I did have an interesting conversation with a customer recently where given ESG they wanted to move relevant workload to zLinux because the power / VM / container ratio was lower than an x86 farm to lower their carbon footprint; I’m not sure if this is an edge case or not.
For many Z customers they do want their supporting software on their platform to ensure that they have control over those services. Again, its a personal choice. My strategy is to move all functions that do not need to run in z/OS available in containers on any platform the customer chooses to optimize their cost on Z. Exciting times. Matt Hogstrom PGP key 0F143BC1 > On Apr 25, 2022, at 23:29, David Crayford <dcrayf...@gmail.com> wrote: > > On 26/4/22 01:03, Matt Hogstrom wrote: >>> Given this constraint, we try to put CPU-intensive activities (such as code >>> page conversion) on the faster-executing platforms. >> and for that reason zCX is very important for ISVs because it does offload >> to Ziips, provides an open technology platform and for customers that do not >> have vZM or IFLs we can deliver software easier using a familiar CI/CVD >> pipeline. > > The elephant in the room is that almost all customers will have x86 platforms > available and it's trivial to stand up VMs using automation such as Red Hat > Ansible. I suppose it comes down to politics again. The mainframe folks at > some sites may be determined to run everything on z. That's not my experience > with our customers who use our streaming products. They all run Splunk, > Elastic, Kafka stacks on x86 and there's practically zero chance of them > moving to z. I'm not knocking zCX. I think it's a great solution and can see > a lot of use cases where it makes sense both technically and economically. > I'm just highly dubious that running medium/large distributed software stacks > is one of them. > > >> >> Matt Hogstrom >> m...@hogstrom.org <mailto:m...@hogstrom.org> >> >> "Never attribute to malice that which is adequately explained by stupidity.” >> - Hanlon’s Razor >> >>>> On Apr 25, 2022, at 12:44 PM, Ed Jaffe <edja...@phoenixsoftware.com> wrote: >>> >>> The z/OS version is not zIIP-eligible and is therefore running on >>> knee-capped CPs in our environment. When running on Linux for Z it runs on >>> an IFL which is fast (and SMT-threaded). The Linux for x86 version is also >>> fast (and hyper-threaded). Given this constraint, we try to put >>> CPU-intensive activities (such as code page conversion) on the >>> faster-executing platforms. >> >> ---------------------------------------------------------------------- >> For IBM-MAIN subscribe / signoff / archive access instructions, >> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN