Ctrip’s Technology Evolution: From Call‑Center Era to Big Data and AI

Author Bio: Li Xiaolin, Vice President of Technology at Ctrip and head of the Platform R&D Center, shares his experience from over twenty years in IT and internet development, originally presented at the 2018 Ctrip Technology Summit.

Ctrip, as a leading online travel agency (OTA), has undergone three major technology system evolutions driven by business model changes and overall internet industry trends. The article aims to provide insights for other internet companies, especially early‑stage ones, to avoid common pitfalls.

Current Technology System

Ctrip now operates with roughly 4,000 R&D staff, releases over 8,000 updates per week, runs more than 10,000 applications on 80,000+ host instances across multiple data centers worldwide, achieving four‑nine (99.99%) availability.

The system consists of three core layers:

System Architecture : wireless front‑end platform, distributed frameworks and middleware, and distributed big‑data storage.

Business Systems : hotel, vacation, flight, and other order‑centric services built on the core framework.

Enablement Systems : big‑data and AI platforms for data mining and recommendation, plus an operations and deployment center for unified resource management and monitoring.

Technology Evolution Roadmap

The evolution can be divided into three stages:

Call‑Center Era : Business driven by offline channels; technology based on Windows/ASP + SQL Server, with tightly coupled code and rapid copy‑paste development.

Internet & Mobile Era : Shift to online and mobile traffic; adoption of layered architectures, service‑oriented APIs, high‑availability clusters, and open APIs across platforms.

Big‑Data & AI Era : Leveraging massive user data for personalization, AI‑enabled recommendation, and intelligent customer service bots; adopting cloud, big‑data integration, and AI on top of the cloud infrastructure (the “ABC” strategy: Cloud → Big Data → AI).

Each stage brought specific challenges:

Rapid business growth outpacing the simple two‑tier architecture of the early stage.

Unclear boundaries when splitting subsystems, leading to duplicated functionality and high coordination costs.

Complex process re‑engineering and the need for distributed transaction handling across micro‑services.

Cache consistency and distributed state management before mature middleware like Redis.

To address these, Ctrip invested two years in SOA‑based subsystem separation, creating shared platforms for payment, messaging, and logistics, which laid a solid foundation for later scaling.

Big‑Data & AI Phase Details

The “ABC” strategy defines Ctrip’s focus:

Cloud (C) : Cloud‑based compute, network, and storage.

Big Data (B) : Integrated data sharing across the group.

AI (A) : Personalization, digitalization, and intelligent services.

AI is applied in two main areas: precise marketing and recommendation to boost conversion rates, and intelligent customer‑service robots that handle a portion of the 10,000‑plus call‑center seats, reducing cost and improving response time despite challenges like speech‑recognition accuracy.

In conclusion, Ctrip’s journey mirrors that of many large e‑commerce platforms: continuous iteration, refactoring, and adoption of emerging technologies have driven its current scale, and the company continues to evolve along this path.