A few weeks ago I found myself in a meeting with the technical team at a major investment company that uses mainframes to support the massive amounts of data they work with every day. I spend just about all of my time talking with mainframers, but this conversation took a bit of an odd twist: they wanted to talk about application programming interfaces (APIs).
It wasn’t what I expected, but after thinking about it more, it makes perfect sense. After all, APIs are driving just about everything these days. It’s no wonder so many people are (unironically) talking about the “API economy.”
+ Also on Network World: The cloud’s silver lining: the mainframe +
Why are APIs so important in the mainframe ecosystem? The simple answer is that they let mainframe shops take care of business in a similar but different way than they did in 1975. You see, instead of creating APIs that are exposed only to other locally run programs, they are creating APIs that can be utilized by both locally running programs, even COBOL, and applications that run on remote machines. These remote machines can be other mainframes, on-premise servers or even a public cloud server.
What kind of APIs? Well, certainly the most popular today are RESTful APIs, followed by NoSQL/MongoDB JSON APIs, and then good ol’ SQL. JSON is the lingua franca for mobile application programmers, while the BI analytic applications still like to treat everything like it was a relational database table—even treating some applications like they are data. Cool, huh?
Beyond cool. It's essential.
But it’s more than cool—it’s critical for organizations that rely on mainframes. That’s because few people in IT organizations (and fewer people under age 50) know COBOL. The trick is not to teach everyone how to program a mainframe, but to make mainframes act like the computers everyone already knows. New tools that allow seamless use of mainframes are rapidly becoming the standard, making big iron more accessible to everyone regardless of experience. To paraphrase the punchline of an old joke, it’s a lot easier to raise the height of a basketball net than it is to lower the floor.
This is where APIs come into the picture.
In the case of the investment company with whom I met, they want to build new applications to access, and update, data where it is currently stored on the mainframe. In their specific case, the data is in VSAM, and they want to provide both a JSON API as well as an SQL API to their application programmers. The application they mentioned was one that allows an investment advisor to change the investment options for a client. How great is it that they can do this and not spend time educating the application programmers on what a VSAM key-sequenced dataset is? The application programmers simply connect to this data store, and the data feels the same as if it were in Oracle, SQL Server or DB2.
APIs essentially open mainframes up to a whole new generation of programmers and engineers. With so many financial and insurance firms relying on big iron for their computing needs, and so many millennial techies unfamiliar with mainframe techniques, APIs serve as a gateway and translator for the skills of folks just getting into tech. Professionals unfamiliar with mainframes can nonetheless program the big machines to serve their organization’s best interests.
APIs make mainframes accessible to any stakeholder, securing the future of big iron and the companies that use it, and introduce new programmers to the potential of these powerful machines. APIs are crucial to the backbone of computing, and with their ability to adapt to changing conditions, mainframes will remain a major market within the API economy.
This article is published as part of the IDG Contributor Network. Want to Join?