Glossary

That utilization value in the selected time interval which 10% of the utilization values fall above, and 90% fall below or are equal to.

The rate at which utilization of a resource is projected to change.

The sum of all the utilization values divided by the number of data points for the selected time interval.

A compute cabinet that can be used as the only compute cabinet in a complex, or as half of a dual compute cabinet complex. A base cabinet is always physically the left cabinet in the pair (when viewed from the front) and is always the cabinet that contains the Service Processor.

The time interval , such as a week or a month, considered when examining trends.

The physical enclosure that contains cells or I/O chassis. A cabinet also includes hardware that provides power and cooling. Some cell-based servers support cabling several cabinets together to form a single complex.

The analysis and planning of workloads on a system or across a set of systems.

Cell controller. A chip located on every cell board that has interfaces to the processors and memory on the cell. The cell controller also has an interface to a system bus adapter and to the fabric. The cell controller maintains data coherency across the cells in an nPartition.

A circuit board that contains processors and memory, all controlled by a cell controller (CC). A cell is the basic building block of an nPartition in a complex.

A server in which all processors and memory are contained in cells, each of which can be assigned for exclusive use by an nPartition. Each nPartition runs its own instance of an operating system.

Central management server. A system in the management domain that executes the HP Systems Insight Manager software. All central operations within HP Systems Insight Manager are initiated from this system.

An nPartition, virtual partition, virtual machine, or resource partition whose resources are allocated by gWLM.

Multiple compartments are grouped to form a shared resource domain (SRD). The compartments all share the resources of the SRD. A compartment can be in only one SRD. Each compartment holds a workload. gWLM manages each workload's resource allocation by adjusting the resource allocation of its compartment.

A complex includes one or more cabinets that are cabled together and all of the hardware resources that they contain. A complex has a single Service Processor.

Any cabinet containing cells. An I/O expansion cabinet is not a compute cabinet.

The actual data-processing engine within a processor. A single processor might have multiple cores.

A set of resource-demand readings made at regular intervals for some period of time. The demand profile of a workload, system, or complex is used when doing capacity planning. Demand profiles can be based on historical data or computed as part of a forecast.

A specially configured compute cabinet that can be connected to a base cabinet to create a dual-compute-cabinet complex. The expansion cabinet is always the right-hand cabinet in the pair (when viewed from the front) and contains a hub to connect it to the Service Processor in the base cabinet.

Within a complex, the interconnect composed of cross-bar chips (XBC) and cells.

A prediction of system utilizations and workload demand profiles for some future time.

A time interval specifying the historical data to use for generating a forecast.

A combination of a forecast data range and a set of annual projected growth rates that is used to estimate future utilization.

Fair-Share Scheduler group. A group of processes that has its processor allocation managed by the HP-UX FSS service. FSS groups allow you to allocate fractions of processor resources, rather than only whole processors, to the processes in the group.

HP Integrity Essentials Global Workload Manager. The VSE Management Software application that allows you to centrally define resource-sharing policies that you can use across multiple HP servers. These policies increase system utilization and facilitate controlled sharing of system resources. gWLM's monitoring abilities provide both real-time and historical monitoring of the resource allocation.

A PCI or PCI-X card cage and associated backplane that contains a system bus adapter and one or more local bus adapters. An I/O chassis may or may not be physically removable.

Data that has been marked as invalid or invalidated by a Capacity Advisor user. Such data is not used in analysis.

I/O expansion cabinet. A cabinet that contains I/O devices (card cages) but no cells.

Local bus adapter. A device that connects the system bus adapter (SBA) to an I/O bus, such as PCI. Multiple LBAs are connected to a single SBA.

A server or other system that can be managed by SIM from a CMS . A managed system can be managed by more than one CMS.

Maximum 15–minute sustained: this is the highest value in the selected time interval that was sustained for at least 15 minutes.

A specific measurement that defines a performance characteristic.

In Capacity Advisor, a combination of the statistical model (such as peak or average) used to calculate the metric and whether it is to be presented as a percentage or an absolute value.

Data that was not collected by Capacity Advisor, perhaps because a monitored system was down during data collection. Such data is not used in analysis.

A partition in a cell-based server that consists of one or more cells, and one or more I/O chassis. Each nPartition operates independently of other nPartitions and either runs a single instance of an operating system or is further divided into virtual partitions.

nPartitions can be used as compartments managed by gWLM as long as several requirements are met. Refer to the gWLM online help for a description of nPartition requirements.

Peripheral component interconnect. A standard for the connection between a processor and attached devices.

Peripheral component interconnect extended. An enhanced version of PCI.

The highest utilization value in the selected time interval.

A collection of rules and settings that control workload resources. For example, a policy can indicate the minimum and maximum amount of processor resources allowed for a workload, and a target to be achieved.

A single policy can be associated with multiple workloads.

The hardware component that plugs into a processor socket. Processors can contain more than one core.

The Profile Viewer provides a visual display of historical utilization data collected by Capacity Advisor along with additional information you have provided. The Profile Viewer also enables you to examine different time intervals and different categories of data.

A collection of processors grouped together for exclusive access by applications assigned to that processor set. Each application runs only on processors in its assigned processor set. On Linux systems gWLM simulates PSETs by using processor affinity masks.

A combination of qualitative and quantitative factors such as up time, response time, and available bandwidth, that collectively to describe how well a system performs. The Quality of Service is frequently embodied in a Service Level Agreement or in a set of Service Level Objectives between or among organizations.

A subset of the resources available to an operating system instance, isolated for use by specific processes. A resource partition has its own process scheduler. CPU resources in the partition may be allocated using Fair-Share Scheduler groups or processor sets. Policies for controlling the allocation of resources to the partition may be set using Global Workload Manager (gWLM).

System bus adapter. The chip in an I/O chassis that provides a connection between the cell controller on a cell and the set of local bus adapters in the I/O chassis.

A possible configuration of systems and workloads under consideration when doing capacity planning.

An independent support processor for HP servers that support nPartitions. The Service Processor provides a menu of service-level commands, plus commands to reset and reboot nPartitions and configure various parameters.

The Service Processor in HP servers is sometimes called the Management Processor (MP) or the Guardian Service Processor (GSP).

HP Systems Insight Manager. The platform and framework on which the VSE Management Software products are deployed.

For Capacity Avisor, a combination of a duration and a starting or ending point that defines the period of time over which the simulation is to be done.

Shared resource domain. A collection of compartments that share system resources. The compartments can be nPartitions, virtual partitions, virtual machines, processor sets (PSETs), or Fair-Share Scheduler (FSS) groups.

A server containing nPartitions can be an SRD as long as nPartition requirements are met. These requirements are detailed in the gWLM online help topic Getting the most out of gWLM.

A server or an nPartition divided into virtual partitions can be an SRD for its virtual partition compartments. A VM Host can be an SRD to its virtual machines. Similarly, a server, an nPartition, or a virtual partition containing PSETs can be an SRD for its PSET compartments. Finally, a server, an nPartition, or a virtual partition containing FSS groups can be an SRD for its FSS-group compartments.

A complex with nPartitions can hold multiple SRDs. For example, if the complex is divided into nPartitions named Par1 and Par2, Par1's compartments could be virtual partitions, while Par2's compartments are PSETs.

Hardware that can run one or more operating systems but does not support dividing hardware resources into nPartitions.

The WBEM services provider for real-time utilization data from managed systems.

A software entity provided by HP Integrity Virtual Machines. This technology allows a single server or nPartition to act as a VM Host for multiple individual virtual machines, each running its own instance of an operating system (referred to as a guest OS). Virtual machines are servers in the Virtual Server Environment (VSE).

A software partition of a server, or of a single nPartition, where each virtual partition can run its own instance of an operating system. A virtual partition cannot span an nPartition boundary.

An HP Integrity server running HP-UX with the HP Integrity Virtual Machines software installed. Virtual machines are manifested as processes executing on the VM Host. Configuration, management, and monitoring of virtual machines is performed on the VM Host.

An HP software product that provides virtual partitions.

The program that manages the assignment of resources to virtual partitions in a vPars-enabled system. To enable virtual partitions, the vPars monitor must be booted in place of a normal HP-UX kernel. Each virtual partition running under the monitor then boots its own HP-UX kernel.

The vPars monitor reads and updates the vPars partition database, boots virtual partitions and their kernels, and emulates certain firmware calls.

The database that contains the configuration information for all the virtual partitions on a vPars-enabled system.

The HP Virtual Server Environment (VSE) is an integrated virtualization offering for HP-UX servers, providing a flexible computing environment that maximizes utilization of server resources. VSE consists of a pool of dynamically sizable virtual servers, each of which can grow and shrink based on service-level objectives and business priorities.

Web-Based Enterprise Management. A set of Web-based information services standards developed by the Distributed Management Task Force, Inc. A WBEM provider offers access to a resource. WBEM clients send requests to providers to get information about and access to the registered resources.

The collection of processes in a standalone server, nPartition compartment, virtual partition compartment, or virtual machine compartment. Global Workload Manager (gWLM) extends this concept to include processor set (PSET) compartments and FSS group compartments. Global Workload Manager enables you to monitor and manage workloads by automatically adjusting the resource allocations of their compartments based on policies.

Cross-bar chip. On some server models each cell in a compute cabinet plugs into a cross-bar backplane by means of a pair of connectors, thereby forming a connection between the cell controller on the cell board and a cross-bar chip. On other server models, cell controllers are directly connected to other cell controllers, thereby eliminating the need for a cross-bar backplane.