Technical Strategies for Scaling and Optimizing JD.com Advertising Systems During the 618 Promotion
The article details JD.com's advertising division's comprehensive backend engineering efforts—including traffic handling, data pipeline upgrades, memory optimization, and disaster‑recovery designs—to ensure system stability and performance during the high‑traffic 618 sales event.
JD.com’s Commercial Promotion Department is responsible for monetizing the platform’s massive traffic, and its advertising teams have organized a series of preparation tasks around the core advertising system to meet the challenges of the 618 promotion.
Traffic Battle: The ad playback system faces tens of thousands of QPS during peak periods. The Advertising Architecture team leveraged years of technical experience to create a mature large‑scale promotion solution, implementing dual‑hot incremental data streams, dual‑hot Kafka sources, migration to CFS storage, flexible degradation plans with rehearsals, and a pipeline upgrade that doubled processing speed and moved batch processing to a clustered model for higher stability.
Memory and Latency Optimizations: Search and external ad services reduced memory usage dramatically, saving over 10 TB of RAM and cutting TP99 latency by 30‑35 ms, which directly improved revenue. Model inference customizations saved thousands of CPU cores and lowered CPU usage by 25 %.
Data Battle: The Data team rebuilt the Shenzhen real‑time data flow architecture and migrated financial data from JimDB to a custom financial state storage service. They performed extensive stress testing, built dual‑active disaster‑recovery clusters, and introduced the Minos data‑quality monitoring platform to ensure high‑quality, consistent data.
Material Battle: The Customer Platform team handled the surge in ad material submissions, introducing merchant tiering, smart bidding tools (eCPC, tCPA), and automated scaling to balance resource consumption with revenue growth.
Across all teams, a systematic five‑step preparation process—requirement gathering, system analysis, resource assessment, performance testing, and contingency rehearsals—was followed, with early communication starting in March. Continuous monitoring, automated disaster‑recovery scripts, and full‑stack alerting ensured high availability throughout the promotion.
The article concludes with a poetic acknowledgment of the engineers’ dedication, highlighting night‑time testing, rapid issue resolution, and the collective spirit that drove the successful 618 campaign.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
JD Retail Technology
Official platform of JD Retail Technology, delivering insightful R&D news and a deep look into the lives and work of technologists.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
