JD Microservice Component Platform Overview and Architecture

JD's Microservice Component Platform serves as the underlying architecture for service calls, message notifications, operations management, knowledge sharing, collaboration, and service evaluation within the JD Group.

The platform is built on three core communication technologies—JSFRPC, JMQ messaging, and a service mesh—supporting synchronous, asynchronous, and hybrid communication across various protocols and languages, handling the majority of the group's backend traffic.

Beyond basic communication, the platform has evolved to enable external empowerment and componentization, connecting developers, product managers, project managers, and decision makers across teams and departments, fostering efficient collaboration.

Its architecture consists of three main parts: the core layer, the ecosystem toolchain layer, and the foundational data service layer, each comprising multiple modules such as infrastructure, framework, system extension, and application layers.

Core Layer

Infrastructure Layer: Utilizes JDOS container clusters as the foundation, integrating with JDOS 2.0 and other infrastructure components.

Framework Layer: Includes JSF SDK, ContainerMesh, JSF Registry, and JMQ, with upcoming enhancements to security.

System Extension Layer: Provides features like service call graphs, flow control, and monitoring in partnership with UMP.

Application Layer: Offers the "Service Marketplace" for knowledge search, tagging, commenting, collaborative editing, and the "Open Platform" for gateway services and one‑stop service publishing.

Ecosystem Toolchain

Visual design tools based on DDD to aid service design.

Rapid development and debugging environments.

Online multi‑party debugging tools.

Configuration center for dynamic settings.

Distributed lock and transaction solutions.

API gateway similar to Zuul/Kong.

Foundational Data Service

Provides unified IP, data center, system, application, and grouping information, consolidating scattered data into a shared service for developers.

Application Layer – Service Marketplace

Search by basic and custom attributes, supporting fuzzy and categorical queries.

Comprehensive knowledge base with parameters, owners, call graphs, version history, and snapshots.

Permission and workflow controls for service access.

Quality tracking with importance, health, and architecture assessments.

Call relationship visualization via service call graphs.

User-defined tags for services and interfaces.

Commenting and interaction between service providers and consumers.

Collaborative editing with change history.

System Extension – Service Call Graph

Analyzes source, entry, path, latency, bottleneck, and dependency of service calls, supporting over 2,200 applications and 46,000 IPs.

Microservice Flow Control

Version‑aware traffic control.

Gradual gray releases, A/B testing, and canary deployments.

Dynamic configuration for thresholds and timeouts.

Full lifecycle support for service deprecation.

Forced downgrade, rate limiting, and circuit breaking under critical conditions.

Governance of obsolete interfaces.

Configuration Center

Offers RESTful K‑V APIs for string, JSON, and properties formats, enabling cross‑language and cross‑platform dynamic configuration.

Service Framework – JSF SDK

Provides probes, extension points, new communication protocols, annotations, and support for extensible service attributes to improve observability and flexibility.

Service Mesh

Developed to decouple business logic from communication and governance, allowing multi‑language interoperability and transparent upgrades.

Service Discovery (JSF Registry)

Supports multi‑data‑center, cross‑WAN, and disaster‑recovery features, handling nearly 30,000 JVM instances and 3 × 10⁴ service interfaces.

Security System

Implements a global unique ID per service for certificate, key management, authentication, and authorization, with namespace and security level controls.

Ecosystem Toolchain

Future tools include distributed locks and API gateways to address challenges in distributed application development.

Foundational Data Service

Aggregates IT data across JD’s e‑commerce, logistics, finance, and cloud divisions to support precise, reliable flow‑control operations.

© JD Technology