Modern businesses are rapidly adapting cloud based services like Salesforce, Workday, ServiceNow. There are many companies that completely operate from the cloud, or more precise clouds. The benefits of cloud based platforms are clear. But, the companies face new challenges of using and operating many cloud services in parallel and have to build complex integration solutions with the goal to extract the biggest operational value form the integrated services. Creating and maintaining those integrations requires a special expertise and become separate infrastructure project that in many cases decrease the benefits of the use of cloud services.
How to integrate new SaaS efficiently
To increase the speed of SaaS connectivity, many companies use a modern integration platform as a service or iPaaS. With an iPaaS, developers can build, run, monitor and govern integration flows between cloud applications. The typical iPaaS use cases are often less technical and easier to solve than those of traditional middleware. For example, a developer connecting SaaS applications via REST APIs will not care about transactions and compensation mechanisms — often the critical concerns of integration specialists. This same developer will face easy-to-understand data models described in JSON or XML instead of a legacy formats such as COBOL Copybooks or SAP BAPI Business Objects. This allows developers across the enterprise to do work that typically bottlenecked within Central IT.
While there is nothing wrong with simplicity, there is a hidden trap in this logic. Many iPaaS systems accelerate the creation of the technical debt by helping developers build point-to-point connections faster than before. Because point-to-point is so easy, work is often repeated over and over again, creating a lack of reuse in the enterprise. A lack of reuse creates a tightly-coupled architecture that makes downstream changes costly and time-consuming. Over time, this can create an accidental architecture orders of magnitude more complex when compared to the few on-premises applications you once connected via your traditional middleware.
As the scale of connectivity grows, iPaaS users are often confronted with complex integration projects. Example case: a customer needs to copy data into an HR operational dashboard tool. This data originates from several sources, including Greenhouse, an applicant tracking and recruiting software, and Workday, a human capital management software. Complexity very quickly comes to the picture: there is a huge amount of data which can’t simply be transferred via HTTP due to constant timeouts, pagination with complex logic, and throttling limits which prevent access to Greenhouse beyond a certain frequency. Although these are things which an iPaaS can handle, it is hard for someone who is not an integration specialist to digest all of that and make it work in a proper way.
iPaaS experts know formats and protocols of major cloud services and can quickly integrate your services and insure continues functionality of the integrated solution.
Connecting applications in a robust yet flexible way require a balance. We don’t want to prevent people from getting their data by reverting back to the old IT model: enforcing complex governance models, using old-fashioned middleware tools, hiring developers with specialized skill sets, and using traditional waterfall models where only IT can deliver the end-to-end solution. IT and other more technical groups must make all integrators successful without losing governance and control of their core systems and assets. They can do this by providing less technical developers with access to only the set of tools and resources (such as APIs and best practice templates) that are meant to let them become self-sufficient and to allow them to succeed.