APIs”). These are instruction sets that enable people or computers to interact with systems, without being in the system.
We will all be hearing a lot about APIs in the weeks ahead because how they are used and who owns them is the center of the currently front page lawsuit between Google and Oracle.