COMPUTER FORUM: Sharing Data Between Plastics Business Applications
When you use your sales or purchase order management software, are you constantly worried about whether it contains the same information as your inventory and production software? In the ideal world, each individual part of your software sets would share the same information with every other part, and you should be able to switch freely between them with confidence that all your information is up to date. Obviously, this doesn't happen in the real world and, because of the diversity of software available and individual preferences for particular programs , it isn't likely to happen easily in the near future. This is the primary reason for the existence of systems integration software.
Does that mean we can solve the problem of information systems integration by simply getting the individual components to talk to each other? Fat chance! It is achievable, however, by setting up a software component whose only job is to perform the transfer of necessary information at the time that is needed. If that sounds too easy to be true, it is. It is a very difficult mission, frustratingly enough, due to the natural tendency of software developers (myself included) to use varied development and data management systems. Sorry, there is no standard at present! To help with this mission impossible, however, a relatively small group of software vendors have already put together programs designed to minimize the hassle. Essentially these programs are very literate, data managers that know (or can be taught) how to read the individual types of data found in various data sources, interpret it, and route it to the diverse data repositories that will use it.
I have to caution before this discussion goes any further, however, that certain specialized programs will not benefit from mere data transfer due to the need for internal "flag" settings that can only be established by introducing new information through the program's own procedures. The solution in this situation is to make that program the data source (not the target) for any information it handles. Most accounting software falls into this category, which is fortunate because this is the first piece of software we all seem to buy when we start up a business. That makes it an ideal candidate as a source of information.
The integration process itself typically consists of three steps: 1) define the types and locations of the source and target data structures, 2) build a map that tells the software where to get and put everything and 3) build an API (a mini application) that tells the transfer software when and what data transformation to run. Thanks to modern visual or graphical user interfaces, most of the tedium of this process is less difficult than it was when professional programmers were an absolute requirement for custom data conversion. However, I'll caution that it is not so easy that a novice to data handling should try it. A fundamental knowledge of database systems and data formats is essential to the understanding of the procedures involved.
A look at two vendors' products and processes will give us an idea of the extremes of complexity levels involved in this type of procedure and, perhaps more importantly, of the cost levels associated with the software involved.
Let's start with Data Junction's® DJEngine®. Data Junction evolved its data integration software over the last 15 years, with an approach that grew out of a nicely designed data conversion utility. As the utility grew more proficient, it became possible to use its very capable data handling engine (software that deals with the management of data formats) for automated conversion. DJEngine runs in Windows NT/95® or UNIX® environments. It requires the use of their visual interface to define a two "spoke" (from their data conversion utility) model of the source and target data parameters. This is not a spokesmodel, like Cindy Crawford, but merely a transformation routine that includes a description of the location , types and transform maps of the data. Once the routine is complete, you are ready to build the API described above. DJEngine® has its own proprietary code (language) necessary for the limited amount of programming performed in the creation of the API. Finally, your interface or business software can be modified to trigger the API to begin the transformation routine.
Now let's look at Mercator Software's Enterprise Broker®, a high-end solution designed for custom ERP (Enterprise Resource Planning) and Internet systems. This software grew out of a comprehensive business management approach and matured as one component of a data warehousing system in a LAN (Local Area Network) environment. It grew to encompass Intranet and Internet architectures as they came online (pun intended). It employs a versatile visual data definition system using the spoke and hub, network or pipeline as well as other topologies, and has programming-free integration flow data type, and data mapping designers to simplify your definition of the transformations. But one of Enterprise Broker's most interesting aspects is its continual monitoring, event-driven data integration control. That means you define the data transformations you want it to perform and it monitors the source data so that when information changes, the change is affected throughout the system automatically. That saves you the need to incorporate programming into your business software to affect the data transformation.
Finally, lets look at costs which I know you have all been waiting to hear. The cost of the DJEngine® software for Windows 2000/NT® is $1,150 with 12 months support ($1,500 with 12 months upgrades). Mercator Software's Enterprise Broker, along with a partner (whom it is sold through) to assist in the integration process can run $50,000 or higher before the integration is tested and complete.
The cost disparity between these two products represents extremes for your comparison. The amount of information that could be written on this subject is too lengthy to include complete coverage or even a list of integration software vendors in this brief column. In a future column, I will attempt to review the topic more fully.
For more information, click on the Authors Biography at the top of this page.