Integration scenarios
Have you ever come across the words “synchronous” or “asynchronous” in the field of integration? These two strange little words are really important for integrations. That’s why we think it’s worth saying something interesting about them.
When designing enterprise information systems integration solutions, the first thing to do is to design integration scenarios. Integration scenarios consist of one or more messages that are exchanged between systems or applications. The fundamental property of these messages is precisely the property of whether the message is “synchronous” or “asynchronous”.
If your business scenario requires high performance and maximum independence between the integrated systems, in this case the so-called “high-speed” transmission is used. “asynchronous” messaging is used. Typically, it is set up for “exactly once” processing, which means that data is transferred from one system to another exactly once and each party acknowledges the successful transmission as well as the receipt of the message. You can think of it as both parties, when handing over an important document, confirming with their signature the handing over and receipt of the document, along with verification that the envelope itself is intact. The recipient, i.e. the one who receives the letter, does not open or read the letter when receiving it. The recipient will only open the letter when he actually handles it. If the sender requires a reply as to whether the data in the message has been read and processed by the recipient, the sender must send a completely new letter with the reply and repeat the whole process of sending a new letter.
The “asynchronous” mode of transmission is often used in integration scenarios to guarantee message delivery while maintaining maximum performance. Therefore, a sender can send a large number of messages over such an integration scenario. Another advantage is that it is not indispensable for the actual functionality of the integration to verify what performance, what controls and or what application processing is available to the target system.
A completely different method of download is used by the “synchronous” message. In this case, the receiver opens the letter in the presence of the sender, reads the entire content, checks the completeness of the message, processes the message. Then, on the basis of what is written in the message, he puts his reply in an envelope, seals the envelope again, stamps it and hands it back to the sender. In this mode of synchronous message processing, the sender has to wait for the receiver to finish processing the message, during which time he can do nothing else and thus cannot, for example, attend to the delivery of the next letter. While such delivery is much simpler in terms of process, technology and organisation, it is much more demanding in terms of time and therefore performance, and neither the sender nor the postman can influence the processing time itself.
Why are these features so important for integrations?
Using the right integration properties in integration scenarios can do real wonders. A good choice is usually rewarded by working for many years after implementation and tuning, and moreover without much maintenance and operational overhead. Conversely, inappropriate use can put wrinkles on the forehead of an enterprise IT department head, in addition to overextending the implementation budget and dealing with problems on a daily basis.
Although the competition between these key integration features has no clear winner, based on our 20 years of experience, the first choice for us is clearly “asynchronous” management.
And do you have a favourite too ?
