Whiteflag API Protocol State

The API has to keep its state of the Whiteflag protocol. This inludes keeping track of its blockchain status and accounts, known originators, encryption keys etc. The state is implemented as an in-memory object which is managed by lib/protocol/state.js. The structure of the state is defined in lib/protocol/static/state.schema.json.

State configuration

All state configuration parameters are in the [state] section of the config/whiteflag.toml configuration file.

State preservation

To preserve the state across restarts of the API, the state object is stored in the primary datastore. When the API starts and initialises the state, the state is restored from the primary datastore. If no state is found in the datastore, the API starts with an empty state.

State encryption

The state contains sensitive data on which the security of the Whiteflag data depends. Therefore, the state can and should be encrypted. A Master Encryption Key (MEK) is used for the encryption of the state. This key is provided with the masterKey parameter in the configuration file.

All secrets and keys that are stored in the state are always encrypted with a unique Key Encryption Key (KEK) which is derived from the MEK, using a HKDF function with a salt and the key identifier. The key identifier is a hash of key specific attributes such as blockchain account and originator address.

The state as a whole can (and should) also be encrypted when stored in a file or in the primary datastore if encryption parameter is set to true in the configuration file. The state will then be encrytpted with a Data Encryption Key (DEK) which is also derived from the MEK.

Currently it is not possible to change the MEK without losing the state. Also, please be aware that losing the MEK results in unobtainable keys, including private blockchain keys and pre-shared encryption keys!