The advantages of DMS RocketMQ are shown as follows:
Accumulated Concurrency
Hundred millions of messages can be accumulated with little impact on queue performance. You can increase the number of queues to improve the concurrency capabilities of the system.
DMS RocketMQ performs well in processing accumulated messages and guarantees its performance through the following features:
l Message storage mechanism: RocketMQ uses a disk-based message storage mechanism to support the persistence of a large number of messages on disks. This way, messages will not be lost.
l Message sending speed: RocketMQ features high throughput and low latency and can quickly process a large number of message sending requests.
l Message consumption speed: RocketMQ supports scale-out and load balancing to achieve high-concurrency message consumption.
l Storage capacity: RocketMQ uses disks to store messages and has a large storage capacity. You can add disks to expand the storage capacity.
DMS RocketMQ features high concurrency that is supported by the following aspects:
l Cluster mode: RocketMQ deploys message queues on clusters. You can add brokers and consumers to improve the system's concurrent processing capability. Each broker and consumer can process separate message flows for concurrent processing.
l Partition mode: RocketMQ divides messages into multiple partitions, and each partition can be processed by a different broker. In partition mode, you can distribute messages to different nodes for load balancing, which improves concurrent processing.
l Asynchronous sending mode: RocketMQ sends messages asynchronously. The sender will not wait for the ack after sending the messages to the broker and can process other services immediately. This method improves the concurrency of message sending, but may lead to message loss.
l Number of consumer threads: You can increase the number of consumer threads to improve concurrent processing. Each thread independently pulls and processes messages from brokers. Note: Excessive threads may have a negative impact on the system performance and shall be adjusted according to the actual situations.
Flexibility and Timeliness
The queue processing capacity can be automatically scaled on demand and messages can be delivered in milliseconds, guaranteeing instant messaging.
DMS RocketMQ is highly flexible and can meet various business needs. Advantages:
l Multiple message models: RocketMQ supports multiple message models, including the message queue model and the publish/subscribe model. In the message queue model, the message sender sends the messages to a queue, from which the message receiver can read the messages. This model applies to scenarios such as ordered and transactional messages. In the publish/subscribe model, the message sender publishes messages to a topic, and all consumers who have subscribed to that topic will receive the messages. This model applies to scenarios such as real-time notification and data distribution.
l Flexible message filtering mechanism: RocketMQ filters the attributes of messages to deliver satisfactory messages to consumers. This way, it achieves dynamic routing and selective consumption of messages and thus improves the flexibility and efficiency of the system.
l Delayed sending and scheduled consumption: RocketMQ allows you to specify a delay time during which messages will not be received by consumers and can only be sent to consumers after this time elapses. This is useful for scheduled tasks and delayed processing.
l Good scalability: RocketMQ uses a distributed architecture and supports master-slave replication and message partitioning. You can dynamically add producers, consumers, and message storage nodes according to your business needs for large-scale message processing and high concurrency access.
High Reliability
Cluster nodes work in the master-slave mode and support automatic master-slave failover upon failures, message persistence, multi-replica redundancy and auto deletion of message data.
The high reliability of DMS RocketMQ is supported by the following key features:
l Master-slave replication: RocketMQ uses a master-slave replication architecture. When the producer sends a message, the message is first written to the master node, and asynchronously replicated to multiple slave nodes. When the master node fails, the slave node can take over the services.
l Reliable message storage: RocketMQ uses the Write Ahead Log (WAL) technology to store messages reliably. A message is first written to a sequential file on the disk and then to the memory. When RocketMQ restarts, you can check the files on the disk and restore unconsumed messages.
l Repeated message storage: RocketMQ uses the unique ID of a message to ensure idempotence. If a message fails to be sent due to network issues or other reasons, RocketMQ determines whether it has been successfully sent based on the message ID so that it will not be re-sent.
l Disaster recovery backup: RocketMQ supports deployment in the broker cluster mode and multi-data center mode for disaster recovery backup of message data. If one broker fails, other nodes can continue to provide the services, ensuring system availability.
l HA design: RocketMQ adopts a multi-broker cluster and achieves high availability (HA) through master-slave replication and fault switchover. If one broker fails, other nodes automatically take over the work to ensure the normal processing of messages.
O&M Management
DMS RocketMQ provides diversified O&M management functions, including multi-dimensional metric monitoring (at the queue level), message query, message tracking, and automatic message deletion upon expiration.