That Old IBM zSystems Magic
Recently, I read an article that got me thinking about the impressive capabilities of the CPU in IBM’s s390x-based boxes. When my laptop (not IBM, they don’t make those any more) had a logic board issue that required sending it back for replacement, I realized this never happens with an IBM zSystems box.
Today’s z CPU, like the new IBM Telum™ dual chip module that runs in the IBM z16, is a true powerhouse. It boasts 16 processor cores, runs at impressive speeds (up to 5Ghz,) and offers extensive performance and capacity across various workloads. The largest model, Model A01, scales up to 200 configurable cores and has up to 40 TB of Redundant Array of Independent Memory (RAIM).
The reason these systems avoid the need for CPU repairs lies in their advanced recovery options. Each CPU core can undergo a transparent core reset, quickly restoring the system’s state without any data loss. Moreover, spare CPUs are ready to take over instantly in case of core issues, ensuring uninterrupted operation. The occurrence is so rare, IBM has actually cut back the number of included spares to 1 or 2 depending on which model you buy.
Additionally, the box is usually bigger than you will need and there are going to be additional unused cores that can be utilized on-demand. These systems are designed to handle any rare CPU core failures completely transparently, without intervention, with ease. So, the need for extensive (read as human technologist) repairs is almost non-existent when you’re talking about the CPUs themselves.
It’s clear that the magic inside IBM zSystems that keeps the system up and running with only minutes of downtime required in a year is quite remarkable. While having such capabilities in a laptop seems far-fetched, the idea of seamless, transparent recovery is undeniably appealing. Wouldn’t you want that option too?