|
|
||||||||||||||||||||
| Home > SOA News > Explaining integration II | |
| SOA News: |
|
||
Market Analysis
Explaining integration II To discuss this let us start with two applications A1 and A2 with their related datastores D1 and D2. It is necessary to distinguish integration requirements from integration implementation (this is another line in the sand that can not be defined accurately). I will first describe some integration requirements then describe some implementations. There may be a requirement for an event in A1 to trigger an event in A2, for example creating an invoice in A1 will trigger the processing of that invoice in A2, and this is an application integration requirement. There maybe a requirement that when some data changes in D1 it is reflected in D2 but there is no requirement for the change in D1 to trigger an event in A2. Datawarehousing is the classic example of this requirement. But there are other requirements such as reference data, where master reference data in D1 is propagated to D2. A2 then has the latest reference data whenever it needs it, and it is assumed that there A2 does not have to do any processing just because of the change in D1. Datawarehousing and reference data are two examples of requirements for data integration. If we now look at the implementation options there are a variety of application integration and data integration technologies. Messaging, Web services, RPC etc are application integration technologies. ETL, data-propagation, data-replication etc are data integration technologies. The confusion factor is that there is not a one to one mapping between requirements and implementation. Let me explain. If the requirement is for A1 to integrate with A2 then the obvious solution is for A1 to use one of the application integration technologies to communicate with A2 but that requires A1 to have been written or modified to do this. If it is a legacy application this may not be possible; but data integration technologies may provide a solution. The change to D1 can be captured by data integration technology and put in a database to be picked up and processed by A2. If the requirement is for changes in D1 to be reflected in D2 then the data integration technology offers a number of implementation solutions requiring no processing by A1 or A2. However increasingly in this service oriented world application are broadcasting messages to anyone who may be interested. In this case messages from A1 can be used to trigger changes in D2. So we are using application integration technology to implement a data integration requirement. I would suggest that application integration technologies provide a more flexible environment and will become increasingly important over time. However data integration technologies have two possible advantages: performance, and the ease of ensuring that all changes get propagated. I hope this short note has spread some enlightenment but I also know that it will raise further areas of discussion and I am sure Phil and I will return to it in these columns.
Copyright 2004. Originally published by IT-Director.com, reprinted with permission. IT-Director.com provides IT decision makers with free daily e-mails containing news analysis, member-only discussion forums, free research, technology spotlights and free on-line consultancy. To register for a free e-mail subscription, click here. For more information:
'); // -->
|
|
|||||||||||||||||||||||||
| About Us | Contact Us | For Advertisers | For Business Partners | Site Index | RSS |
|
|
|
|||||||