XML Schemas Pave the Way for Enterprise Integration
If you're involved in any way with Enterprise Application Integration (EAI), the recent emergence of XML Schema as a standard should be heartening. Although XML is undeniably the lingua franca for B2B, its adoption as a common data layer within the enterprise has been limited. The reason is simple: Schemas, which are critical for enterprise data management, have been a missing piece in the XML puzzle until this point.
XML Schemas and EAI
Enterprise data sources such as relational databases are typically rich in terms of schema definition. Databases have a data definition language (DDL) that define data structures and types. This weakness in XML becomes apparent in EAI, where the primary task is data integration. Schemas allow for data integration to occur as a simple extension of metadata mapping, as shown in Figure 1. When applied correctly, schemas allow for data transformations to be type-safe.
XML Schemas applied in Integration Architecture
Currently, EAI is conducted in other less-optimal ways. One approach binds both data sources through APIs and code. This leads to a brittle system, one that is inflexible and that can be invalidated with changes to the enterprise data models. Newer EAI products such as Sagavista take a giant leap forward by enabling you to map message definitions between messages. This is important because you are managing data integration through metadata mapping and transformations. However, although these newer products can produce and consume XML documents, they do not use XML natively as a canonical form because document type definitions (DTD) provide inadequate schema-management facilities. This changes with the introduction of XML Schema. Let's examine how XML Schema is superior to DTDs.