How AI and Big Data Powered the Hit Series ‘The Longest Day in Chang’an’

The Chinese drama “The Longest Day in Chang’an” not only dazzles with its historical scenery but also showcases a sophisticated “Tang‑era black tech” stack built on AI and big‑data systems.

AI Recommendation System (Beidouxing)

Alibaba’s “Beidouxing” AI brain analyzes scripts, selects actors (e.g., casting Lei Jiayin), predicts audience emotion curves, and optimizes trailer creation, turning uncertain content creation into a data‑driven process.

Content creation understanding

Predictive capabilities during production

Data‑driven audience matching in promotion

Interactive Viewing – Cool Look Mode

Viewers can interact with the series via multi‑stream synchronized playback, dynamic rendering, and on‑screen “black‑tech” features such as encyclopedia tips and storyline maps.

Multi‑stream same‑screen display

Intelligent smooth switching

Frame‑level video synchronization

4K/HDR Remaster

Youku applied cloud‑based HDR rendering and a 2K‑to‑4K super‑resolution algorithm to deliver ultra‑high‑definition experiences on OTT devices, even enabling near‑HDR playback on 70% of smartphones.

Smart Bitrate Selection

The system splits video files into 5‑10 second segments and dynamically selects the appropriate resolution and bitrate based on network conditions, eliminating buffering and delivering seamless playback.

Real‑time bitrate adaptation

Low‑latency decision engine

Anti‑Leech Architecture

Facing billions of leeching requests, Youku deployed a full‑link, multi‑dimensional anti‑leech framework that combines watermarking, request validation, and traffic analysis to protect the content.

Media Asset Platform

The media‑asset middle platform stores, queries, and searches all of Youku’s media data, handling billions of daily API calls with high availability, low latency, and scalable architecture.

Underlying storage and domain model design

Search system architecture

Data aggregation and computation pipelines