Cyberspace
ORMS has a Farm — EAI, EAI, O!
By ManMohan S. Sodhi
"Our job is 20 percent model-building and 80 percent politics," an OR manager from a large West Coast utility company told me when I was a doctoral student at UCLA. The "politics" was about getting access to databases that were controlled by different departments; data his group needed for such models as capacity planning. Later, when I was an e-business consultant, a recurring challenge was providing access to information "owned" by different departments or even different companies. Happily, access in such situations may become available using software in a new category called enterprise application integration (EAI).
Recall that supply chain planning, at least in the form of packaged solutions from the likes of i2 and Manugistics, and later SAP, did not really become widespread until ERP systems had been implemented. The reason is that the data needed for planning software was now in one system, namely ERP. It is more problematic for OR professionals to build other applications because: (1) existing ERP implementations tend to be limited in scope due to expensive implementation, (2) ERP-like software does not exist for most business processes, and (3) even when some data resides in the ERP systems, integrating with these systems remains a challenge. This is where EAI can help OR professionals.
How EAI Software Works
There are different types of software based on different levels of
integration [1]:
Platform Integration: This is the lowest level of integration across heterogeneous hardware, operating systems and application platforms using messaging for asynchronous connectivity, remote procedure calls (RPCs) for synchronous connectivity and object request brokers (ORBs) for either type of connectivity. Typically, the logic for connecting to each application and routing messages at this level is coded manually.
Data Integration: Data integration products include: (1) database gateways that provide structured query language (SQL) access to heterogeneous data sources, and (2) tools for extracting, transforming, moving and loading data. The latter are generally batch solutions suitable for the initial loading of the warehouse or large batch transfers. Both sets of tools bypass the application logic by extracting and loading data directly from the databases. The resulting solution is a set of system-to-system data maps, which means that changes in any application will impact the mapping to every other system with which it is integrated. OR applications that need to read data from standard databases could use EAI tools at this level.
Application Integration: Unlike data integration tools, software in this category aims to integrate the applications built on top of databases. To do this, they use lower-level technologies to varying degrees to provide near real-time integration. These technologies include: (1) underlying platform integration technology, (2) event integration through message brokers that provide data translation, (3) transformation and rules-based routing in near-real time, not batch, (4) interfaces or adapters to SAP and other leading ERP packages, and (5) custom applications. Application integration is more maintainable than data integration because each application is mapped to a canonical form. Changes in any application impact only this form, not the mapping to every other system with which it is integrated.
Process Integration: This is a higher and more nebulous level than application integration. EAI solutions at this level are intended to enable business managers to define, monitor and change business processes through a graphical modeling interface. Business process integration requires application integration and the other underlying integration services. Some vendors' products sit on top of other EAI software.
Business-to-Business Integration: The next jump is to integrate processes across firms, but in reality it is application integration beyond the four walls of a company. For instance, a supply chain solution for a company may require integration of its applications with those of its customers and suppliers. Some vendors' technologies focus on a company wishing the integrating with others, while others concentrate on creating technology for online trading communities. While the concept of application integration and B2B integration is the same, there are contextual differences, B2B integration is more likely to use the Internet and eXtensible markup language (XML) data formats and involve additional security.
The Role of XML
The Internet and XML together have many advantages over other options for integration. The Internet and intranets are ubiquitous and provide real-time access to users. And XML, being plain text, is platform-independent and human readable. This makes it easier to share data across diverse applications. XML-based software provides standardized message transformation between consumers and producers, and there already are industry-specific standards. Like other new technologies, EAI software may not always match the sales hype. Still, EAI using Internet technologies is worthy of OR professionals' attention so they can spend 80 percent of their time building models and applications and only 20 percent playing politics.
References
- Gold-Bernstein, B., "EAI Market Segmentation," EAI Journal, www.eaijournal.com, Aug. 6, 2001
Dr. ManMohan S. Sodhi (MohanSodhi@AOL.com) is vice president at Gandiva.
OR/MS Today copyright © 2001 by the Institute for Operations Research and the Management Sciences. All rights reserved.
Lionheart Publishing, Inc.
506 Roswell Street, Suite 220, Marietta, GA 30060, USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com
Web Site © Copyright 2001 by Lionheart Publishing, Inc. All rights reserved.