The Technology Pendulum: From traditional APIs to cloud services

Information Technology has a pendulum-like history that’s pretty simple to catalog: when the tools for building things become easier to use, more people begin to use them. For example, when the PC’s operating system changed from text-based usage to a graphical user interface, the use of PCs in the work environment skyrocketed. The same is with the case of the REST API – it lowered the technical barrier to entry, and many more engineers began utilizing them.

In the mid-90’s, if you told software engineers they needed to integrate an API into their program, you were often met with groans and subdued demeanors. This was because, more often than not, it was going to be a painful process. You might receive twenty pages of step-by-step documentation on how to do the integration, only to discover part way through there were undocumented errors, and have to start over. Or, you might receive a single page of documentation which basically told you to figure it out on your own. On the other side of that equation, the software engineers building those APIs were not facing a pleasant set of tasks either.

So, when the term Representational State Transfer (REST) was formalized and the REST API (aka web service) was born in 2000, software engineers around the world cheered. The technical barrier to creating and utilizing certain types of APIs was greatly lowered. Systems integration and distributed services using this new approach were all the rage. Much too hyped at first, these web services have become a mature offering. However, RESTful APIs are not a perfect solution either. Issues around security, reliability, data durability, authentication, authorization, non-repudiation, and change management can make any web services implementation into a resource-intensive quagmire.

So now it is time for the pendulum to swing back again. Cloud computing is certainly a far cry from mainframe computing, but notice how Saas, Paas, and other aaSes are all centrally managed and controlled just like the mainframes and dumb terminals (green screens) of yesterday. With each swing of the pendulum, we realize the benefits of the “new” approach and sometimes forget or ignore the benefits of the “old” approach for a little while until we remember that there are always pros and cons to each approach. Ultimately, we need to strive to learn how to combine the best of both.

The repeated history of lowering the technical barrier to entry and striving to distill the benefits of multiple approaches were all on the table in January 2007 when I completed the architectural requirements of our product, Linxter. We labored to take the best of the old world and merge it with the best of the new. Linxter combines cloud computing, web services, and a series of key old school features such as local transactional queues and file chunking to name a few. We’ve designed Linxter so that it is easy to use, can be integrated in minutes, and does not require any specialized coding. So that I do not sound like a brochure, you can read more about Linxter on your own or watch some of our video tutorials and have me or our voice-over talent read to you.

Does the pendulum stop here? Of course not. As engineers, we know that change is inevitable, but we also need to know that it’s up to us to keep designing adaptable tools that make development easier.