The Mainframe Endowment Effect: The Reality of Computing

1 May 2019

In psychology and behavioral economics there is the notion of an "endowment effect". This notion suggests that people attribute MORE value to something they already have than something new.

Many mainframe customers firmly believe that no other computing platform can replace the mainframe. There is no question that this has been a stalwart computing environment for the last 5 decades. That doesn’t mean that it is the only technology that can handle the kind of business workload normally associated with this platform. So, the question is whether mainframe proponents underestimate the replaceability of this platform based on this effect.

This is not a debate as to whether the mainframe is and has been a good computing platform. This debate is whether mainframe proponents are unwilling to openly evaluate alternatives because they have an inflated view of this platform and give little consideration to the possibility of obtaining similar performance and reliability characteristics on different infrastructures.

Supporting enterprise workload at scale

It is true that applications on the mainframe have been written to take advantage of the scale-up nature of the mainframe. Need more power, add more MIPS! And the reliability of the platform has enabled organizations to write large scale, high data volume applications that consume large amounts of MIPS and data by a single program. One would write this differently in a modern scale-out environment, and in so doing obtain good performance and limited impact of failures. A new architectural mindset is required, for sure. That does not mean that modern x86 infrastructure cannot run much of existing mainframe transactional or batch workload right now. It can! Further, cloud service providers existence depends on supporting enterprise workload at scale. Let’s not dismiss the 99% of workload that CAN run in this environment, because of the 1% you believe can’t.

One of the interesting anecdotes from our recent production SDM implementation at Swisscom is how several of the mainframe system programmers were completely opposed to this effort. They refused to believe it would work or that it could be as good as their mainframe! Interestingly, after Swisscom had full production implementation of about 2,000 – 2,500 installed MIPS on Software Defined Mainframe, these same doubters expressed the opinion that they wished they had made this move over a year ago! Innovation overcame the endowment effect! This team is now running this workload utilizing only 4 cores of an 8 core system, with performance equivalent to the mainframe. Continued tuning will likely improve the results even more.

Opportunity for innovation on modern platforms

The familiarity of the mainframe can easily blind one to the possibilities of new computing platforms. One shouldn’t be naïve about the move. Don’t let the endowment effect limit your opportunity for innovation on modern computing platforms. The Lzlabs Software Defined Mainframe® is the first step to a modern, lower cost, innovation-centric computing environment. Take a look. What can it hurt?

Want to find out more about modernizing legacy mainframe applications? Read our white paper 'A Graceful Path to Legacy Mainframe Modernization'

Popular Articles

Related Topics