-
Notifications
You must be signed in to change notification settings - Fork 6
run off the crud
RUN off the CRUD is an alternative to the traditional CRUD approach to defining basic operations on resources in data management systems (e.g., databases, APIs). RUN stands for Read, Update, Nullify and bears a nuanced approach to deletion.
Why is it preferred ('run off') over the CRUD (Create, Update, Delete)?
Consider the need to protect 'authentic data' in a decentralized environment.
In a decentralized control model, the data always originates from a controller (aka client). The data created (sourced) by the controller follows the principle of 'Non-Interactive Replay Monotonicity' to be able to protect the data from a replay (events are changed) or a deletion (some events are deleted) attacks. That is to say, the data (or events comprising it) is never deleted, it's rather always added to via updates. Each update, therefore, forms a verifiable, continuous log ( e.g. by providing growing sequence number, date timestamp, etc for each update). To enable invalidation of data, a special update, called Nullify, is used.
The client, therefore, updates the server (it's peer or peers), which just maintains the log following certain rules (see BADA - Best Available Data Acceptance).
To summarise, the server can only Read the log, add Updates to it, including Nullifying ones. So no Create or Delete.