The USA released a comprehensive national health insurance enrollment system within the last year as a method for implementing the new laws and regulations for health insurance. The release was poorly received because the application didn’t work effectively on many fronts.
I do not know any specifics about the application. But there is a lot that can be known due to the implementation of the law. There is no single payer, meaning that multiple health insurance companies are responsible for paying health insurance claims. As a result, multiple health insurance companies would have to integrate their systems into the single health enrollment system in order to facilitate the automated government enrollment process.
No longer is a single system, under the control of a single company, responsible for all the pieces creating a comprehensive user experience. From my experience, this is a daunting task. It reflects the perspective of many that micro kernel implementations on this scale become increasingly more complicated. Micro kernel applications consist of many small applications, doing small things, interacting with each other, forming a larger application than the individual pieces themselves. Any one system alone is pretty easy to manage. However, as the number of micro kernels increase, the complexity increases exponentially.
I’m not surprised that the health care enrollment system experienced difficulties. I’m more surprised that it wasn’t expected. The kind of collaboration necessary to make large scale projects work is difficult to achieve in a single company let alone multiple companies with individual goals.
What have we learned about micro kernel architectures? Are they the only solution for some applications? How do we manage them? What does the future hold? Get into the conversation with your comments or Email to btaylor@sswug.org.
Cheers,
Ben