Architecture Essentials and the Multi‑Layer Design of Sina Weibo's Large‑Scale System

Before discussing the essence of architecture, the author reflects on the scale of million‑level websites, emphasizing strategic importance and tactical humility, using Uber's order volume as an example to illustrate the magnitude of traffic and data storage challenges.

The concept of architecture is described as an abstract "rack" that holds business logic and algorithms, emphasizing abstraction, reuse, and forward‑looking design based on experience and industry insight.

Key capabilities for architects include strategic decomposition, abstraction (de‑duplication), classification (service modularization), and algorithmic performance optimization across CPU, memory, I/O, and network.

The article then examines Sina Weibo's overall architecture, which follows a three‑tier model: client (Web, Android, iOS), an interface layer providing security isolation, traffic control, and platform differentiation, and a backend consisting of platform services, search, and big‑data processing.

Technical challenges are analyzed using an orthogonal decomposition method, separating horizontal layers (interface, service, data storage) and vertical concerns (business architecture, technical architecture, monitoring, governance).

Common design principles highlighted include RPC services, message middleware for asynchronous decoupling and traffic shaping, and configuration management for graceful degradation and gray releases.

Statelessness at the interface layer is discussed, noting that state is shifted to caching or storage layers to maintain scalability.

Physical‑to‑logical mapping is illustrated, showing how teams are organized vertically (business groups) and horizontally (infrastructure groups) to improve collaboration.

The multi‑level cache architecture of Weibo's feed is detailed: L1 caches (small, high‑QPS) and L2 caches (large capacity) work together to reduce database load, with dual data‑center deployment for high availability and geographic latency reduction.

Distributed service tracing is introduced to address stability and performance monitoring at massive scale, using unique request IDs propagated across RPC calls and AOP‑based instrumentation to achieve low‑intrusion full‑stack observability.

Operational practices for handling peak traffic (e.g., Chinese New Year) include graceful degradation plans, full‑stack load testing, and shared Docker clusters for resource efficiency.

Finally, the author outlines a personal learning roadmap: mastering Java, the JVM, operating systems, design patterns, TCP/IP, distributed systems, data structures, and algorithms, encouraging continuous study to develop one’s own architectural principles.