Scalable filtering of XML data for web services - Internet Computing, IEEE

Scalable content-based routing architectures for Web applications can handle the growing number of XML messages associated with Web services. A s the Web gains prevalence as an application-to-application communication medium, organizations are deploying more Web service applications to provide standardized, pro-grammatic application functionality over the Internet. Web services use open standards based on the Extensible Markup Language (XML, www.w3.org/TR/REC-xml), such as the Web Services Description Language (WSDL, www.w3.org/TR/ wsdl12) for service definition and the simple object access protocol (www.w3. org/TR/SOAP) for service invocation. Users can reach the services using SOAP and construct requests using the Web ser-vice's WSDL information. A wide range of domain-specific specifications are also based on XML, such as ebXML for business to business interactions and FpML for financial data exchange. Even HTML, arguably the most widely used data format on the Internet, has recently been rewritten as an XML-based specification, Extensible Hypertext Markup Language (XHTML, www.w3.org/TR/xhtml1). Filtering, classifying, and routing the growing number of XML messages associated with Web services requires scalable mechanisms. Enterprise application servers, for instance, must scale to numerous clients, provide high throughput, and support a variety of XML-based protocols. As Figure 1 illustrates , a Web server (generally associated with a firewall) receives XML data, and one or more XML routers filter it. These routers dispatch XML data, according to its type or content, to the appropriate back-end server, possibly using load balancing, selective multi-cast, or another routing scheme. XML routers can also act as a sophisticated firewall by filtering out unauthorized or invalid XML messages.