In his previous EAPJ article “Application Landscape Views and SOA” (https://eapj.org/application-landscape-views-and-soa/) Borjan Cace stated that the common way of examining application landscapes insufficiently reflects the importance of the proliferation of API interactions. There, he proposed alleviating this problem by architecture views based on “observation”; i.e. information obtained from IT operations, mainly based on retrieving live data from the components of the integration infrastructure. In this article he goes a step further by proposing architecture views based on the same principle, covering not only API’s but also all other sorts of application-to-application interconnections. Additionally, and unlike the first article, the matter is not approached from the perspective of SOA, but rather the scope covers interactions that are not necessarily modeled with service-orientation in mind.
This is an intriguing subject. What is envisaged is the ability to have a model of the runtime topology of interconnected applications that is accurate at all times. That model would then provide for expressive and insightful views into the application landscape by employing the power of 3D spatial metaphors and other visualization options. What’s more, since the model is accurate, we would be able to examine the structure of the “as is” application landscape programmatically, like analyzing coherence and coupling at the enterprise level.
As appealing as that sounds, there are challenges:
- How to gather the needed information, and how to maintain it to ensure accuracy at all times?
- How to model that information so that it can be used at all levels in the organization, from operations to strategy?
- How to present it?
This article focuses on modeling, but as the solutions to these challenges are interdependent, the author also addresses the other two points above.
Use the button below to access Borjan’s paper, and please leave your thoughts below.