Today marks the official 50th birthday of one of the original development environments that started it all, which might make for a good time to consider where COBOL and mainframes are headed for the next 50 years.
For all the noise generated about the death of COBOL and the venerable mainframe platform that the applications built using this language mostly run on, both COBOL and the mainframe seem to be doing just fine.
The reason for this is that in terms of creating business-class applications, no language has really emerged that replaces COBOL. We have a plethora of language choices today, most notably Java and C, but neither one has been able to truly match COBOL's ability to handle lot of transactions simultaneously. Of course, much of COBOL's longevity has to do with the transaction capability of IBM's CICS software running on the mainframe. Both CICS and COBOL will forever be joined at the hip.
What's interesting today is that mainframe is undergoing something of a renaissance these days. When IT organizations don't have a lot of money to spend, they concentrate on improving what they have instead of replacing it. This has led to a lot of "application renewal" work that has led to some modernization of COBOL applications that usually focuses on upgrading the user interface and integrating applications with the Web. In fact, as companies look to extend their business processes out on to the Web, it's not unusual to find out that the back-end processing system behind these business processes is the same COBOL application that's been in place for 20-plus years.
The mainframe is benefitting from other trends as well. IBM has brought Linux to the mainframe and created versions of the platform that are optimized for Java and database applications. According to International Data Corp.'s most recent high-end server quarterly tracker, IBM System z market share has nearly doubled on a rolling four quarter basis - growing from 17 percent to 30 percent of the high-end segment-since 2000.
The mainframe, in particular, has been benefitting from a server consolidation trend that should be amplified by the continued adoption of virtualization, and it gets a lot of points for being a greener approach to enterprise computing compared to deploying thousands of Intel-based servers.
The other knock against COBOL specifically and IBM is that nobody is developing any new applications. Micro Focus takes issue with that statement by noting that thousands of developers, particularly in India and Russia, are building new COBOL applications. In addition, Micro Focus notes that it has over 50 universities in the U.S. participating in its Action program to promote the development of COBOL programming skills.
IBM adds that more than 1,000 new or updated applications were introduced for the IBM mainframe in 2008-bringing the total to more than 6,200 unique applications available on the System z platform in 2009. Nearly 3,000 of these are Linux applications, they said. At the moment, IBM says, there are over 1,600 software vendors creating applications for the mainframe today.
IBM adds that it has partnered with 600 universities worldwide to teach various mainframe skills to over 50,000 students over the past four years. And CA has committed itself to a Mainframe 2.0 strategy for delivering a set of management tools for the mainframe over the next several years that will be much more intuitive and simpler to master. That work, they say, should increase the pool of people available with sufficient skills to manage mainframes.
To put all this in perspective, other platforms dwarf the mainframe in terms of popularity. But the mainframe and the applications that run on them still define the term "business critical." So say what you will about COBOL and mainframes, they will be with us another 50 years.