The article explains why the last six digits of Taobao order numbers are constant, revealing the “gene” method that embeds a hashed user identifier to bind orders to specific shards, thereby improving query performance, ensuring balanced data distribution, and supporting scalability in large e‑commerce systems.
The article explains how Taobao embeds a user‑specific “gene” in the last six digits of its order numbers, enabling efficient sharding routing, uniform data distribution, and idempotent order handling while maintaining global uniqueness in a large e‑commerce system.
This article examines common order‑number generation rules and compares four practical solutions—UUID, database auto‑increment, Snowflake algorithm, and Redis INCR—providing code examples and best‑practice recommendations for building globally unique, fast‑producing identifiers in distributed backend systems.
This article presents several approaches for generating unique, scalable order numbers in Spring Boot, including UUIDs, database sequences or auto‑increment IDs, timestamp‑based strings with random components, and distributed Snowflake IDs, each accompanied by Java code examples and discussion of their advantages and drawbacks.
After encountering duplicate order IDs during a system upgrade, the author analyzes the shortcomings of the original timestamp-and-random-based generator, demonstrates concurrency tests revealing collisions, and presents a revised Java implementation using atomic counters, shortened timestamps, and IP suffixes to reliably produce unique order numbers even under high load.
The article analyzes a real‑world incident where duplicate order IDs were generated under high concurrency, demonstrates the shortcomings of the original Java implementation, and presents a thread‑safe redesign using AtomicInteger, Java 8 date‑time API and container IP suffix to guarantee unique identifiers across parallel requests and clustered instances.
This article explains how to design a high‑throughput payment system using database sharding, Snowflake‑style globally unique order IDs, eventual consistency via message queues, high‑availability architectures, data tiering, and coarse‑fine traffic control to handle massive request spikes.
This article explains how a high‑throughput payment platform uses database sharding by user ID, Snowflake‑style globally unique order IDs, asynchronous replication for eventual consistency, multi‑level data caching, and coarse‑fine traffic pipelines to achieve millions of requests per second with robust high‑availability.
LeTV upgraded its payment architecture in 2015 by sharding databases, designing a Snowflake‑based globally unique order ID, implementing asynchronous replication for eventual consistency, building high‑availability master‑slave clusters with LVS and KeepAlive, tiering data caches, and adding a coarse‑fine traffic pipeline to reliably handle up to 100,000 orders per second.
The article details how LeTV upgraded its payment platform in 2015 to handle up to 100,000 orders per second by implementing database sharding based on user IDs, a Snowflake‑style globally unique order ID, asynchronous consistency via message queues, multi‑level data caching, and high‑availability architectures for both master‑slave and load‑balanced deployments.
This article explains how LeTV's payment platform handled a hundred‑fold surge in traffic by redesigning its architecture with database sharding, a custom order‑ID scheme, asynchronous consistency, high‑availability clustering, data tiering, and a coarse‑fine request pipeline to sustain up to 100,000 orders per second.
