Instant Recall With XML Data Caching
By Tony Amos
When a Web application loses momentum as it gains more users, customer complaints aren't far behind. That's when Web developers look for ways to optimize application performance. But occasionally, even a Web application that has a solid design from the outset is burdened by performance problems.
Because the typical three-tier Web application design employs one or more Web servers and a single data store, one all-too-common malady is slow response to database requests. By caching database information on the Web server or application server, you can relieve the database server of some of its repetitive work.
One way to do this is to create an in-memory database on the Web server that maintains a copy of static, read-only information drawn from the database. Not only does this enhance performance by reducing the database load, but you also gain greater flexibility in how your application can use the data. For example, the application could then perform its own sorting, key lookups, and operations on data subsets.
This isn't as hard as it may sound. You can find a prebuilt solution in most developers' toolkits: A good XML parser coupled with an XSLT processor delivers everything you need, and more.
XML Parsers
Prominent XML parsers include Microsoft MSXML, Sun Java API for XML (JAXP), and Apache Xerces. Xerces is the most flexible of the three, because it runs on most platforms and supports Java, C++, and COM. JAXP is limited to the Java platform, and MSXML is limited to the Windows/COM platform.<>
