Interfaces
Interfaces ingested by Riverbed IQ are displayed on the Interface Explorer page. This page allows you to search all the interfaces known to Riverbed IQ and to tag them using custom properties. You can also forecast Interface Utilization using Predictive AI.
Click Explorer > Interfaces to display the Interface search page, which lists all of the interfaces known to IQ.
Interfaces Page
The Interfaces page shows these columns for each interface:
-
Name — The name of the interface.
-
IP Address — The IP address of the interface.
-
ifindex — The Ifindex of the interface.
-
IF Alias — The IF Alias of the interface.
-
IF Description — The IF Description of the interface.
-
Type — The type of interface.
-
Location — The physical location of the interace.
-
Inbound Speed — The inbound speed of this interface.
-
Outbound Speed — The outbound speed of this inteface.
-
Custom Properties — The list of custom properties assigned to the interface.
-
Data Source Name — The name of the data source(s) that reported the interface.
Columns can be shown and hidden by clicking on the column chooser button. The column chooser button is located on the upper right corner of the table. Clicking on the column chooser will launch the column chooser dialog.
To add columns to the table, select columns in the list of Available Columns and click the Add button and they will move into the list of Selected Columns. To remove columns from the table, select columns in the list of Selected Columns and click the Remove button and they will move back into the list of Available Columns.
Searching For Interfaces
Interfaces may be filtered using the Filter section or by doing a text search. To do a text search just enter the search team in the search box and click Search. Notice that the number of search results are always displayed below the search box.
There are a large number of interface properties that can be used for filtering the interfaces. To add a filter, just expand the filter section and select the items you want to see in the search results and click Apply Filters.
Note that the filters that were selected appear below the search box. Not all filter facets show up by default. To add and remove filter facets, click on the ... next to the Filters label and select Customize list. This launches the Facets Chooser which looks very much like the Column Chooser.
To add facets to the filter bar, select facets in the list of Available Facets and click the Add button and they will move into the list of Selected Facets. To remove facets from the filter bar, select facets in the list of Selected Facets and click the Remove button and they will move back into the list of Available Facets.
Setting Custom Properties
Custom Properties can be assigned to an interface. These custom properties can then be used as filters when you search for interface and the custom properties can be displayed in the runbook tables. To set a custom property on an interface, click the Set Custom Property button in the toolbar.
This launches the Set Custom Property dialog. To select the custom property to tag the interfaces with select the name of the custom property in the Name dropdown. Select the value of the custom property in the Value dropdown. You can choose to apply the custom property to only the interfaces currently selected in the table or you can apply the custom property to all the interfaces in the current search results using the radio buttons. When you are done, click Apply.
Predictive AI
The Predictive AI - Forecast feature allows a user to select up to 10 interfaces and, through the use of Machine Learning algorithms, predict when those interfaces will have utilizations above a customizable threshold, making it ideal for use in capacity planning. To use the Predictive AI feature:
-
Navigate to the Explorer / Interfaces page:
-
Select up to 10 interfaces to run the forecast
-
Click on “AI Forecasting” to bring up the forecasting dialog:
Here, the user can choose how far into the future to run the prediction on all selected interfaces (1, 3 or 6 months), as well as have IQ highlight the point in time when the predicted utilization will cross the given threshold. The threshold must be a value between 0 and 100, inclusive.
The predicted value can be thought of as an average, or normal value, based on what the algorithm learns by analyzing the interface.
-
Click “Show Forecast” to submit a forecast request:
The AI Engine to examine up to a month’s worth of historical data for each interface, for both in- and out-utilization, learn each interface’s traffic patterns, and then predict what that traffic will look like 1, 3 or 6 months into the future.
The learning and prediction are calculated on the fly, so it may take several seconds for the results to be ready. Once they are, the selected interfaces will be summarized:
Here, the in- and out-utilizations are summarized for each selected interface, as well as the predicted percentage change (up or down) for the next 7 days and 30 days. If the expected high value of the prediction range exceeds the threshold, the number of days until that happens will be shown in the May Exceed column. If the expected value of the prediction exceeds the threshold, the number of days until that happens will be shown in the Likely to Exceed column.
It is possible that interfaces that are new, or are lightly used, do not have enough data to make accurate predictions as to their utilizations in the future. When this occurs, “View Details” will show a red icon, which the user can hover over for more information.
-
Click on the blue icon in “View Details” to render the historical data used to train the algorithm and the predicted values:
The historical data is shown in the yellow region, and the predictions, with the expected normal range, based on the natural variability of the data, are shown in the green region.
Once the summary table is closed, the forecast data will be lost, but can be regenerated by entering the desired time range and thresholds can clicking the “Show Forecast” button.