Apache Dubbo Website

Dubbo Rust is a high-performance, extensible RPC framework designed to align with the Dubbo3 core feature set. It serves as a data plane component within the Dubbo Mesh architecture, receiving configurations from a control plane to provide service governance and RPC communication capabilities.

Key features include:

• High-performance communication based on the HTTP/2 Triple protocol.
• IDL-based RPC definitions using Protobuf for code generation and Serde for serialization.
• Support for multiple communication models: request-response (unary), request/response streaming, and bi-streaming.
• An extensible architecture supporting components like Listener, Connector, Filter, Protocol, and Invoker.

dubbo-go is the Go language implementation of Apache Dubbo. It is a microservices development framework designed to solve RPC communication and service governance challenges in the Go ecosystem.

Key features include:

• API and RPC Protocols: Provides HTTP/1/2 based communication and streaming models.
• Microservice Governance: Built-in support for service discovery, traffic control, observability, full-link tracing, and logging.

Apache Dubbo is a high-performance, lightweight microservices framework designed for building distributed systems. While originally developed in Java, it has expanded to support multiple languages including Go, Rust, and Python, enabling cross-language and cross-platform communication.

Key features include:

• Service Governance: Includes service registration, discovery, load balancing, fault tolerance, and call chain tracing.
• High Performance: Optimized for efficient communication in distributed architectures.

The Sentinel console provides a unified interface for managing traffic control and circuit breaking rules. Key features include:

• Single Device Monitoring: View individual machine details including device name, IP address, port number, health status, and heartbeat time.
• Link Monitoring: Analyze real-time execution conditions for specific client resources. It supports two modes:
  • Tree Structure: Displays resource call links in a hierarchical tree.
  • Flat Mode: Displays resource execution status without hierarchical distinction.
• Aggregate Monitoring: Provides real-time, second-level accuracy monitoring for all machines belonging to the same service.
• Rule Configuration: Allows real-time viewing and configuration of flow limiting, degradation, and system protection rules.

Apache Dubbo 3.3 introduces several key enhancements for modern workloads:

• Native Image AOT Support: Supports Ahead-of-Time compilation to create native binaries, significantly reducing startup time and memory usage (ideal for serverless).
• Project Loom Support: Optimizes thread management using virtual threads, improving concurrency handling and simplifying asynchronous programming.
• New Routing Rules: Provides more flexible traffic control and service governance for large-scale deployments.

The dubbo2.js library provides the following capabilities for Node.js developers:

• Zookeeper Support: Integration with Zookeeper for service discovery.
• Native Dubbo Protocol: Full support for the standard Dubbo communication protocol.
• Direct Connection: Ability to connect directly to services without a registry.
• Full-link Tracing: Support for distributed tracing across the call chain.
• Interface Generation: Automatic generation of Dubbo interfaces for easier consumption.

The Dubbo Go development roadmap includes the following key objectives:

• Dubbo Router Implementation: Swiftly implementing the router by leveraging enhanced governance configurations that support parameter delivery via configuration centers like Zookeeper and Apollo.
• Distributed Tracing: Introducing a mainstream tracing solution using OpenTracing as the standard to integrate with the OpenTracing ecosystem.
• Kubernetes Operators: Implementing service invocation through operators to create a new Kubernetes-centric registration center.
• Service Mesh Integration: Assimilating into the cloud-native ecosystem by aligning with Istio to define Dubbogo's role within a service mesh environment.
• Compatibility: Aiming for functionality that meets Dubbo 2.7 compatibility.

Dubbo provides tools to address the complexities of distributed microservices R&D:

• Service Testing: The Dubbo RPC protocol supports curl access, facilitating easier development collaboration and manual testing.
• Service Mocking: Tools are available to mock services for isolated development.
• Dubbo Admin Console: A centralized console to visualize operations, manage documents, and handle single-machine operation and maintenance.

Dubbo3 is a cloud-native service framework designed for high scalability and performance. Its core features include:

• New Service Discovery Model: Application-level service discovery designed for cloud-native infrastructure and heterogeneous systems, enhancing cluster scalability.
• Triple Protocol: A next-generation RPC protocol based on HTTP/2 that is compatible with gRPC. It supports strong gateway penetration, is multi-language friendly, and supports Reactive Streams.
• Unified Traffic Governance Model: Provides uniform governance rules across SDK, Mesh, VM, and Container environments.
• Service Mesh: Supports both Sidecar Mesh and Proxyless Mesh (as of version 3.1.0) to reduce migration costs and provide architectural flexibility.

Dubbo 3 provides significant performance improvements in service discovery compared to version 2.x, specifically regarding resource utilization and Garbage Collection (GC) overhead.

Resource Utilization

• Interface-level service discovery: Resident memory on a single machine decreased by approximately 50% compared to version 2.x.
• Application-level service discovery: Resident memory on a single machine decreased by approximately 75% compared to version 2.x.

Garbage Collection (GC) Performance

• Interface-level service discovery: GC consumption during address changes decreased by an order of magnitude (from hundreds of GCs down to tens).
• Application-level service discovery: The number of GCs approached zero, with Full GC (FGC) counts declining significantly from hundreds in version 2.x to zero.