The following table compares TDMQ for Apache Pulsar with open-source Apache Pulsar.
|
| Pay-as-you-go is used, ensuring elastic billing. Costs are controllable, and no dedicated Ops manpower is required. | Self-built Pulsar clusters cannot be used elastically, resulting in low resource utilization. Self-built Pulsar clusters require human maintenance, resulting in high Ops costs. |
| It is highly flexible and easily scalable. Customers do not need to worry about the scaling process and can fully leverage the scale effect to handle sudden heavy loads. | Broker nodes can be scaled flexibly. However, manual BookKeeper cluster scaling is complex and may lead to misoperations, affecting data. |
| Cross-AZ deployment is adopted in regions with multiple availability zones (AZs), and messages are stored across regions with three replicas. Tencent Cloud guarantees over 99.95% availability and provides optimized cluster traffic throttling to prevent clusters from being overwhelmed by heavy traffic. | Customers need to deploy clusters across regions to ensure availability, and need to ensure cluster availability under heavy traffic. |
| Security protection is natively supported with Tencent Cloud security products. | Open-source plugins need to be installed and configured. |
| Monitoring alarms are natively supported with Tencent Cloud products with monitoring and alarm capabilities. | Open-source plugins need to be installed and configured. |
Summary | Pay-as-you-go supported, and no need to worry about configurations. Ops-free, and no need to worry about underlying components. Sending and receiving messages using the HTTP protocol of cloud APIs, which is simple and easy to use. High SLA guarantee, and targeted parameter tuning. | Many dependent components, and heavy Ops workload. No Service Level Agreements (SLA) guarantee. Limited security protection capabilities. Unable to precisely manage configurations, leading to resource waste. |
Major Features
Message retry and dead letter mechanisms
Message tagging, supporting tag-based message filtering
listenerName identifier added on the client, supporting multi-network access
Optimized cluster restart time and jitter issues on the server, reducing the impact of restarts on businesses
Unique Features
Complete message queries and traces
TDMQ for Apache Pulsar supports the complete message query and tracing feature from production to storage and consumption, enabling users to quickly locate the status of abnormal messages.
Proactive message repushing on the server
Whether proactively pushing messages not been acknowledged for a long time can be configured on the server. This prevents messages from being lost when messages are not acknowledged due to business failures and prevents excessive message backlogs caused by overlooked acknowledgments.
Single-instance traffic throttling by tenant
Tenant-level traffic throttling for the production and consumption rates and traffic is supported.
Fine-grained metric monitoring for memory usage, internal data pulling traffic, and rate of concerned objects
Fine-grained monitoring of core memory resources has been implemented to facilitate observation and statistics of the memory usage of different resources. The rate and traffic for reading messages from the bookie are monitored.
Visualized monitoring of bookie data compression
Complete information on the bookie data compression process is displayed, including compressed ledgers and the time required for processing each ledger.
Read/Write traffic throttling and dynamic configuration capabilities for bookie data compression
Read traffic throttling is supported during compression to prevent excessive disk bandwidth consumption. In addition, the compression traffic throttling capability can be adjusted dynamically.
Optimized the bookie client to reduce the fault recovery time for AZ disaster recovery
Faulty bookie nodes can be removed quickly, improving the overall fault recovery speed of the cluster.
Others
Background maintenance
Many community features are not supported in earlier versions. The TDMQ for Apache Pulsar team closely tracks community developments and selects valuable features and bug fixes to merge.
Support and expert services
Services, such as product upgrades, new service launches, and major marketing campaigns, are provided to ensure smooth business operations.