ADO.NET Architecture

The main design goals of ADO.NET are:

• Customer-driven features that are still backwardly compatible with ADO.NET 1.x
• Improving performance on your data-store calls
• Providing more power for power users
• Taking advantage of SQL Server–specific features

ADO.NET addresses a couple of the most common data-access strategies used for applications today. When classic ADO was developed, many applications could be connected to the data store almost indefinitely. Today, with the explosion of the Internet as the means of data communication, a new data technology is required to make data accessible and updateable in a disconnected architecture.

The first of these common data-access scenarios is one in which a user must locate a collection of data and iterate through this data just a single time. This is a popular scenario for Web pages. When a request for data from a Web page that you have created is received, you can simply fill a table with data from a data store. In this case, you go to the data store, grab the data that you want, send the data across the wire, and then populate the table. In this scenario, the goal is to get the data in place as fast as possible.

The second way to work with data in this disconnected architecture is to grab a collection of data and use this data separately from the data store itself. This could be on the server or even on the client. Even though the data is disconnected, you want the capability to keep the data (with all of its tables and relations in place) on the client side. Classic ADO data was represented by a single table that you could iterate through; but ADO.NET can be a reflection of the data store itself, with tables, columns, rows, and relations all in place. When you are done with the client-side copy of the data, you can persist the changes that you made in the local copy of data directly back into the data store. The technology that gives you this capability is the DataSet class, which is covered shortly.

Although classic ADO was geared for a two-tiered environment (client-server), ADO.NET addresses a multitiered environment. ADO.NET is easy to work with because it has a unified programming model. This unified programming model makes working with data on the server similar to working with data on the client. Because the models are the same, you find yourself more productive when working with ADO.NET. This productivity increases even more when you use some of the more recent tools such as LINQ to SQL or Entity Framework.