Having worked ‘on The Mainframe’ for almost 20 years at a variety of clients and sites I’ve come about a number of taboos within the Mainframe ‘ecosystem’ that are sometimes just doing my head in and just need to be broken.

These taboos can be seen across the playing field and when I say across, I do mean across. Ranging from banks to insurance companies, from airlines to pension funds and from governmental institutions to ‘narrow-lane truck and heavy container warehouses’ and heard through the vox populi. Yes there are actually global warehouses running their stuff on The Mainframe (and for darn good reasons too!), but that’s maybe something for another blogpost…

This is not meant to be (or interpreted as) a ‘naming and shaming’ rant, but should be read as a guide for improvement.

That being said : here’s the Top-7 list of Mainframe Taboos that should be broken. (and if you can you should help me break them!)

1. It’s too complex to understand and not cool

Really? So setting up a VLAN to accomodate for a cassandra-cluster storing data on volume-groups aggreated over a GlusterFS backend serving data to a memchached REST-API used by a single-sign-on, two-factor-authenticated webgame to crush some annoying candy is an ‘easy-peasy-lemon-squeezy’-task, considered to be cool and doable by any random guy with some basic IT skills? IT is serious business, if it were easy and cool everybody would run a datacenter…

2. We should not IPL too often

I keep hearing this. No we can’t do change “this-or-the-other” because we’re not allowed to IPL more than once a month.
Come on! This is the 21st century, we’re all running sysplexes here (or at least we should be…). A rolling IPL to activate some ‘trivial’ changes or just clear the remnant storage from ECSA will do no harm!

3. It’s an obsolete platform

Go and stand in the corner! There is only one sane reply to taboos like these and it can be summarised as follows. On a more friendlier tone (and subject of an up and coming blogpost here) imagine a world with no running mainframes for a day then tell me again how obsolete it is……

4. Nobody likes a green-screen

Because we all want a sexy and slick GUI, integrated IDE’s with point&click, drag&drop interfacing?
False! Most of the developers I know (this includes developers working on the real sexy stuff like opensource-repositories, the nerds doing frameworks/modules and all in between) are happily chugging away in a green-on-black vim screen. If you don’t believe me : check them out here. Nothing wrong with a green-screen. UserFriendlyness is a personal thing. GUI’s are not a requirement, they never were and never will be.

5. It only runs Legacy Runtimes

Seriously, the online dictionaries define Legacy as follows : “… relating to old or outdated computer hardware, software, or data that, while still functional, does not work well with up-to-date systems…” That being said, The Mainframe has been interfacing with all the ‘up-to-date systems’ as they came and went. It made Legacy 50 years ago whereas the ‘new platforms’ create it on a daily basis. It’s been running Linux since kernel 2.2.13 (yup that’s over 15 years ago, mostly the same time-period the average ‘Senior Linux Specialist’ I come across has on his CV). It has been doing XML and JSON for quite some time and then I haven’t even mentioned ‘dockering‘ or the Java Stack (raw via java -jar or containered via WebSphere). Did I mention TCP/IP already? It’s been doing that from when the rest of the ‘modern platforms’ were still running IPX/ODI…

6. It’s too Expensive!!!

It was not easy not putting this one as the #7 taboo as it’s only just that much less important than the taboo that did reach the #7 spot. Expensive is not the right word. A lot of money, sure but please bring the figures and explain them to me. 9/10 sites will have non-mainframe stuff pushing on The Mainframe budget. Heck, there are companies out there charging the company canteen on The Mainframe’s budget. Also (and if not more pressing) are the hidden-costs of ‘the other platforms’ which mostly are not contributed to that platform (power- and  labor costs to be the biggest if you ask me) resulting in an uneven (and thus unfair) cost-comparison. Besides all these ‘factual’-rebuttals we should keep pushing towards a world where IT is on the ‘value side’ of the Ledger opposed to the spot on the ‘cost side’ it now occupies.

7. There’s no Mainframe to play on

This is the biggest taboo you all need to help (me) break. There are hardly any playgrounds/sandboxes for us mainframe geeks to play on. Au-contraire to the surplus amounts of ‘sandboxes’ for the other platforms. It’s not rocket science to see the amount of ‘innovation’ and ‘new stuff’ coming on the Linux environments is correlated with, if not directly resulting from, the (almost zero cost) availability for everybody to hack away at one whenever they please. IBM has provided us with an ‘x86 emulator’ (zPDT) but this is (cost-wise) still far out of reach for the majority of us. From a legal-point-of-view (I’m talking licenses and what one can and can’t do with it) it’s miles away from just downloading a Linux .iso and ‘hacking away’ on it. Imagine the (extra) innovation achieved on z Systems once an ‘open playground’ will be made available for free so all the geeky guys and girls like you and me can hack away…

Demand your free mainframe environment today! (either on-premisis, at IBM or your local community (SHARE,GSE))

Originally published on Mainframe Debate.

Henri Kuiper is a DevOps Engineer Mainframe-security at Belastingdienst. From early in his career, he worked his way up from a PL/1 Application Programmer to a full-fledged Systems Programmer on the z/OS platform and has since evolved into an all-around Enterprise IT specialist. He is the Founder and CEO of 'zdevops b.v.', a company with only one slogan: Keeping the "z" in Enterprise IT. Among his specialties is his ability to quickly adapt to changing or new situations.
4 thoughts on “7 Mainframe Taboos We Should Break”
  1. Remember, a Legacy System is “Any system that has been successfully implemented”. Does anybody REALLY believe that converting logic from COBOL to Java actually simplifies anything, or that the newly-created legacies will be any simpler to maintain after a few years when their developers have moved on? The fact that COBOL is a primitive language is a minor factor.

    I agree that point #7 is a major issue. The cost of mainframe access is my largest cost in the development of MANASYS Jazz (see https://www.jazzsoftware.co.nz), a system to slash development complexity.

Leave a Reply

Your email address will not be published. Required fields are marked *