Mainframe Battle

Is there battle going on in your organisation over whether you should ditch the mainframe or not ? There could be some big mistakes made in the future if the arguments are not clearly understood.

Cost of migration

Think of you worst case figure and double it. May be an exaggeration, but big systems migrations nearly always hit trouble and cost escalation. There will be whole areas of legacy code that no-one understands, or maybe worse, people think they understand it but in fact have got it totally wrong !


Security and processing speed

So you’ll maybe save some operational costs, but that may be at the expense of reduced processing speed and lower levels of security. The mainframe is the mother of security and speed and you need to think carefully and clearly before moving away from all that.

Skills

COBOL is just a programming language – it does not take too long to train people up to code in COBOL. It’s the individual system functionality that tends to be the complex area that takes time to understand before being competent to design, make and test changes. This system complexity remains whatever the language it is written in so migration to snazzy new application operational and development environments can in the end be disappointing from a developmental point of view. Let’s face it you’ll have a new legacy system in a year or two whatever you do !
The COBOL skills shortage is a myth, rates are down, and if demand goes up substantially due to mass baby boomer retirement then more people will quickly train up in what is essentially an easy to learn discipline. What we may be short of is computing skills in general, as the youth seem to be staying away in droves.


Conclusions

Think seriously of making a mainframe upgrade rather than throwing it out. Look at IFLs zLinux, zAAPs and ZIIPS , hypersockets and Capacity on demand. Then try to make a comparison between the two.