I remember attending the Debut of Dot Net back in the late ‘90s, attending a presentation in Indianapolis, IN, USA. They made a statement at the time I have not forgotten. I didn’t quite believe, or understand it. But it stuck with me. “Microsoft believes the future for software is going to be software as a service, not software as a product. Clearly, they have continued with that strategy.
Some obvious examples:
Azure: a product only available in the form of Software as a Service. Sure, you can rent Infrastructure as a Service, hosted in the Azure infrastructure. However, to truly take advantage of the powerful services available in Azure, many are only available as SAAS, such as Machine Learning.
Office 365: For those companies where Google Docs may seem too open, or not powerful enough, Office 365 provides a great alternative.
There are many other Microsoft products as well as other Microsoft partners, such as Oracle, with full featured stacks that can be spun up on the Azure platform.
There’s lots of benefits. No need to list then again here. From a software developer perspective, this is a great way to mitigate unlicensed software installation. You get paid for the use of your efforts.
The model that seems to be working best, in my opinion, is the one where software is nearly given away. A useful base product is available, with big enough limitations to allow a true evaluation of a product. Revenue is achieved when the product is used in a commercial environment, or even sometimes when it exceeds a number of installations. For example, Team City can be installed with a limited number of servers, and is fully functional. Once you exceed the limit, you purchase a license.
Other software may be free. What you purchase is consulting services, or management software. Cassandra is one example of this model. The engine is essentially free. If you want to put in the time to build your own management tools, and become and expert with the product, the software cost may be smaller. Your out of pocket expense may be more than if you purchased consulting help up front.
I am not sure who first came up with the perception that software as a service would become a norm. I am sure that the prediction appears to be quite accurate.
The next prediction I’m waiting for came in the early ‘90s. They were saying back then that you would not be able to tell where your program was running. Any CPU with available resource on the network could be supporting your work. we have the predecessor to that model in application fabrics such as Appistry, Azure, or even Amazon EC2. Who knows, as network bandwidth increases, this could come true in more than specialized implementations. We could have real time distributed computing. Imagine the power of millions of idle cpus. The only thing holding us back is security and bandwidth.
That’s enough dreaming for today. Why not share your favorite prediction in or comments today?
Cheers,
Ben