Episode 26 – Modern Day Mainframes: Interview with Craig Mullins
Ever since we did Episode 7 -Why are Mainframes Still Around? we were left with the realization that really all the podcast involved was a bunch of non-mainframe programmers making speculations about why the dinosaurs didn't become extinct in our world?
Granted, I did a bit due dillegence prior to Episode 7 via StackOverflow to get others opinions on the matter but still, my journalistic spidey senses tingled at the fact Episode 7 was a one-sided debate for the most part.
To remedy this, a friend put us in touch with Mainframe Consultant and Author: Craig Mullins. It was an eye opener of a podcast, Mainframes truly haven't lost their "super computer" moniker even in today's advances with cloud computing.
Around the time I was arranging this interview I thought to get in touch with a previous colleague named Glen Sepke. We went to grab a beer and I told him my aforementioned feeling about Episode 7. He agreed it was too one-sided. So I offered to have him on the podcast so it wasn't just us against Craig Mullins 
So Why Do Mainframes Still Exist? As "modern developers" we really need to ask these types of questions and get credible answers. Otherwise, people tend to alienate that which they don't know. It's why developers (and, yes, I was in this camp), snicker at the term "COBOL" and "Mainframe". Its because we truly don't understand them, and thus a lot of FUD is created. Mainframes are perhaps becoming even more relevant now that they can run Java and talk to the outside world via Web Services.
So give this episode a listen and open your mind to allowing Mainframes their rightful place in today's modern computing arena.
Listen here:
Enjoy!
About the Author
On 
By the way, do you know how the name HAL was created? Think IBM, and then go one letter back with each letter:
I –> H,
B –> A,
M –> L.
@Craig S. Mullins
Interesting, I didn’t know that actually. I liked the books, perfectly depicts how 1950s-60s people viewed their place on the technology curve and extrapolated that we’d be colonizing the moon and venturing out to the planets by now.
In summary: where’s my flying car?
I just came upon this podcast and thought it was pretty interesting. Many of the questions could have been answered easily by simply reading IBM’s sale literature for any recent model. I would have liked to have heard more opinions and comparisons of mainframes vs. ordinary server clusters, by the two experts and less chit-chat from the interviewers.
Thanks for listening Pierre. Unfortunately if you don’t like “chit chat” you’ll probably hate the rest of the casts, it’s a pretty laid back “off the cuff” type show. That being said, I agree with your assessment of this cast in that we need someone from the non-mainframe world who knows a lot about current server techs, cloud and distributed processing techniques to really nail a Mainframe’er on some of the claims they have on performance vs. cost. If you know any, send them our way!
I don’t think IBM would be eating up and shelfing companies like PSI [1] if it weren’t shit scared of the possibilities; not to mentioning threatening legal action against OSS Mainframe emulators [2]
[1] – http://www.theregister.co.uk/2008/07/02/ibm_buys_psi/
[2] – http://arstechnica.com/open-source/news/2010/04/ibm-breaks-oss-patent-promise-targets-mainframe-emulator.ars
@craig
Interesting links. I had no idea that hercules was in a state to challenge IBM–maybe they’re hoping to suffer the same fate as psi?
As for cost/performance, I think it would be impossible to resolve that debate. What is the value of fault tolerance including soft logic faults, being able to run full db2, redundant and hotswappable processors, database transactions per second as fast as you want, centralizing all IT in an organization, full backwards compatibility to 1960, reliable parts and service, and on and on? vs. the probability that your organizational needs will change and you’ll be somehow locked in and unable to upgrade to more appropriate tech?
Even though I don’t really care about any of those things whilst doing my personal computing, a used G5 s/390 runs about the same price as a high-end laptop these days and the inner geek in me finds that hard to resist, even with all the electricity it would consume. (I am trying to convince the wife that it could heat the house in the winter.)
To me at least, the mainframe seems to fit the centralized (also monopolist?) business model of certain other large, centralized businesses well. The biggest harm that comes from an IBM monopoly, at least from my point of view, is loss of innovation amongst mainframe makers and that other businesses are forced to move their servers towards being more mainframey. The distinctions are starting to blur, and mainframes are probably losing ground. In a more competitive environment, I think we wouldn’t even think of asking: are mainframes really better? It would be obvious they were.
I wish your site had podcast transcripts.
Worlds fastest CPU? IBM zSeries may have the highest clocked CPU in the world, but that does not mean that it’s the fastest one. IBM does not like benchmarks, and when someone doesn’t want to be compared against the competition, there’s usually a reason for that.