· 4 Min read

Logical Partitions

PR/SM empowers the z13 for logically partitioned operations, boasting support for up to 85 LPARs. Each LPAR operates its own OS image independently, allowing for flexible management including addition, removal, activation, and deactivation without disruption or the need for Power-On Reset (POR). However, it's important to note that not all facilities may be accessible across all operating systems due to potential software prerequisites.

Processor Resources

Each LPAR is equipped with resources akin to a real Central Processing Complex (CPC), including:

  • Logical Processors: These can be categorized as CPs, IFLs, ICFs, or zIIPs, and allocated as either dedicated or shared among LPARs. Shared processors have a defined weight to manage resource distribution and an optional capping feature to limit processor consumption.
  • Dynamic Management: The LPAR CPU Management function, part of the Intelligent Resource Director (IRD), allows for real-time adjustments in processor weight and online logical processor numbers to meet specific partition and system-wide goals.
  • Provisioning Architecture: The z13's provisioning architecture, enhanced by PR/SM, supports physical processor capacity limits for individual LPARs, ensuring efficient resource utilization.

Memory and Storage

  • Dedicated Memory: Each LPAR must have dedicated main or expanded storage available at activation time. Reserved storage can also be defined for non-disruptive memory adjustments.

Channels and I/O

  • Shared Channels: LPARs can share channels by listing the partition name in a channel-path identifier (CHPID). The configuration is managed through IOCP or HCD, with tools like the CHPID mapping tool (CMT) for physical channel ID mapping.

Workload and Licensing

  • Soft Capping: For z/OS Workload License Charges (WLC) and related metrics, LPAR defined capacity supports soft capping, allowing software license fees to be based on actual processor utilization.
  • Dynamic Capping: Workload Manager (WLM) monitors a 4-hour rolling average of processor usage, dynamically applying or removing caps based on defined capacity limits.

For detailed insights into software licensing and workload management, refer to the System Programmer's Guide to: Workload Manager, SG24-6472, and the section on software licensing considerations.

Dynamic CHPID Management

The Dynamic CHPID Management (DCM) function, part of the IRD, optimizes FICON channel allocation in response to changing requirements, enhancing overall system efficiency.

Copyright IBM Corporation