Babel: An XML-Based Application Integration
Hua‐Xin Zhang, Eleni Stroulia · 2002
One of the major problems in integrating independently developed applications is the divergence between the data and control-of-processing mod- els assumed by these applications. Research on database integration has focused on establishing and maintaining a canonical schema on top of the schemas of the underlying databases. At the same time, web-accessible software systems have been adopting a multi-layer architecture style, with databases in the lowest tier, business in the middle tier and user interfaces in the top-most tier. However, as the time-to-market window shrinks, new software is presented with the challenge of reusing and integrating the functionalities of existing whole applications, instead of simply their database back-ends. The Babel framework n- alities they deliver and the data they manipulate. In addition, it supports the specification of the logic defining how these functionalities should be inte- grated. Based on these specifications, Babel produces a run-time mediator that monitors the behavior of the underlying applications, evaluates the defined on the global state of the integrated system, and generates triggers for new functionalities to be accomplished according to these rules. As the number of alternative technologies underlying software development increases, the need to develop methods and tools to support the integration of heterogeneous software assets becomes more pressing. Research efforts to that end aimed originally at integrating heterogeneous databases. As the object-oriented paradigm usurped the more traditional procedural approach to software design and development, object-oriented databases were proposed as an alternative to relational databases. As a result, a new area of research was born, f o- cusing on methods for developing object-oriented views of relational data (20). How- ever, even when developed with the same design paradigm for the same application domain, different databases differ in terms of the syntax they adopt for the representa- tion of (potentially the same) data and the semantic (7) interpretation of the data by the applications supported by these databases. Thus, a substantial research effort has been invested in integrating the schemas of heterogeneous databases or constructing unified views of their data (2,3,4) and creating new query languages to query distributed het- erogeneous databases (10).