Disruptions to the Software Business by Mark Stone
Previous: Welcome to the Cloud
Imagine during a spike in gasoline prices that your local auto mechanic offered a deal to retro-fit your car's engine so that, for a small fee to your mechanic, your car would go from getting 25 miles to the gallon to getting 60 miles to the gallon. That would be a great deal for you, but a scary prospect for anyone selling cars or gasoline.
This is essentially the position that Intel found itself in 7 or 8 years ago, when virtualization was moving from theory to practical reality. While the idea of making better use of wasted computing power was compelling to businesses, it also held out the prospect of being able to stretch existing computer infrastructure further and thus being able to defer the purchase of new hardware. Not good for hardware companies like Intel.
Its no coincidence then that the retirement of Intel's Pentium line of processors, and the rollout of Intel's current multi-core technology took place just as virtualization was taking off. Multi-core allows many processors to function in parallel off of the same circuit board, and is exactly the architecture you'd like your computer to have in order to optimize it for virtualization.
In efffect, Intel trumped the local IT shop by saying "Yes, your IT pro can get you from 25 MPG to 60 MPG with your existing machine. But with our new machines, you'll get 100 MPG. And go from 0 to 60 in 4 seconds." Multi-core enabled Intel to finesse what otherwise could have been a downturn in their business, and make a smooth transition to the cloud computing era.
But the cloud computing disruptions don't stop at the hardware level. By limiting computer idle time, cloud computing is also limiting software idle time. In other words, getting more productivity comes from getting more efficient use out of existing software, not purchasing more software licenses. So software companies face the same downturn threat from cloud computing that hardware companies have faced.
The response is similar. While existing enterprise software can be run in a virtual environment, the software is not optimized to be run in that manner. A better approach is to replace traditional client software (Microsoft Office, for example) or traditional client-server software (Peoplesoft, for example) with software as a service. This is precisely why companies like Microsoft and Oracle are eager to stake a claim in the cloud computing space. As old business models around software licensing become less relevant, they are looking to ramp up new business models around software leasing.
There will be winners and losers. Not all software is easily amenable to the SaaS model. I personally have had a tough time explaining to folks at Microsoft that the real competition for Word is not programs like Open Office, or even Google Docs, but rather programs like FireFox (the open source web browser), MediaWiki (the software behind Wikipedia) and WordPress (the software behind LiveJournal). When most of our content lives on the Web, we want easy web processing software, not easy word processing software. This is what web browsers, wikis, and blogs provide, and it is where most content is being created and viewed today. Adapting to this change is more complex than just stitching an SaaS interface onto Microsoft Office.
Individual consumers stand to be either big winners or big losers too, and which direction the cloud will take us in is still, so to speak, up in the air. Here's why.
Next: Open Data