Literally, it's hard to understand. Let's give a classic example: someone withdraws money from an ATM machine, which is a use case in itself, but the level is relatively high and it is subdivided. A rough idea, there are several:
(1) Query balance.
(2) withdrawal.
(3) transfer.
(4) Deposit. These four points can be an independent use case, and the executors are all people. Simply put, a use case is a collection that describes the interaction between the executor and the system.
Extended data:
The event stream processing mode realizes real-time operation by integrating various mechanisms. First, inbound processing is supported, that is, the input event stream begins to flow through continuous queries as soon as it enters the system. As they flow, queries transform events and continuously give results, all of which are done in memory.
Read or write operations to disk storage are optional and are handled asynchronously in many cases. Input processing overcomes the limitations of traditional output processing used in traditional database management systems.
In the export process, data must be inserted into the database and indexed before any processing is started. By excluding disk storage from the critical path of processing, compared with traditional processing methods, the event flow processing mode has achieved obvious performance improvement.