Control of Web Based Applications

Having implemented a regime of software testing for all PC based applications I now find that I am faced with another hole - web based applications.

One of the key concerns with traditional software testing is the impact on the network both localand wide-area. Whenever a new version is received we submit it for testing before we deploy.

In the case of web based applications a user is free to go to a URL, pull the appropriate applets (if required) and run the software. When a new version is released the upgrade work is largely hidden from view.

Does anybody have an approach which addresses the testing and change control of web-based applications?