There are three classes of boost: IPL (startup) boost, recovery process boost, and shutdown boost. Each class has different capabilities.

Prerequisites

• IPL boost is enabled by default and delivers extra processor capacity after an IPL to get you back up and running faster.

  Learn more about IPL boost

• Recovery process boost provides increased short-duration processor capacity for the acceleration of sysplex recovery situations.

  These recovery process boosts are enabled by default:

  • HyperSwap
  • Coupling Facility data-sharing member recovery
  • Coupling Facility structure recovery
  • Sysplex partitioning
  • HyperSwap configuration load (IBM z16 and later)
  • Dynamic I/O activate (IBM z16 and later).

  Beginning with the IBM z16, you can enable additional recovery process boosts for SVC dumps and middleware region startup.

  Learn more about recovery process boost

  Enable SVC dump boosts with a dump size threshold

  Enable middleware region startup boosts with the WLM service definition

• Shutdown boost enables a faster shutdown by delivering extra processor capacity upon indication that a shutdown is in progress.

  Update your shutdown automation to automatically use the START IEASDBS system command during your shutdown process to use shutdown boost to speed up the shutdown.

  Learn more about shutdown boost

There are two boost types, speed boost and zIIP boost.

• Speed boost is a capability of System Recovery Boost that improves the recovery time of exploiting operating systems when running on a subcapacity CPC. If you are running on a subcapacity CPC, then while System Recovery Boost is active, z/OS will request that the CPC firmware increase the CP speed of the image to full capacity model speed for the duration of the boost. After the boost ends, the image will return to the subcapacity model speed.

  Speed boost applies to each of the three boost classes: IPL boost, recovery process boost, and shutdown boost.

  Speed boost applies only to the image being boosted; all other images not being boosted will run at subcapacity model speed.

  If you’re running on a subcapacity machine, there is no reason not to run with speed boost. Any processing activity that uses CPs will benefit from speed boost. Use the following links to learn more:

  Turning on System Recovery Boost

• If your system has z Integrated Information Processors (zIIPs), then zIIP boost can improve z/OS recovery time, assuming zIIP capacity is available to the image.

  z/OS is the only operating system that can exploit the zIIP boost capability, as it’s the only OS that can natively exploit zIIPs. While zIIP boost is active, z/OS makes most non-zIIP eligible work zIIP eligible, thus allowing most work to run on zIIPs if there isn’t sufficient CP capacity available. This provides additional capacity and parallelism to accelerate processing during the boost periods. IBM refers to this as blurring the CPs and zIIPs together.

  While zIIP boost permits certain kinds of general purpose (GP) work to run on zIIP processors, the zIIP processors retain their characteristics as zIIPs during a boost period, and certain types of GP work continue to run only on GP processors, even with zIIP boost active.

  zIIP boost applies to each of the following boost classes: IPL boost, recovery process boost, and shutdown boost.

  Learn more about zIIP boost

  There are variations of zIIP boost, depending on how the CPC is configured.

  The most basic type of zIIP boost is when you have defined initial zIIPs and no reserved zIIPs that are backed by real processors. In this case, the zIIP blurring happens for the duration of the boost on your defined, entitled zIIPs and at the end of the boost period. Normally, non-zIIP eligible work stops running on zIIPs.

  When you have some zIIPs that are defined as reserved, and some number of those can be backed by physical zIIPs. Those reserved zIIPs that can be backed by physical zIIPs are brought online by the system at the beginning of the boost period and then taken offline when the boost period is over.

GDPS® performance and parallelism increases the speed with which GDPS drives hardware actions, along with enhancing the speed of the underlying hardware services. The GDPS enhancements are available in GDPS 4.2 and above running on an IBM z15 or higher.

GDPS takees advantage of the HyperSwap exploitation of recovery process boost.

IBM recommends reading the following GDPS manuals, which describe how to implement these RTO improvements:

• GDPS Metro Planning and Implementation Guide (ZG24-6755), section 4.7.2, SYSPLEX script statement.
• GDPS Global - GM Installation and Customization Guide (ZG24-6757), section 4.5.7, SYSPLEX script statement.
• GDPS Global - XRC Installation and Customization Guide (ZG24-6759), section 4.6.9, SYSPLEX script statement.