I have been doing Enterprise Application Integration my whole career, and while there is certainly a great collection of tools and resources available to us, there’s something missing. Regardless of language and technology stack, there’s a need to answer certain questions before beginning any integration project. This simple, obvious and crucial process has been overlooked and I’ve decided to fix that.
The list below contains the questions I ask myself before starting any integration, regardless of ESB. It is by no means exhaustive, and in order to keep it useful, I’ll maintain it here with new questions that inform best practices in design patterns. Furthermore, I’m sure those of you reading this have an experience where answering a question beforehand could have saved a lot of trouble. Hopefully between all of us, we can keep this a useful, living document to help create successful integration projects.
Integration Checklist
- Transport protocol (S/FTP, MQ/JMS, HTTP/S, etc…)
- Asynchronous vs. synchronous
- Frequencies/schedules
- Peak volumes/times
- Response times expected
- Logging/reporting needs (if any)
- Exception notifications (what needs to happen on failures)
- Correlation of messages; message expiration and/or message sequencing requirements
- Sample data sets (CRUD – create, update, retrieve, delete as required)
- XSD/WSDL’s/database schemas if they exist
- Point of contact (SME)
- Security requirements (username/pwd’s, SSL certs, etc…)
- Test clients if one exists today
- Canonical data representations
If you have any other items that I missed on this list feel free to message me and I will update accordingly.
