Many real-world applications need to interact with a database. The .NET Framework provides a rich set of objects to manage database interaction; these classes are collectively referred to as ADO.NET.
ADO.NET looks very similar to ADO, its predecessor. The key difference is that ADO.NET is a disconnected data architecture. In a disconnected architecture, data is retrieved from a database and cached on your local machine. You manipulate the data on your local computer and connect to the database only when you wish to alter records or acquire new data.
There are significant advantages to disconnecting your data architecture from your database. The biggest advantage is that you avoid many of the problems associated with connected data objects that do not scale very well. Database connections are resource-intensive, and it is difficult to have thousands (or hundreds of thousands) of simultaneous continuous connections. A disconnected architecture is resource-frugal.
ADO.NET connects to the database to retrieve data, and connects again to update data when you've made changes. Most applications spend most of their time simply reading through data and displaying it; ADO.NET provides a disconnected subset of the data for your use, while reading and displaying.
Disconnected data objects work in a mode similar to that of the Web. All web sessions are disconnected, and state is not preserved between web page requests. A disconnected data architecture makes for a cleaner marriage with the Web.