Building enterprise architecture is a journey, not a destination. Technology changes so quickly that it’s difficult for even the most cutting-edge companies to keep up. However, the evolution of enterprise architecture toward modularity is a concept that’s both been around awhile and also has stood the test of time.
Companies that have successfully made the transition stand a better chance of growing and prospering in an increasingly technical world, so it’s a worthy goal to have for your company.
An API-based enterprise architecture truly brings plug-and-play into the mainstream. In older, legacy systems, for example, if you want to extract certain data or implement a business rule across functional areas, you have to customize multiple systems to collaborate in a non-intuitive way because there is no flexibility in the software.
However, with an API-based infrastructure, extracting that information in real time is easier. So why isn’t everyone adopting an API-based enterprise architecture? The reason is because more established companies have a conglomeration of legacy systems that have been cobbled together in a haphazard manner over time. Data is replicated across many systems, and each one is generally not aware of the others within the environment. There is a lot of the technology equivalent of glue and duct tape holding them together.
Consider the recent outages in the airline industry. As one industry watchdog said after the latest outage in February, “Some of the airlines are still running old IBM mainframes, AS/400 and using software that was written in COBOL.” COBOL, really?
But think about the age of mainstream airlines, how many mergers and/or acquisitions they’ve been through and the degree to which systems are automated. It would be easy to think, “Well, just start over from scratch,” but how does a company put its business on hold while it adopts a totally new software system?
So while plug-and-play, modular software systems should be the goal, it’s a project that should be tackled one area at a time.
Connections to and from cloud applications is a good example of how plug-and-play should work. APIs link your systems to the cloud systems, offering maximum uptime and on-the-go software updates.
If building modular systems were so easy, every company would already be there. But morphing existing systems and their connections into plug-and-play modules is like trying to change a tire while the car is careening down a winding road at 60 mph. With existing businesses, nobody can afford the downtime necessary to complete the task or build it from scratch.
Most of the clients Pyramid works with have either tried and failed to implement an API-based architecture themselves or just don’t know where to start. Any plug-and-play project should begin with a thorough understanding of the existing architecture to determine where and how APIs could be implemented, how they fit together and how to replace parts of the existing system without the entire IT infrastructure failing. Pyramid also helps customers establish roadmaps and ecosystems to engage stakeholders and customers and identify additional business opportunity that API-based architectures offer.
This isn’t the type of project you want to ramp up staffing for, only to let most of the workers go once the project is complete. Rather, you should find a competent partner that has the people and the know-how to help.