IBM To Transition Their z/OS, POWER + AIX Compilers To Being LLVM/Clang-Based

[dlcusa]

Really, I'm sick of people that try to look important by pumping up the thing they are working on, that's why I'm low-tolerance.

Don't look in a mirror, then. Really, you have motivated me to ignore you henceforth.

[starshipeleven]

Really, you have motivated me to ignore you henceforth.

Mission accomplished

[dlcusa]

Recently I have seen more and more use of Z/Linux because they want the flexibility and compatibility of Linux based applications, but want the controls of Z/OS. These cases have focused on applications that have a large degree of MQ/DB2 integrations.

Indeed. IBM has long marketed IFLs tailored for Z/Linux that cannot run legacy MVS/VSE/VM/TPF modules, are priced very reasonably, and apparently are highly accepted in the market. I'd love to know how many non-z/OS and non-z/VSE CPCs customers are running these days.

[bridgman]

You must be unaware of just how much z/OS offers out of the box vis-à-vis major Linux distros with thousands of supported packages. Just the size of the z/OS documentation is intimidating. For a flavor, look at just one manual and uncover the capabilities of the WLM (workload manager): https://www-01.ibm.com/servers/resourcelink/svc00100.nsf/pages/zOSV2R3sc278419/$file/izua300_v2r3.pdf and compare to any distro's capabilities in this area.

Yep. I think the real answer is that z/OS and its predecessors have been offering the same kind of features that are just appearing in Linux now for decades. I grew up on 370 systems (my first full-time job was performance-tuning channel programs for a 370/158 system) and 35 years later still find answers (or at least inspiration) for today's problems in what IBM was doing decades ago.

Yes, over time Linux will probably gnaw away at z/OS in the same way that it wiped out proprietary Unix systems, but it's worth understanding that IBM has been offering many of today's cutting edge Linux capabilities to S/3xx and z/OS customers for a long time, and those customers are not all ready to jump ship yet.

Will it last forever ? Probably not, and I'm sure IBM's management knows that. Hardware and associated software represent a fairly small part of IBM's total revenues - maybe $10B USD today - but that is still a big and important business.

Disclaimer - I grew up in awe of the organization and technical expertise that IBM brought to the market, and still use that experience as the benchmark for what we should all be doing for our customers.

That organization and expertise has diffused over the years into a series of spin-offs, copies and derivatives, and at the moment it appears that x86 and maybe ARM along with GPUs and other specialized accelerators are going to be the end-state solution, but to a large extent we are all newcomers fighting over a market that IBM created and serviced for decades.

[dlcusa]

my first full-time job was performance-tuning channel programs for a 370/158 system

Shiny!

[dlcusa]

[...] IBM has been offering many of today's cutting edge Linux capabilities to S/3xx and z/OS customers for a long time, and those customers are not all ready to jump ship yet.

Will it last forever ? Probably not, and I'm sure IBM's management knows that. Hardware and associated software represent a fairly small part of IBM's total revenues - maybe $10B USD today - but that is still a big and important business.

There are factors to keep it going that might sway the outcome. The multi-national corporations in the customer base fully understand the competitive advantage the technology gives them and are very glad the diminished market makes the cost of entry seemingly prohibitive for smaller competitors, perhaps enough to endure profit margins even more favorable to IBM. As you said, for all its turmoil at the top in recent decades, IBM still values the the black numbers Z dependably produces. It will be interesting to see what happens to Z if the rest of IBM tanks.

Intel, AMD, and ARM have all benefited from the expiration of IBM's patents from the 1960s onward, especially those VM-related (that have lapsed as the Constitutional Convention intended--what we've done with copyright extensions is an outrage). While Z virtualization is still king of the hill, it seems IBM does not believe in R&D as much this century as the last. If Z stops innovating...

[zxy_thf]

Mission accomplished

of showing your ignorance of mainframe and fanboism of cloud?

[starshipeleven]

of showing your ignorance of mainframe and fanboism of cloud?

of being ignored by people that aren't giving me useful answers.

I think I made clear by now what I wanted and I really don't think it was hard to guess to begin with.

I can pick up and read manuals even without his "assistence", but I really don't think it makes sense to add another time-intensive hobby to my already limited free time (as there is really near-zero chance of me ever coming close to a position where I'm supposed to operate a Z/OS system, given my current skillset and location).
I just wanted some tl:dr from an expert, that's it. I know already that Z/OS is a big and complex field, why the fuck you think I'm asking other people's opinions? Because it's wildly impractical for me to RTFM and learn stuff on my own on that particular subject.