What Was Wrong with ADO?
Microsoft extended ADO to ADO.NET to address the massive scalability of the Internet. The potential is that millions of users may eventually be using an Internet application simultaneously. ADO was designed for hundreds (perhaps even thousands) of users, but the connected nature of ADO ultimately resulted in an upper limit on the number of simultaneous connections.
ADO.NET relies on a disconnected model. If the data is disconnected, an infinite number of clients can theoretically be running an application. Additionally, the best way to move data around networks is still as small text files. Microsoft took advantage of XML as the internal storage mechanism for ADO.NET. Due to XML's self-describing nature and in lieu of it being text, it is very easy to transport self-describing data using standard Internet protocols such as HTTP.
ADO.NET means database transactions across the Internet using XML. It means huge scalability because browser clients don't have to hang onto connections to a single database, which promotes database Internet applications running on Web farms. And because ADO.NET is built on XML, ADO.NET integrates seamlessly with XML Web Services.