The specific terms used in message queueing services are defined as follows:
Name | Definitions |
---|---|
topic | A topic can be interpreted as an actual queue. A certain cache space is set for each topic, and there are completely isolated data between different topics. |
Instance | Services and charging are based on entities that can provide independent services under a user account. You can have multiple topics in one instance. |
sub (subscription) | Sub can be interpreted as an actual broker which sends data consumed to client. Multiple brokers are supported by a queue, which are mutually independent and can consume messages from the queue in different modes. |
Message | An independent datagram queued is called a message. |
Message production | An act of sending messages to queue |
Producer | An object that sends messages to a queue |
Consumption | An act of retrieving messages from queue |
Consumer | An object that reads messages from queue |
Message ID | A unique number assigned to queued messages, which actually represents the first byte offset of message data in the queue |
Consumer position | The first byte offset of data in the next message read by the consumer from the queue |
Cache usage for messages | The number of messages calculated by the server per unit of 1k, based on the data traffic produced/consumed by the queue |
Clustering | Sub clients who subscribe to the same topic are considered to be in a cluster group, and the server will send messages to every client in the cluster on average. |