ADT dataset is an essential data source designed to include the device and command data (including device configuration files and CLI commands). ADT dataset, embedded in ADT, can be repetitively used and shared with users in the same domain.

In ADT, four methods of building ADT Dataset are introduced to add device configuration files and CLI commands as links into dataset column, serving as data source for intent creation, intent decoding, intent replication, and intent execution. The following diagram illustrates the workflow of Dataset. 

Why do you need this feature?

ADT dataset is developed to achieve the following goals:

• ADT dataset and Benchmark data can be repetitively used for building and debugging intent-base automation even when live network is not accessible or is only accessible in a certain period, increasing the scalability. This will benefit the following workflow:

• Continuous network assessment with network-wide observability dashboard (auto-updating).
• Build baseline for the entire network.
• Provide scalability solution for network auto-diagnosis via TAF to achieve network-wide monitoring while reducing network load due to high-frequency network data retrieval. In TAF, a ticket may trigger live network diagnosis and historical data diagnosis (e.g., analysis of yesterday data, last known-good data or latest baseline configuration). In this case, ADT Dataset can serve as data source for TAF to provide historical data. This is a recommended TAF configuration to support more TAF calls.

• By using ADT Dataset, you can not only analyze current network status via automation feature such as auto intent, but also analyze the network status based on history data or known-good data. For example, you can view the diagnosis results of 'known-good' application path.
• For detecting network change, the previous network data can be stored in dataset, and then be compared with the data obtained later to determine whether network change occurs.

Use Case

Case 1: When customers do not have network access or only have network access in certain period, they can use dataset for intent replication by completing the following major two steps.

1. Perform intent replication by using dataset without live network access.
2. Debug intent replication with dataset: Retrieve data from live network once and save the data into a dataset, then debug intent replication with the dataset. In this way, customers don’t need to repetitively retrieve the same data from live network.

Case 2: Customers can build network baseline with dataset as data source to verify network changes. Specifically, customer will make network change on weekends and wants to detect unexpected Config drift after network change by running intents. To achieve this, customers can follow the steps below:

1. Set up a benchmark task to get network device configuration on every Friday and save the data into a dataset.
2. Use the data to update intent baseline.
3. Run these intents to compare data from live config files with the baseline after a network change executed on Sunday.

Main User Flow

• For the consultants who do not have network access, the workflow includes the following main steps:

1. Consultants provide useful CLI command files to NetBrain Engineers.  
2. Import the CLI command files to the system via Import File Wizard, then create ADT dataset using the imported files as data source.
3. Carry out intent replication using the ADT Dataset in step b as data source.
4. Run intents using the same ADT dataset.

• For consultants who only have network access in certain time period, they can work on cached data with the following main steps:

1. Create an ADT and decode intent to create the intents by NIT when network is available.
2. Create ADT dataset using the manually retrieved data by selected intent column of the ADT.
3. Manually retrieve the command data in the ADT dataset when network can be accessed.
4. Run intents via dataset when they do Not have live network access.

 

See Also:

• Automation Data Table
• Create and Run Created Intent