<p class="p"><strong class="ph b">Flexibility and high efficiency</strong></p> <p class="p">Message Queue Kafka supports multiple message types without being bound to any products. It is 100% compatible with open-source communities. Service codes can be smoothly switched to Ping An Cloud Kafka without any modification.</p> <p class="p"><strong class="ph b">High scalability</strong></p> <p class="p"> As a distributed system, Kafka is easy to scale out. Message Queue supports hot expansion of the cluster scale. With distributed deployment of producers and consumers, the cluster supports zero-downtime scaling.</p> <p class="p"><strong class="ph b">High performance</strong></p> <p class="p">In the process of data publication and subscription, Kafka can guarantee high data throughput. It maintains stable performance even for the TB-level data storage.</p> <p class="p"><strong class="ph b">High reliability</strong></p> <p class="p">Kafka ensures high reliability with its partitioning, replication, and fault tolerance mechanisms.</p> <p class="p"><strong class="ph b">High security</strong></p> <p class="p">Message Queue Kafka provides standard AK/SK identity authentication and ACL permission control of Ping An Cloud. Both the tenant-level network isolation and account-level instance access protect the user data from being stolen or altered.</p> <p class="p"><strong class="ph b">Asynchronous communication</strong></p> <p class="p">The sender of a message can return immediately after sending the message to the message queue without waiting for the receiver's response. The message will be stored in the queue until it is fetched by the receiver. The sending and processing of messages are completely asynchronous.</p> <p class="p"><strong class="ph b">Process decoupling</strong></p> <p class="p">Message queues help reduce the coupling between two processes. As long as the message format remains unchanged, no changes are required on the sender even if the interface, location, or configuration of the receiver changes. The sender does not need to know the message receiver, which simplifies the system design. For processes that use the remote procedure call (PRC) or socket connection, one party must modify the request configuration if the interface, IP, or port of the other party changes.</p> <p class="p"><strong class="ph b">Message routing</strong></p> <p class="p">The sender does not need to establish a direct connection with the receiver. Message queues guarantee that messages can be routed from the sender to the receiver. Message routing is even available for services that are not easily interoperable on the network.</p> <p class="p"><strong class="ph b">Multi-terminals</strong></p> <p class="p">Multiple parts of the user's system can send or receive messages at the same time, and control the availability of messages based on the message status.</p>