Types of Application Integration IST 421 Fall 2006

Types of Application Integration IST 421 Fall 2006

n Enterprise integration should enable organizations to become more agile and flexible. This means: Continuously monitor market demand n Quickly respond by providing new products, services & information n Quickly introduce new technologies n Quickly modify business methods n Lee, Siau, & Hong, Enterprise Integration with ERP and EAI, Communications Of the ACM, Vol. 46, Number 2, February 2003, 54 -60.

Coupling n n n The act of bringing together Binding logic with data; level of interdependency between the two components Problems: Creates one application and database out of many, each dependent on the other n Over time, a change to any system means a change to the coupled systems as well n

Cohesion n Act of “sticking” together Applications and databases are independent Changes to any source system should not affect the others

Coupling vs. Cohesion n n The goal is to have loosely coupled components so that there is little or no dependency Integration solutions should be cohesive rather than coupled

Integration Models n n An approach and configuration used to integrate software It emphasizes one or two of the following attributes: Simplicity of performing the integration n Reusability of integration for different configurations n Breadth of possible approaches to integration n Expertise required to perform integration n

Integration Models n There are four general categories to consider for integration: Information-oriented n Business process integration-oriented n Service-oriented n Portal-oriented n

Information-Oriented n Information-oriented approach suggests that integration should occur between databases Data replication n Data federation n Interface processing n

Information-Oriented n Data Replication Move data between two or more databases n Provide the infrastructure to exchange data n Advantages to this technique: n Simple n Low cost n

Data-Level Integration

Data-Level Integration n A variety of tools and data access middleware have been used to access and integrate at the data level Batch file transfer n Open Database Connectivity (ODBC) n Database access middleware – provide connectivity to distributed databases n Data transformation – convert data from source database format into target database format; EBCDIC to ASCII n

When to Use it n n n Combine data from multiple sources for analysis and decision making Provide multiple applications with read access to a common source of information Allow data to be extracted from one source and reformatted and updated in another

Data-Level Integration n Advantages: Tools and techniques currently exist n Relatively easy n Requires no changes to the application logic or database structure n Provides access to a wider range of data n Allows data to be reused across other applications n

Data-Level Integration n Disadvantages: n Need to rewrite business logic

Data-Level Example n n n Customer data from Sybase, IBM DB 2, and SAP R/3 databases combined into a call center application. Middleware would be used to access the data and integrate it into the call center application Call center application would send a request to the middleware

Data-Level Example n Middleware would send the request to each of the databases in the correct format and wait for responses. Responses would be passed back to the call center application.

Other Data Source to Consider n n n Relational databases IDMS IMS VSAM ISAM Flat files

Presentation Integration n One of the simplest models Integration of multiple software components is accomplished through an application user interface Application appears to be a single application although it may be accessing several legacy applications

When to Use it n n Put a PC-based user interface on an existing terminal-based application Provide a single application that is in fact a composite of several applications

Presentation Integration n Advantages: Very easy and quick to do n Developer focuses on new presentation n n Disadvantages: Can have performance bottlenecks n Most limiting model n

Information Integration solution set that works: Identify the data n Catalog the data into a data dictionary n Reason for the data n Ownership n Format n Security n Role in the data structure n n Build the enterprise metadata model n Used as a master guide for integrating

Information Integration n n Maintain integrity issues Data latency Real time n Near time n One time n

Information Integration