|
|
Statistics collection in Release 10 of CWM is provided by the Statistics Collection Manager (SCM). SCM consists of a Java graphical user interface (GUI) and two types of back-end servers, a control server and a collection server. This chapter describes the operational features of the SCM.
The SCM control server communicates with the Java GUI and provides control of the statistics data collection. The SCM collection server performs the data collection from the various network nodes. One SCM control server provides the controls based on user input via the Java GUI. Multiple collection servers are distributed across a network. Each SCM collection server provides collected statistics to an associated SCM statistics parser which writes the data into the SCM statistics database.
Figure 1-1 shows the SCM system architecture and the basic communication paths for data and control within the SCM.
You use the SCM graphical user interface to control the statistics data collection and monitor the collection status. The SCM GUI is launched from the CWM Desktop along with the other CWM desktop applications. The SCM GUI can run on any workstation capable of running the CWM desktop GUI and connects to the SCM control server in the CWM workstation specified by the desktop. The SCM GUI is generic; it handles statistics collection independent of node hardware. The list of statistics data applicable for various types of cards and nodes is provided by the SCM control server.
The SCM control server provides the control for data collection based on your input via the SCM GUI. For example, you use the GUI to tell the SCM control server to enable or disable collection of various statistics and to start or stop statistical data collection.
The SCM control server runs of the CWM workstation and uses the CWM database for storing SCM metadata and the list of enabled statistics for all nodes. The SCM metadata provides the list of enabled statistics for various types of cards and nodes.
The SCM control server does not run in a primary or secondary role like the CWM workstation architecture. The SCM GUI can connect to any SCM control server. The primary or secondary status of the CWM workstation, however, is used during a CWM cold start to initialize data control.
The SCM collection server performs the data collection from the various nodes. The SCM collection server is distributed to various hosts and receives requests from the SCM control server to start and stop data collection from a node. After an SCM collection server begins collecting data, the collection server is independent of the SCM control server.
An SCM collection server acts as a primary, secondary, or tertiary data collector for a given node. The SCM collection server automatically provides a default primary, secondary, or tertiary SCM collection server when you enable statistics data collection from a node. The secondary and tertiary data collectors are optional, and you can change the collection server if desired.
The SCM collection server forwards the collected files to the SCM statistics database hosts via file transfer protocol (FTP). If the SCM collection server and the SCM statistics parser are both on the same CWM workstation, you can select a no-FTP option and the SCM statistics parser will process the file directly from the incoming directory of the SCM collection server.
The SCM collection server begins operation during the initialization process during workstation startup, if statistics collection has been enabled and started.
The SCM statistics parser parses the statistics data collected by the SCM collection server and writes the data to the CWM statistics database. One statistics parser is associated with each SCM collection server.
The SCM statistics parser begins operation during the initialization process during workstation startup, if statistics collection has been enabled and started. When you add a new SCM collection server via the SCM GUI, the SCM control server updates the SCM statistics configuration database. When a new SCM collection server is started, a SCM statistics parser automatically starts, too.
The SCM statistics database is a separate relational database from the CWM database used for storing statistics. For most installations, only one database is needed, but more than one can be used, if desired.
The SCM statistics ager removes old data from the database and file system.
Figure 1-2 shows the relationships among the SCM components. An SCM control server, which is the statistics master for a given node, manages an SCM collection server which is the primary statistics collector for that node. The SCM collection server works with a statistics parser which processes the data
Posted: Thu Sep 28 16:16:59 PDT 2000
Copyright 1989-2000©Cisco Systems Inc.