Contents

1   Understanding Performance Management

1.1   Key Performance Management Concepts

Performance Management (PM) provides a management interface for performance measurement data handling in the Managed Element (ME) including the following:

• Collection and aggregation of performance measurement data
• Creation of performance measurement reports
• Reporting of performance threshold alarms

Performance measurement data is collected and recorded by the ME regularly. This activity is performed to verify the physical and logical configuration of the network and its functions, to understand how they are used, and to locate potential problems as early as possible. The performance measurement reports can be used by management systems to understand performance trends, to balance traffic load, and to identify problem areas.

The Performance Management managed area, Pm, can be found in the Managed Object Model (MOM). For general information about the MOM, Managed Object Classes (MOCs), cardinality, and related concepts, refer to Managed Object Model User Guide.

Performance threshold alarms can be used by management systems to identify problem areas needing immediate attention.

1.2   Measurement Types

The measurement types define what measurement data the ME can collect and what collection method and aggregation method are used. A measurement type supports one collection method and one aggregation method.

Table 1    Measurement Type Collection Methods

Collection Method	Description
Cumulative counter	An ME uses a cumulative counter for counting events. Each time the event occurs a value is added to the counter. The counter is reset to a well-defined value (usually 0) at the beginning of each Granularity Period (GP). The GP is the time between the initiation of two successive collections of measurement data.
Gauge	A gauge is used for representing a dynamic value that can change in either direction. It can be used for continuous measurements over many GPs, for example, representing a temperature. It can also be used to get a maximum or minimum value during a GP. In this case, the gauge is reset at the beginning of each GP.
Discrete Event Registration (DER)	Using DER, data related to a particular event is captured (same as a gauge). Every nth event is registered, where n can be 1 or larger. The value of n depends on the frequency of occurrence of the event being measured. DER measurements are reset at the beginning of each GP and only have a valid result at the end of the GP.
Status Inspection (SI)	An SI is supported in the same way as a gauge but cannot be used for continuous measurements over several GPs. Status inspection measurements are reset at the beginning of the GP and only have a valid result at the end of the GP.

Table 2    Measurement Type Aggregation Methods

Aggregation Method	Description
Sum	All measured values are added during a GP.
Average	An average value is calculated using all values collected during a GP. For a gauge, an average value can be calculated over several GPs.
Minimum	The minimum value during a GP is collected.
Maximum	The maximum value during a GP is collected.
Last update	The last value during a GP is collected.

Measurement types are categorized and logically grouped into PM groups.

1.3   PM Groups

A PM group is a grouping of the measurements into logical grouping. Grouping is done according to what the measurements are characterizing. For example, a Traffic Counters group would contain measurement types which quantify the traffic on an interface (#requestsReceived, #requestsRejected, #failedConnections, #errors). A PM group corresponds to NBI CM MOC if CC has OAM model.

1.4   Measurement Jobs

The execution of measurement data collection, aggregation, and reporting is performed by measurement jobs. A measurement job applies to one or more measurement types or to an entire PM group. A measurement job executes the measurement types respective collection method and aggregation method at regular time intervals as defined by its GP.

The supported GP values are 1 minute, 5 minutes, 15 minutes, 30 minutes, 1 h, 12 h, and 24 h. The default value for GP is 15 minutes.

The Measurement reader defines the measurement being counted in a particular measurement job and ties it to the specific measurement types or a PM group.

A measurement job collects measurement data and generates reports as performance measurement XML files.

A measurement job can have three different priorities; low, medium, or high. The measurement jobs with the highest priority have their reports output first or its threshold condition checked first.

1.5   Reported Data / Performance Measurement XML Files

Performance measurement XML files are compliant with 3GPP® TS 32.435. The files are of type "A", which means that the file contains results of measurements collected from a single ME during a single GP, according to the 3GPP TS 32.432 V10.0.0 standard. The performance measurement XML files also handle multi-value counters, which are an addition to the 3GPP standard. For more information, refer to Performance Management Report File Format.

The performance XML file can optionally be compressed using GZIP to reduce the amount of disk space used.

A new performance measurement XML file is created after regular aggregation intervals at the end of each GP.

The collection of performance measurement data is started when the time is 0:00. Four jobs are set up. Two (job 1 and 2) have a GP of 5 minutes and the other two (job 3 and 4) have a GP of 15 minutes.

Table 3    Example of Generation of Files

Time	Number of Generated Files	Content
Between 0:00 and 0:05 (a)	0	No file is generated, measurement time collection has started at 0:00, and the GP has not finished.
Between 0:05 and 0:10 (b)	1	Measurement results for jobs 1 and 2 during period (a)
Between 0:10 and 0:15 (c)	1	Measurement results for jobs 1 and 2 during period (b)
Between 0:15 and 0:20 (d)	2	Measurement results for jobs 1 and 2 during period (c)
		Measurement results for jobs 3 and 4 during period (a) + (b) + (c)
Between 0:20 and 0:25 (e)	1	Measurement results for jobs 1 and 2 during period (d)

The creation of a performance measurement XML file after the end of a GP can take up to 5 minutes. After this time, the absence of an expected performance measurement XML file or the absence of measurement data for an active performance job is to be considered as a potential fault situation.

This default performance file generation can be modified using the concept of job groups. A job group is defined using a simple text string associated with each PM job. The PM jobs with a common string name are all to be written into a performance management XML file prefixed by that string name, creating a group of PM jobs.

The created performance measurement XML files are exposed in the logical file system under MO FileGroup= PerformanceManagementReportFiles and can be transferred to external systems as any files in the logical file system. For more information, refer to Handling Files.

File Management preventive maintenance policies is not needed to performance measurement XML files. PM provides instead its own internal preventive maintenance policy. Up to 1000 files can coexist in the ME. When this limit is reached, the ME always deletes the oldest file before creating a new one.

1.6   Threshold Alarms

Threshold alarms can be used along with the gauge and cumulative counter collection methods. A threshold monitoring job monitors measurement data and raises an alarm when a threshold is crossed, but do not collect measurement data. A threshold monitoring job raises or clears an alarm based on a comparison of the counter value with the configured high and low threshold values.

At the end of the GP, if the counter value is higher than the high threshold value and no alarm is active, then an alarm is raised. However, if the counter value is lower than the low threshold value and if an alarm is active, then the alarm is cleared.

If two or more threshold monitoring jobs are associated with a measurement type, then a change alarm can be raised to indicate that the counter value has crossed a higher or a lower threshold than previously shown in any earlier alarm.

1.7   Used Interfaces

PM uses the following:

• NETCONF or the Ericsson Command-Line Interface (ECLI) for configuration
• A file interface for performance measurements
• The Ericsson SNMP Alarm Management Information Base (MIB) interface for performance threshold alarms

2   Basic Performance Management Procedures

PM is accessed using NETCONF or the ECLI to manipulate the Management Information Base (MIB).

• List measurement types, PM groups, and PM jobs

  The list of measurement types and PM groups informs the user about what measurement data it is possible to collect on the ME. The list of PM jobs informs the user about what PM jobs can be started or are already running on the ME. The procedures in List Performance Management Groups and Measurement Types and List Performance Management Jobs provide further details on how to perform these operations.

• Create, start, stop, and delete measurement collection jobs

  A new performance measurement collection job can be created by the user to collect performance data according to supported measurement types that are currently not collected on the ME. As part of maintenance situations, the user can suspend a measurement collection job to lower the resource use in the ME and start it again. The procedures in Create Measurement Collection Job, Start Measurement Collection Job, Stop Measurement Collection Job, and Delete Measurement Collection Job provide further details on how to perform these operations.

  Note:  

  The automated suspension of jobs in, for example, overload situations, is not supported.

  
  

  A performance measurement collection job that is no longer of interest because of, for example, a new network performance monitoring strategy can be deleted by the user.

• Create, start, stop, and delete threshold monitoring jobs

  A new threshold monitoring job can be created by the user to raise threshold alarms on supported measurement types that are currently not monitored through alarms. As part of maintenance situations, the user can stop a threshold monitoring job and start it again. The procedures in Create Threshold Monitoring Job, Start Threshold Monitoring Job, Stop Threshold Monitoring Job, and Delete Threshold Monitoring Job provide further details on how to perform these operations.

  A threshold monitoring job that is no longer of interest because of, for example, a new network performance monitoring strategy can be deleted by the user.