Technical Systems Behind JD Logistics for the 11.11 Global Shopping Festival

JD Logistics is renowned for its rapid order delivery, with 90% of warehouse‑distribution orders arriving the same or next day, and the article explores the technical systems that support this performance during the 11.11 Global Shopping Festival.

After a user places an order on JD Mall, the order is split, transferred, and routed through a production pipeline that includes warehouse outbound, network‑wide routing, and optimal path planning for delivery personnel, enabling fast movement of goods, information, and funds.

Warehouse System – Global Reach and Technological Growth The warehouse system supports small‑item, large‑item, cold‑chain, B2B, and cross‑border operations, handling inbound, storage, and outbound processes. Since 2012 it has evolved from single‑machine applications to internet‑scale distributed systems and now to modular deployments, accelerating warehouse expansion.

By applying big data and AI for product layout, intelligent storage, pick‑path optimization, human‑machine collaboration, smart scheduling, and consumable recommendation, JD Logistics achieves lean management, higher efficiency, and lower costs. IoT devices such as unmanned warehouses, AS/RS, conveyors, sorters, and AGVs enable smart park implementations.

Distributed architecture, database optimization, service tuning, degradation control, full‑link stress testing, and monitoring give the system high concurrency, high availability, and elastic scalability, allowing it to handle the massive order spikes during major promotions.

Order‑Routing System – GIS Engine as the Intelligent Core The GIS‑based precise routing platform quickly matches delivery addresses to stations and couriers. During the 11.11 event it processed over 10 million orders per minute with >99% accuracy, using a self‑trained address segmentation model and optimized TextCNN + LSTM structures to build a standardized address tree.

Performance improvements such as multi‑level caching, optimized R‑Tree construction, and grid‑based Geohash encoding keep address‑fence query latency (tp99) under 2 ms.

Delivery System – Multi‑System Collaboration and Efficiency The Qinglong system handles billions of waybills daily, integrating third‑party carriers via a shared platform. Sub‑systems manage merchant waybill integration, data storage, scheduling, sorting (with advanced automatic sorters and edge computing), transportation, and last‑mile delivery, all contributing to peak performance during promotions.

Promise Fulfillment – Real‑Time User Progress The Promise system provides users with delivery time estimates, freight calculation, pickup point recommendations, and various service‑level options (e.g., “Jingzhun Da”, “Jingshun Da”). It uses multi‑level caching and parallel processing to boost throughput.

During 11.11, the checkout API peaked at 99.74 million calls per minute, with total daily calls reaching 165 billion; product‑detail page calls peaked at 24.82 million per minute, totaling 93.7 billion daily. The intelligent scheduling system predicts warehouse and station capacity, transforms unordered production into orderly, real‑time controlled processes, and provides real‑time alerts.

Conclusion JD Logistics’ massive, finely partitioned technical ecosystem—covering weak‑point optimization, slow‑SQL tuning, stress testing, emergency drills, dependency upgrades, information security, and smart hardware—enabled it to maintain the industry’s fastest delivery times with zero incidents during the 11.11 promotion, offering valuable insights for supply‑chain optimization.