Collation and analysis of operational data
ERAMON’s Performance Management Module (EPM) collects data from different network components or servers, and their interfaces. The data is then being prepared graphically for information and analysis purposes. The aim is to improve performance and status assessments of devices, and to display any relevant information relating to these errors and their interfaces clearly. Thus EPM supports tasks such as the planning of WAN-connections or file server capacities by recording traffic flow and storage volume values, and reporting any violations of threshold levels predefined in ERAMON’s event system. This enables the reliable recording of short-term capacity patterns and an identification of long-term trends, which in turn allows expansion measures to be carried out in good time. In addition to this, the quality of services rendered may be monitored and documented. EPM provides a considerable extension of the possibilities to display operational data currently available in the market. Scalability to the granularity level required by the user allows better handling as well as more accurate analysis.
The polling of measurements is based on a ruleset that can be recorded granularly. Based on this ruleset, the relevant device is automatically polled, for example the rule for setting up a measurement for the “CPU Load” for all Cisco Catalyst 4506 switches. The information that is to be polled is collated by the satellites at predefined intervals, usually via SNMP. Many MIBs or example measurements have already been integrated in the NMS, these can be expanded at any time. A dedicated database for backup and archiving of the collected data is available.
Depending on the number of active satellites (SATs) installed, ERAMON’s scalable client/server architecture (SAT/CENTER) allows an almost infinite amount of polled values. The type of data to be monitored depends on the MIBs of the device manufacturers and spans from ATM Soft-PVCs to switch buffers. Depending on the SNMP version implemented on the operating system of the devices, 32-bit- or 64-bit-information can be polled. EPM is configured using templates which can be rolled out accordingly. Certain settings can still be adjusted after the rollout. The requested information is either displayed historically in a graph, or you will see the actual current value. It is also possible to specify threshold values for alerts, as well as minimum/maximum values to improve the clarity of the display.
The following functions are available when displaying measurements:
- 95% Rule
The 95% average is shown
- Trend Analysis
The trend analysis applies the exponential smoothing method with a smoothing factor of 0.3 to approximate the curve. The curve is then displayed using linear regression
- Forecast Value
The forecast extrapolates the values from the linear regression to the selected time period. The system uses data from the measurements during the displayed time period as the basis for these calculations.
in CSV, pdf or png format
- Free scaling via the mouseover
Using the EPM Baselining facility, a range of historical values can be applied to define baseline values. Any violations below and/or above the set limits calculated on the baseline values would then generate the necessary alerts. The baseline is the mean average of a certain number of historical values. The polled values are then individually compared to the baseline and you will be alerted of any violations of (upper or lower) limits, these are specified in percent.
- Dynamic Views
With the help of a search mask you can view any of the measurements together, such as all QoS statistics of a service class for several devices.
- IP SLA
Currently already pre-defined for: cisco, huawei and juniper
- Top-N Report
Manually via GUI or automated