Cloud computing—the delivery of processing power over the internet from vast warehouses of shared machines—will further blur the lines between proprietary and open-source software. Most of the firms peddling this model, such as Amazon and Google, use open-source software, since having to pay licensing fees would make the business unprofitable. But their services also rely on code developed in-house, which is not given away free. Microsoft, meanwhile, is building a huge cloud using its own software. If computing becomes a service delivered over the internet, it will hardly matter how the underlying software is developed.
Does this mean that the quest for openness in software is obsolete? On the contrary. If they are not careful, companies and consumers could get locked into a cloud even more tightly than into a piece of software. This is because data residing in the cloud can be hard to move to another service. “If you have a gigabyte somewhere, it develops a certain inertia,” says Mike Olson, the boss of Cloudera, which recently found it could not switch from a poor storage service because there was no way to move the data.
This sort of problem has spawned an open-data movement. In March a group of technology firms led by IBM published an “Open Cloud Manifesto” that has since received the support of more than 150 companies and organisations. It is only a beginning, but perhaps this time around the industry will not have to go through a long proprietary period before rediscovering the virtues of openness.