Getting different applications to talk to each other is a problem faced by many companies. Application integration is particularly troublesome when legacy systems are involved, and transferring data between old mainframe systems and more modern applications can be cumbersome and time-consuming. A good deal of time is spent on ensuring this happens, and software is often available to help with the process.
There is a lot of middleware around designed to help overcome these problems, and this is often very helpful. Failing this, a procedure to download, from one database, convert the data, then upload to another database might be necessary. Such procedures are often unstable and demand the attention of highly skilled people.
Originally it was felt that object oriented databases would overcome these problems for new applications. However, many applications are developed independently, and the problem of duplicate entities persists. The solution would be to have one database, with all applications making use of it, especially for common data.
However, packaged solutions would still depend on their unique databases. This means so there will still be considerable redundancy, necessitating communication via some form of middleware. With packages often being preferred, there is really no solution in sight for this dilemma.
The best solution would seem to be a central repository database which houses all data common to several applications and provides the functionality to maintain and inquire on the data. New applications or packages would interface to that master database. Data used by only one system could be handled as appropriate.
Unfortunately this scenario is very much one for the future. At the moment the world of application systems is very imperfect, as companies still struggle to come to terms with their model T applications. What is needed is the same sort of attitude as with machinery, where equipment that still works is quickly replaced by more efficient machines: an approach that drove America to the forefront among manufacturing nations.
A disproportionate amount of time and money is absorbed by application integration. Many smaller companies cannot deal with the complexities, and simply choose to live with the problem. In time it is possible that a move might be started towards rationalizing the whole approach to applications. Maybe one day everybody will look back and laugh at the current state of the art, after computers have fully emerged from the Dark Ages of systems development.
There is a lot of middleware around designed to help overcome these problems, and this is often very helpful. Failing this, a procedure to download, from one database, convert the data, then upload to another database might be necessary. Such procedures are often unstable and demand the attention of highly skilled people.
Originally it was felt that object oriented databases would overcome these problems for new applications. However, many applications are developed independently, and the problem of duplicate entities persists. The solution would be to have one database, with all applications making use of it, especially for common data.
However, packaged solutions would still depend on their unique databases. This means so there will still be considerable redundancy, necessitating communication via some form of middleware. With packages often being preferred, there is really no solution in sight for this dilemma.
The best solution would seem to be a central repository database which houses all data common to several applications and provides the functionality to maintain and inquire on the data. New applications or packages would interface to that master database. Data used by only one system could be handled as appropriate.
Unfortunately this scenario is very much one for the future. At the moment the world of application systems is very imperfect, as companies still struggle to come to terms with their model T applications. What is needed is the same sort of attitude as with machinery, where equipment that still works is quickly replaced by more efficient machines: an approach that drove America to the forefront among manufacturing nations.
A disproportionate amount of time and money is absorbed by application integration. Many smaller companies cannot deal with the complexities, and simply choose to live with the problem. In time it is possible that a move might be started towards rationalizing the whole approach to applications. Maybe one day everybody will look back and laugh at the current state of the art, after computers have fully emerged from the Dark Ages of systems development.
About the Author:
Peggie K. Lambert is an expert in data management and integration. Need to learn more about application to application integration? If so the author recommends you visit www.liaison.com.
No comments:
Post a Comment