Tokenization

See Tokenization for more information about this algorithm framework.

Creating a tokenization algorithm via API

1. Retrieve the frameworkIdfor the Tokenization Framework. This information can be retrieved using the following endpoint:

   CODE

   algorithm   GET /algorithm/frameworks

   The framework information should look similar to the following:

   CODE

   {
       "frameworkId": 13,
       "frameworkName": "Tokenization",
       "frameworkType": "STRING",
       "plugin": {
           "pluginId": 7,
           "pluginName": "dlpx-core",
           "pluginAuthor": "Delphix Engineering",
           "pluginType": "EXTENDED_ALGORITHM"
       }
   }

2. Create a Tokenization algorithm instance via the following endpoint:

   ^{algorithm POST /algorithms}

   Configure a new algorithm using the JSON formatted input similar to the following:

   CODE

   {
       "algorithmName": "exampleTokenization",
       "algorithmType": "COMPONENT",
       "frameworkId": 13,
       "algorithmExtension": {
           "ivLength": 16,
           "fallback": "CHARACTER_MAPPING",
           "cmCharacterGroups": [
               "[A-Za-z0-9+/]"
           ],
           "cmMinMaskedPositions": 1
       }
   }

Tokenization algorithm extension

• ivLength(default=16, minimum=0, maximum=16)

Integer The length of the initialization vector (IV) used for AES in CBC-CTS mode. The default length is 16, which offers the most security. The tradeoff is that this increases the length of the masked result. Selecting a lower IV length decreases the length of the masked result. It is recommended that you only select an IV length of 0 if you require the masked value for each input to be consistent between jobs and for the same input to only mask to one output.

• fallback(required, no default)

String This specifies how to handle masking a value where the encrypted result does not fit in the column size. If an AES encrypted result is too long to fit into the field, there are two fallback options: - NONE - the job fails if the masked result is too long - CHARACTER_MAPPING - the Character Mapping algorithm is used to tokenize the value, which produces a result that is the same length as the input

Extension for Character Mapping fallback

• cmCharacterGroups(default=["[A-Za-z0-9+/]"])

Array of Strings A list of String values defining the characters to be masked. Each group must be either: - a Java regex style character group beginning with '[' - a String of the literal characters that comprise the group.

Duplication of characters within or among groups is not permitted.

• cmMinMaskedPositions(default=1, minimum=0)

Integer The minimum number of positions that must be replaced for masking to be considered successful. Non-conformant data handling is triggered whenever fewer positions are masked. Inputs containing only whitespace never trigger non-conformant data handling.