From Personal Homepage to Billion-Visit Sites: Key Lessons in Scalable Web Architecture

A website is like a person, growing from infancy to adulthood; nurturing it requires different methods at each stage, yet shares common principles.

1: Accumulation Is Essential

Architects are not made in a day. In 1999 I built a personal homepage with Dreamweaver, table layouts, a DB connection, and a few lines of PHP, then uploaded via FTP. By 2000 I expanded into network fundamentals, Linux, AIX, FreeBSD, and set up services such as RealServer streaming, FTP, Battle.net gateway, Apache, DNS, and Qmail for school users.

When traffic reached ~100k PV, I introduced MySQL master‑slave replication and index optimization. The site’s 2‑tier architecture persisted through many projects, with JSP vs PHP debates irrelevant because I still used a simple 2‑tier model: JSP pages directly accessing MySQL without transactions or fault tolerance.

In 2005, at a financial project, I encountered MULE ESB and realized middleware’s role in isolating specific system functions. By 2006 at Alibaba Software, I experienced MVC’s three‑layer architecture, appreciating its extensibility and team scalability.

From 2007 onward, I recognized the need for service‑center and persistence layers as sites grew to tens of millions of daily visits, leading to the adoption of three‑tier designs, user‑center projects, and SAAS concepts.

In 2009 I joined the newly formed Aliexpress team, confronting the challenges of scaling from small systems to billion‑PV e‑commerce platforms.

2: Knowledge Structure

Website architects come from diverse backgrounds—computer science, art, biology, physics, even policing—bringing cross‑disciplinary perspectives. Core competencies include programming experience, algorithms, design patterns, multithreading, remote calls, and data source handling.

Deep understanding of networking, protocols (HTTP, cookies, keep‑alive), data formats (JSON, XML), and performance‑critical mathematics for QPS, latency, and resource planning is essential.

Familiarity with JVM memory management, GC behavior, and database concepts (indexes, storage structures) is also crucial, as is avoiding direct DB access from front‑end services in high‑traffic scenarios.

Continuous learning of emerging middleware (e.g., Dubbo, HSF, Redis) and adapting architectures to fit current scale and operational costs is a must.

3: Design Philosophy

Architects aim for at least one‑year foresight, preferring horizontal scalability. Common tactics include:

Asynchronous over synchronous : isolate async services for degradation and improve QPS.

Centralized to distributed : start with monolithic modules, then split services as traffic grows.

Layered architecture : evolve from 2‑tier to 3‑tier (application, service‑center, persistence) and beyond.

Functional decomposition : separate high‑traffic, latency‑sensitive, or CPU‑intensive modules.

Service‑centerization : encapsulate business logic as stateless services without UI.

Node monitoring : implement snapshot, baseline, and key‑metric monitoring, balancing overhead.

4: Core Infrastructure

Scaling to billion‑PV sites demands careful capacity planning, fault tolerance, multi‑region disaster recovery, and latency considerations (e.g., 10 ms intra‑region, 200 ms inter‑continent). Infrastructure must support high QPS, separate read/write paths, and employ sharding, caching, and distributed storage.

Monitoring systems provide health dashboards, baseline trends, and targeted business‑logic metrics, while acknowledging the performance cost of instrumentation.

5: Software Engineering

Beyond code, architects establish release processes, gray‑scale deployments, coding standards, mentorship, and prioritize business impact. Different architect roles (PM‑type, consulting, performance‑focused, middleware‑focused) emerge based on organizational needs.

6: Variations Across Business Types

Each website’s architecture must reflect its specific domain—e‑commerce, logistics, payment, social media—so copying another’s stack wholesale is ineffective. Tailor middleware, data models, and integration patterns to the unique workflow.

7: Emerging Trends

Open platforms, social networking, group‑buy models, mobile/tablet experiences, recommendation engines, cross‑border trade, IoT devices, and long‑tail dynamics all influence architectural decisions.

8: Last but Not Least

Successful architecture requires the right mindset: openness, attitude, and the ability to persuade stakeholders. Proactive architects who anticipate problems are more valuable than those who merely react.