Automated Capacity Management_ACM
Relevant to you?
If your datacenter has sub-capacity pricing arrangements with IBM or participating independent software vendors, Automated Capacity Management (ACM) allows your datacenter to significantly lower monthly software license fees, by controlling the batch load at peak times, based on the rolling 4-hour average (R4HA) of CPU usage.
ACM Extensions for Tailored Fit Pricing provide additional controls. You can control batch load based on CPC busy percentage, MVS busy percentage, and percentage of CPC Purchased Capacity. The latter is useful for controlling batch load consumption when IBM's Tailored Fit Pricing Hardware Consumption Solution has been implemented in an installation. CPC busy percentage or MVS busy percentage limits can be useful in reducing the CPU consumption and provide savings in a Tailored Fit Pricing consumption model environment. They can also be used to automate prevention of batch load affecting online response in periods of high utilization.
Any combination of R4HA limits and these new limit types provided by ACM Extensions may be used. In a Tailored Fit Pricing environment, R4HA limits may not apply and therefore do not have to be used.
How does it work with R4HA?
With ACM enabled, ThruPut Manager AE is aware of capacity limits, whether defined as LPAR Defined Capacity, LPAR Group capacity limits, or both, and monitors the rolling 4-hour average CPU usage. You can specify actions and job selection decisions that will automatically be taken to constrain or defer low importance batch work as the system approaches its capacity limit.
You may also define "LPAR Sets", subsets of the LPARs on a CPC, and assign a R4HA limit. As with Defined Capacity and Group Capacity, ACM monitors these limits and allows you to specify actions to be taken as the limits are approached or reached. For those installations that are taking advantage of Country Multiplex Pricing (CMP), Multiplex LPAR Sets can be defined across all CPCs in the multiplex with an associated overall Multiplex limit that will also trigger specified actions.
ACM will also keep track of mobile and public cloud related workloads and discount the R4HA of those workloads automatically as specified by IBM's Mobile Workload Pricing (MWP) and Z Workload Pricing for Cloud (zWPC) pricing options.
By taking these actions, ThruPut Manager AE prevents you from running hard into your capacity limit because of low importance batch work, which can have negative effects on all your workloads, including online. Instead, the selection of low importance workload you identify as eligible is slowed down so that your important batch work can be completed in time and your online response times can be maintained.
How does it work with non-R4HA limits?
With ACM enabled, ThruPut Manager calculates rolling averages for each of CPC busy, MVS busy, and MSU consumption. The latter is used to calculate the percentage usage of CPC Purchased Capacity. You can configure the rolling average interval value for each of these averages independently - the shorter the interval, the more responsive ACM will be to the changes.
As with R4HA limits, you can specify actions and job selection decisions that will be taken automatically to constrain or defer low importance batch work as the system approaches its capacity limit.
See the Automated Capacity Management Guide for more detail.
Implementation considerations
With Automated Capacity Management, you use the ThruPut Manager AE dialogs to identify those service groups that are to be constrained when the R4HA approaches the capacity limit. LPAR Sets are also defined using the dialog.
If you are implementing Multiplex LPAR Sets across multiple Sysplexes, you will need to configure TCP/IP communications which includes specifying the TMSS initialization statement "TM TCPIP" described in the Base Product System Programming space.
This facility is explained in more detail in the Automated Capacity Management space.