Auto Calculating Golden Baseline

Besides Defining Golden Baseline Manually, the system provides an auto-calculation mechanism to set Golden Baseline for parser variables based on historical data generated through scheduled tasks. Once a Golden Baseline is manually modified with customized rules, its auto-calculation function will be disabled. To re-enable the dynamic calculation, you need to clear the manually defined rule first.

Note: Only the data retrieved through scheduled benchmark tasks, scheduled Qapp/Gapp tasks, or scheduled DVT/Parser tasks, can be involved in Golden Baseline dynamic calculation. The data generated by on-demand tasks will not be involved in Golden Baseline calculation.

 
This section introduces the following topics:

Enable/Disable Dynamic Calculation of Golden Baseline

Calculation Logic for Different Variable Types

Initial Value for Golden Baseline Calculation

Enable/Disable Dynamic Calculation

1.Click the Start Menu and select Golden Baseline Manager. On the Dynamic Calculation tab, click Enabled.

2.Click the Start Menu and select Variable Mapping. On the Global Variable tab, select the target global variable, and check Golden Baseline Dynamic Calculation. Then all its mapped parser variables will participate in dynamic calculation.

Tip: Dynamic Calculation of Golden Baseline for a specific parser variable can also be enabled/disabled individually during the definition of a data view template. See Variable Settings for more details.

Note: Only the dynamic calculation of Golden Baseline for single-value variables, table variables and interfaces’ table column variables can be enabled. To enable the dynamic calculation for interfaces’ table column variables, an interface key is required to be enabled.

Dynamic Calculation Logic for Different Variable Types

Golden Baseline for string/number/table-type variables can be dynamically calculated.

String-Type Variable:

The Golden Baseline for the following two string-type variables can be dynamically calculated.

Equals any of – at least requires data at 10 different time points for calculation, and the count of instance status must be no greater than 3.

Equals – at least requires data at 4 different time points for calculation, and the first three values must be exactly the same. See the following table for the example of IOS version.


1st Time

2nd Time

3rd Time

4th Time

5th Time

6th Time

7th Time

8th Time

Data (IOS Version)

12.1

12.1

12.1

12.1

12.2

12.2

12.2

12.2

Golden Baseline

Calculating

Calculating

Calculating

Equals 12.1

Equals 12.1

Equals 12.1

Equals 12.1

Equals 12.2

Alert

N/A

N/A

N/A

No Alerts

Alert Detected

Alert Detected

Alert Detected

No Alerts

 

Number-Type Variable:

The Golden Baseline for the following number-type variables can be dynamically calculated:

Available Type

Required Data Points

Equals

at least requires data at 4 different time points for calculation, and the first three values must be exactly the same.

Equals any of Range

at least requires data at 10 different time points for calculation.

No Change - Equals Last Value

Continuously Increase

Table-Type Variable:

According to the predefined compared columns, the system generates hash values in parsers, and convert into single-value strings for comparison and analysis. The analysis logic is the same as that for string-type variables.

Take the following table for example, if the “BGP Neighbor”, “Version” and “AS Number” columns are predefined as compared columns, then only the corresponding hash values for these three columns will be generated for comparison.

Compared Columns

Ignored Columns

BGP Neighbor

Version

AS Number

MsgRCD

MsgSent

InQ

10.100.1.1

4

200

26

22

0

10.200.1.1

4

200

51

23

0

Initial Value for Golden Baseline Calculation

The convergence of Golden Baseline dynamic calculation for range-type variables can be accelerated by specifying an initial value, for example, “0 – 90%”. With a specified initial value, its golden baseline can be gradually calculated with convergence ever since the first cycle of data retrieval.