9. Database Connectivity | Java Servlet Programming (Java Series)

Java Servlet Programming, 2nd Edition > 9. Database Connectivity

< BACK

CONTINUE >

Chapter 9. Database Connectivity

It's hard to find a professional web site today that doesn't have some sort of database connectivity. Webmasters have hooked online frontends to all manner of legacy systems, including package tracking and directory databases, as well as many newer systems like online messaging, storefronts, and search engines. But web-database interaction comes with a price: database-backed web sites can be difficult to develop and can often exact heavy performance penalties. Still, for many web sites, database connectivity is just too useful to let go. More and more, databases are driving the Web.

This chapter introduces relational databases, the Structured Query Language (SQL) used to manipulate those databases, and the Java Database Connectivity (JDBC) API itself. Servlets, with their enduring lifecycle, and JDBC, a well-defined database-independent database connectivity API, are an elegant and efficient solution for webmasters who need to hook their web sites to backend databases. In fact, both of your authors started working with servlets specifically because of this efficiency and elegance. Although elsewhere in the book we have assumed that you are familiar with Java, this chapter breaks that assumption and begins with a quick course in JDBC.

The biggest advantage for servlets with regard to database connectivity is that the servlet lifecycle (explained in depth in Chapter 3) allows servlets to maintain pools of open database connections. An existing connection can trim several seconds from a response time, compared to a CGI script that has to reestablish its connection for every invocation. Exactly how to maintain the database connection depends on the task at hand, and this chapter demonstrates several techniques appropriate for different tasks.

Servlets in the Middle Tier

One common place for servlets, especially servlets that access a database, is in what's called the middle tier. A middle tier is something that helps connect one endpoint to another (an applet to a database, for example) and along the way adds a little something of its own.

The most compelling reason for putting a middle tier between a client and our ultimate data source is that software in the middle tier (commonly referred to as middleware) can include business logic. Business logic abstracts complicated low-level tasks (such as updating database tables) into high-level tasks (placing an order), making the whole operation simpler and safer.

Imagine a client application that places an order. Without middleware, the application has to connect directly to the database server that stores the order records and then change the database fields to reflect the order. If the database server changes in any way (by moving to a different machine, altering its internal table structure, or changing database vendors), the client may break. Even worse, if someone makes a minor change to the client (either intentionally or accidentally), it's possible for the database to record orders without first receiving payment or to reject perfectly valid entries.

Middleware uses business logic to abstract the ordering process. Middleware accepts information about the order (for example, name, address, item, quantity, and credit card number), sanity-checks the information, verifies that the credit card is valid, and enters the information into the database. Should the database change, the middleware can be updated without any changes in the client. Even if the orders database is temporarily replaced with a simple flat file order log, the middleware can present the same appearance to the client.

Middleware can improve efficiency by spreading the processing load across several backend servers (CPU servers, database servers, file servers, directory servers, etc.). Middleware can also make more efficient use of bandwidth: instead of having a client perform the back-and-forth communication with the server over what might be a slow network connection, the client can tell the middleware what it needs and the middleware can do the work using a fast network connection and probably pooled database connections.

On the Web, middle tiers are often implemented using servlets. Servlets provide a convenient way to connect clients built using HTML forms or applets to backend servers. A client communicates its requirements to the servlet using HTTP, and the business logic in the servlet handles the request by connecting to the backend server. (More information on applet-servlet communication is coming up in Chapter 10.)

Another advantage of servlets over CGI and many other technologies is that JDBC is database independent. A servlet written to access a Sybase database can, with a two-line modification or a change in a properties file, begin accessing an Oracle database (assuming none of the database calls it makes are vendor-specific). In fact, you should notice that the examples in this chapter are written to access a variety of different databases, including ODBC data sources such as Microsoft Access, Oracle, and Sybase.

< BACK

CONTINUE >