Managing queuing network model scenarios

To create QNM scenario and run the prediction

1. Navigate to Workspace > Domains, Services & Applications > Domain > Works.
2. In the Works folder, select the queuing network model for which you want to add a scenario. The model details are displayed.
3. In the Prediction Scenarios table, click Add scenario. A new scenario is created automatically using the default options and the scenario details are displayed. If required, you can edit the scenario. 
   QN scenario - details page before the prediction is run
   
4. In the scenario details page, enter the required parameters to perform the prediction.
   • In the Business Drivers table, update the business driver metric values that you want to estimate:
     • Current value: The latest value that is available for the business driver metrics. By default, the average value of the business driver is displayed as a current value. You can change this statistic and time filter by using the Estimate current values option. 
     • Target growth factor: The expected increase in the value. It is the ratio between the target and the current values. By default, the growth factor of 10 is applied.
     • Target value: The estimated values based on the growth factor. It is calculated as the current value multiplied by the growth factor. By default, it is set to 10 times the current value.
   • In the Systems table, perform a what-if analysis to simulate scenarios, such as hardware and system changes, and find the appropriate configuration that suits your requirements. You can choose the multiplicity of every existing system/resource and calculate its speed variation, based on the built-in benchmark data.
     • Current system: Displays the current system details.
     • What-if system: When you select a different vendor and hardware model while calculating a speed variation, the selected system is displayed as a what-if system. For more information, see Benchmark comparison.
     • Current multiplicity: For each system, it displays the multiplicity value for the number of servers that currently makes up the system. It is defined by the SYS_MULTIPLICITY property in the system's configuration table.
     • What-if multiplicity: A different number of servers that will make up the system. Perform a what-if analysis by entering a value different than Current multiplicity. 
     • Current quantity: For each resource of the system, it displays the number of parallel processing units available for the resource. For example, the current number of cores for the CPU resource. In the case of memory, it is fixed to 1 as the memory bus is unique across an entire system. 
     • What-if quantity: For each resource, specify a value different than the Current quantity.
     • Scenario speed-up: For each resource, perform a scenario speed variation. Click the Calc option to open the Compute speedup page. For more information, see Benchmark comparison.
5. Click Run prediction to run the model in the configured scenario.

Benchmark comparison

While performing what-if analyses, you might want to estimate what would happen if a server was substituted with a different one.

The Compute speed-up page enables you to define a different hardware configuration.

• The Compute speed-up page displays the details of the current system, such as description, vendor, and cores number. 
• You can select a vendor and hardware model for the what-if system. A description of your what-if system is displayed to help you in your selection.
• You can also choose whether you want to normalize the benchmark results by taking into account the number of CPU cores.

Click Save to confirm your changes and return to the scenario detail page. The Scenario speed-up percentage value and What-if system in the Systems table is automatically updated with the result of the Compute speedup settings.

Calculating speedup through benchmark data

To view the prediction results

After the prediction is run, BMC Helix Continuous Optimization updates the Business drivers and Systems tables with the estimated results, and displays charts based on the queue length, response time, throughput and resource utilization.

QN scenario - details page after the prediction is run

For every metric, three estimations that were previously marked as N/A are now available in the Business drivers table, indicated in red or green):

• Supported value: The maximum value of the business driver metric before a threshold is violated.
• Supported growth factor: By how much can the current business driver increase before a threshold is violated.
• Residual capacity: The actual unused business driver capacity percentage, where the full capacity (100%) is represented by the Supported value. For example, if the current value is 2000 webpage-hits/hour and the predicted supported value is 8000, then the residual capacity is 75%.

The colored semaphores are also displayed in the last column of the Business drivers and Systems tables.

Semaphores in the Systems table can be:

• : The given resource has violated the utilization threshold and is a bottleneck.
• : No thresholds have been violated.

The following scenarios can arise in the Business drivers table:

• If there is at least one red semaphore in the Systems table, the colors of the semaphores in the Business driver table will be .
• If all semaphores in the Systems table are , then semaphores in the Business driver table can be:
  • All 
  • One or more  semaphores, meaning that the business driver violated the response time.

The prediction summary in the form of graphs (chart and data table) is also displayed at the end of the page. The chart in the graph displays two different colored areas:

• White: Represents normal load.
• Red: Signals the threshold violation zone. The vertical red line between the two areas corresponds to the specified threshold.

To edit QNM scenario

1. In the details page of the scenario that you want to edit, click Edit. Alternatively, you can edit the scenario by clicking the  icon in the Prediction Scenarios table.
   The Edit scenario page is displayed.
2. Edit the following parameters, as required:
   • Name: Name of the scenario.
   • Description: Brief description of the scenario.
   • Time resolution: Choose the preferred time resolution, for example, "day".
   • Utilization threshold: Set the utilization thresholds that you wish to use in the model; can be global, customized for each system, or disabled.
   • Response time threshold: Specify the threshold that you wish to define for the estimated response time, in seconds.
   • Aggregation mode: Choose if you wish to use aggregated or detailed performance indices for the forecast.
   • Update policy
     • Manual: The prediction scenario must be executed manually.
     • Automatic on page access: The scenario will be automatically updated when accessing its detail page.
     • Automatic on related analysis changes: The scenario will be automatically updated if a system (or business driver) is added to the domain used by the analysis, or any changes are made to the analysis.
3. Click Save to confirm your changes.

Every time a prediction is run or a scenario is edited, BMC Helix Continuous Optimization deletes the previous results.