Most of the projects require some or other sort of integrated data for a better experience and improved value of data to the business users or end users. You would need to collate data from one or more systems, sanatise them, process them to create useful information. This data from various systems can be integrationed online (in real-time) or offline also known.
Off-line integrations are essentially driven on the concepts on Extract – Transform – Load and mostly run on a schedule/frequency. Offline integration are performed manually (E.g. using file transfers) in some business / (parts of businesses) where the value the automation brings about-out values the spend/cost. The choice to automate an offline integration usually depends on the frequency at which integration is required how much manpower it saves by automating the process.
On-line integrations are meant for data that are mission critical. By mission critical I mean it impacts quality of service of business users or end users. This is usually the complex bit where we need to understand the various parameters impacting/ contributing to the integration strategies such as event driven integration, publisher subscriber model, event streaming. When the data being integrated is very critical in nature it requires to be ensured delivery on the recieving end. Ensuring delivery becomes challenging when one of the systems being integrated to is offline. This brings in the need for persistence of data in the integration layer so that when the participating system recovers integration can resume, thereby ensuring delivery.
Another factor to keep in mind while designing integrations is reusability. Since it takes considerable time and effort in building the right integrations it’s very important that the integration built is reusable. The fundamental principle that can ensure reusability is separation of concerns (E.g.: into system, process and experience APIs). Each integration needs to perform one and only one atomic unit of task. This concept can be greatly appreciated by OOPs developers from the past. It is the core of the 4th generation tools we are using and still holds great importance when it comes to building good quality reusable code.
Another crucial aspect is Master data management. If you build integration flows that are converse to the master data management fllow you would end up putting the participating systems at a risk of having inconsistent data. When the data is inconsistent across systems it holds little to no value you your business as it would never reconcile across systems. E.g.: A contact record in the CRM that you identify in one system may never be able to match the account it belongs to with the account Id on the contact since it does not reconcile with an Account record in the ERP in case you have incosistent data. These are real life problems that can be resolved through good integrations.
At SaaSTrix we understand the core fundamentals of building right integrations that can help your business transform especially when you are working towards catapulting your business growth it will help ensure your data is still organised.