Customer-Facing Analytics,
Powered by Real-Time Data

Real-Time OLAP for the JD.com Search Box

“Replacing Flink’s window computing with Doris can not only improve development efficiency, adapt to dimension changes, but also reduce computing resources.”

Real-time analytics for the JD.com search box: overall search traffic, live A/B test monitoring, and trending search terms, all at SKU-level granularity for business analysts.

• 10 billion rows processed per day
• 10,000 QPS with query latency as low as 150 ms
• Real-time ingestion at 1 million rows per second

Real-Time Recommendations for Xanh SM

“Apache Doris and its Compute-Storage Decouple Mode allowed us to run both workloads from a single storage layer, cutting infrastructure cost and complexity without sacrificing performance.”

Xanh SM, Vietnam’s leading EV ride-hailing platform, runs analytics and real-time serving on Apache Doris to power personalized destination recommendations.

• 175,000 records written per second with sub-second latency
• Average query latency around 76 ms across most time windows
• P9999 latency only occasionally above 100 ms

Real-Time Analytics for ZTO Express

ZTO Express rebuilt its real-time analytics over 500 million daily record updates, with inverted indexes on high-frequency filter fields serving monitoring, multi-dimensional analysis, and precise filtering.

• Multi-dimensional filter queries down from over 1 minute to under 1 second (60×); complex aggregations from 5 to 10 minutes to under 1 minute
• Concurrency on critical queries up from under 50 to 100+, with timeouts down from 30% to under 5%
• One-third of the original hardware, with 500 million daily updates visible in queries within seconds

ZTO EXPRESS