Monday, April 09, 2007

The Impact of "IT on demand"

Amazon are known for selling books, CD and recently also grocery. Amazon have millions of customers, and in order to attend all those customers, they have built up a scalable infrastructure with a huge amount of storage and computation capacity. Based on their experience they recently have started to rent out parts of their IT infrastructure as Web Services. Currently 10 web services are available. The three most important ones are:
  • Simple storage. Amazon S3 is storage for the Internet. It is designed to make web-scale computing easier for developers.
  • Mechanical Turk. Amazon Mechanical Turk provides a web services API for computers to integrate “Artificial” Artificial Intelligence directly into their processing by making requests of humans.
  • Elastic computing clouds. Amazon Elastic Compute Cloud (Amazon EC2) is a web service that provides resizable compute capacity in the cloud. It is designed to make web-scale computing easier for developers.

This trend is important for several reasons. For future providers of such web services, it creates new business models for companies who have invested significant budget in IT infrastructure. What was viewed as a necessary evil is turned into a new source of revenue.

The benefit for potential buyers/users of such web services, is that it frees them from having to think about and invest in IT and computing infrastructure. The barrier to start up an ecommerce business becomes lower, by making much more flexible and controllable the associated IT costs.

2 comments:

Stefano Bertolo said...

Hi Richard,
so, do you think that developers of Semantic Web technologies are thinking about this opportunity as much as they could/should? If not (as I believe to be the case) why not?

Richard Benjamins said...

I don't think that developers of ST think about this a lot. Semantic technologies pay off when there are many services out there, when you get lost in a sea of services. There are still not so many services out there, so it is still feasible to manage this at the "non-semantic", manual, keyword level. I don't think this will change in the coming year, but yes in the long run. There is a "Web2.0" opportunity here to involve -from the start- a community to semantically annotate services. This requires tools, which is i think something ST developers should be interested in now.