AI System Services for IBM z/OS
Will they be supplying the entire contents of the internet on punched cards in order to train it?
IBM has, as promised, announced an imminent upgrade to z/OS, the operating system for its Z-series mainframes, and promised an infusion of AI enablers. IBM IBM shrinks z16 and LinuxONE systems into standard rack configs READ MORE The upgrade will see z/OS jump from version 2.5 to 3.1. According to Big Blue's announcement of …
About few years ago the buzz word was "smart" every home applicance to cars would have something "smart" built into it. Like smart water heater, smart bikes..
Now that the buzz word is AI every company is jumping into this bandwagon and calling all their products AI. Open AI took close to a decade to get their AI in a decent state and it still has errors.
But these IBM and all others just came into AI scene and they say their products are AI?? REALLY??
IBM has had watson, but you can go to IBM watson home page and chat with it. It's DUMB as a rock, even on questions about IBM mainframe and products.
I think it's all a ruse to appease share holders
I could go into a litany of sins IBM has foisted on the community, but I will keep it to a few of the more agaregarious things IBM has done to the community and some good things IBM has done over the years.
Great things IBM has done: IBM was probably the first manufacturer to stand up and say enough is enough to egregious holes in their operating systems by standing up and issuing a statement of System integrity. The statement enabled IBM to clean up its act and forced vendors to clean up their act—more than a few vendors at the time flouted system integrity. Thus causing systems to fail (crash) or create incorrect output. This statement was a godsend to systems people and their managers to enforce system reliability. When MVS first came to our data center, I had to give the programming managers repeated classes on what that meant to them and what will happen as a result. After implementing MVS across our four mainframes, it was clear that IBM promises had been kept, and out outages were dramatically reduced by 85 percent. It took a while to get there as we were fighting a vendor that was politically connected to the company. Every time we would get a stand-alone dump, we would go through it and find the culprit that caused the system to crash. We would promptly take the dump to the vendor and show the person what their software did to crash the system. The dumps were stacked 10 feet high. Once we showed senior management what was happening because of the politically connected vendor, they were given six months to clean up their act, or they would be shown the door.
Another good thing from MVS was the mind-numbing manual effort of applying fixes to the operating system. IBM came out with SMP (Systems Management Program). The first iteration was slow as anyone could imagine. The second version was much faster and also easy to use.
The side issue was that there were quite a few fixes to the operating system, a lot of System Programmer time, and other facets of the Operating system that kept our department busy and several new disks (real and virtual). We had a small lean group, and many a time, I would spend a 100-hour work week. Along the timeline, there were many improvements to the operating system (like support for multi-processor): new disk/tape drives, and too many things to mention here. We had many battles with Vendors and, yes, IBM. We learned early on that the Guide/Share user groups were an excellent way for you to communicate with IBM issues that systems programmers had to face daily and for improvements IBM could make to their software/hardware. One item that we fought and lost with IBM was IBM manuals.IBM was always outstanding with its documentation. IBM decided to go Object Code Only ( no more (or little)) source would be made available to customers. We fought long and hard, but IBM was vehement that we lost that major battle, and I think that is what turned the tide (for the worse) for IBM. IBM promised to keep manuals up to date, but it was not enough. IBM broke their promise for some of it (not all). IBM is still fighting battles with PSF (to name one of them), and I was proud to be a First Customer Ship, and I have the scars to prove it.
I was on the phone several times a day with the PSF people attempting to fix code that wasn't (IMO) tested properly before shipping the product. I was also in the FCS (First Customer Ship) program with their "new" way of shipping and applying maintenance to their new CBPDO. Of course, I found a minor bug that put me behind in the schedule for our acceptance of new disk drives and Printers from IBM. If I hadn't been going to SHARE then, I would not have known who to call at Gaithersburg. I called the person and was immediately put in touch with the right person at the CBPDO time. The person agreed with me (long story and difficult to explain), and he was able to fix the issue 3 days later. A tape showed up, and everything went smoothly after that. I got to system up, but I missed a week for my deadlines.
Over the years since then, there have been one or two attempts from IBM to simplify the systems programmer's job. The latest (is, in my opinion, not a good step for IBM). I have found that the new way of applying maintenance is, at best slow and cumbersome. The people from IBM think they can teach any desk jockey can install a new operating system in one or two days. Sorry IBM, it takes much knowledge to do this, and there is much more than IBM is willing to admit to install and maintain an operating system. IMO, IBM is trying to satisfy management to reduce system programmer costs. The long OT is just one function of the system programmer's job. We have to hold hands with many different people in the IT department. And we analyze memory dumps, talk to vendors in a technical role, and watch out for fly-by-night vendors.
In the 'things that IBM got right' category I'd place a later product, OS/2. Not just OS/2 as such, but its approach to support.
When you called in, the first thing you were asked was your support license code. It turned out that IBM had tied a qualification system to that number, which resulted in me always immediately put through to second level support because they had already worked out I would not call for things first level would be able to solve. Frankly, it was the most efficient support facility I have ever dealt with.
I've not dealt with IBM in any other way (although that is about to change, apparently) so I don't know that that is their norm, but I appreciated this as it saved a LOT of time.
IBM has an 18 month hype cycle, where their sales droids get some new buzz words to sell to the largely IT Illiterate CxO suite.
Whether its Blade Servers, or "Virtual Machines" or CAMSS or "Hybrid computing" or SoftLayer or Watson or "the Cloud" or Blockchain or redhat or AI or another mainframe, there's always something new for the suckers to spend a lot of money on...