Resource types in Workload Manager

WLM manages three types of resources on a percentage consumption basis.

Item	Description
CPU utilization of the threads in a class	This is the sum of all the CPU cycles consumed by every thread in the class.
Physical memory utilization for the processes in a class	This is the sum of all the memory pages which belong to the processes in the class.
Disk I/O bandwidth for the class	This is the bandwidth (in 512-byte blocks per second) of all the I/Os started by threads in the class on each disk device accessed by the class.

Every second, WLM calculates the per-class utilization for each resource during the last second, as a percentage of the total resource available, as follows:

For the CPU, the total amount of CPU time available every second is equal to 1 second times the number of CPUs on the system. For instance, on an eight-way SMP, if all the threads of a class combined consumed 2 seconds of CPU time during the last second, this represents a percentage of 2/8 = 25%. The percentage used by WLM for regulation is a decayed average over a few seconds of this "instantaneous" per-second resource utilization.
For physical memory, the total amount of physical memory available for processes at any given time is equal to the total number of memory pages physically present on the system minus the number of pinned pages. Pinned pages are not managed by WLM because these pages cannot be stolen from a class and given to another class to regulate memory utilization. The memory utilization of a class is the ratio of the number of non-pinned memory pages owned by all the processes in the class to the number of pages available on the system, expressed as a percentage.
For disk I/O, the main problem is determining a meaningful available bandwidth for a device. When a disk is 100% busy, its throughput, in blocks per second, is very different if one application is doing sequential I/Os than if several applications are generating random I/Os. If you only used the maximum throughput measured for the sequential I/O case (as a value of the I/O bandwidth available for the device) to compute the percentage of utilization of the device under random I/Os, you might be misled to think that the device is at 20% utilization, when it is in fact at 100% utilization.

To get more accurate and reliable percentages of per-class disk utilization, WLM uses the statistics provided by the disk drivers (displayed using the iostat command) giving for each disk device the percentage of time that the device has been busy during the last second. WLM counts the number of total blocks which have been read or written on a device during the last second by all the classes accessing the device, and how many blocks have been read or written by each class and what was the percentage of utilization of the device. WLM then calculates the percentage of the disk throughput consumed by each class.

For instance, if the total number of blocks read or written during the last second was 1000 and the device had been 70% busy, this means that a class reading or writing 100 blocks used 7% of the disk bandwidth. Similarly to the CPU time (another renewable resource), the values used by WLM for its disk I/O regulation are also a decayed average over a few seconds of these per second percentages.

For the disk I/O resource, the shares and limits apply to each disk device individually accessed by the class. The regulation is done independently for each device. This means that a class could be over its entitlement on one device, and the I/Os to this device will be regulated while it is under its entitlement on another disk and the I/Os to this other device will not be constrained.

WLM supports accounting and regulation of resources on the basis of total consumption. There are two types of resources that can be regulated in this manner: class totals and process totals.

class totals: Per-class limits can be specified for the number of processes, threads, and login sessions in the class. These limits are specified as the absolute number of each resource that can exist in the class at any instant. These limits are strictly enforced; when a class has reached its limit for one of these resources, any attempt to create another instance of the resource will fail. The operation will continue to fail for any process in the class until the class is below its specified limit for the resource.
process totals: Per-process limits can be specified for the total amount of CPU time, blocks of disk I/O, and connect time for a login session. These limits are specified at the class level, but apply to each process in the class individually (each process can consume this amount). These consumption values are cumulative and thus represent the total amount of each particular resource that has been consumed by the process during its lifetime. Once a process has exceeded its total limit for any resource, the process will be terminated. The process will be sent a SIGTERM signal, and if it catches this signal and does not exit after a 5 second grace period, will be sent a SIGKILL signal. When a login session has reached 90% of its connect time limit, a warning message will be written to the controlling terminal to warn the user that the session will soon be terminated.