Sharing insights on technical analysis and exchange, making life better through technology.
This article explains how Redis and Zookeeper implement distributed locks, compares their performance, consistency, and reliability, and provides guidance on choosing the appropriate lock mechanism for different backend scenarios.
MySQL auto‑increment primary keys are expected to be sequential, but deletions, manual IDs, transaction rollbacks, unique‑key conflicts, and bulk‑insert allocation strategies can all cause gaps, and this article explains each scenario in detail.
This article explains how to efficiently verify whether a user nickname is already registered on high‑traffic websites by combining Bloom filters with Redis caching, reducing database load and improving concurrency handling.
This article explains how to use segmented (sharded) locks and Redis‑based routing strategies to parallelize inventory deduction and order creation, dramatically increasing orders processed per second while maintaining atomicity and fault tolerance.
This article explains why Redis, a single‑threaded in‑memory database, can become blocked, covering five common culprits such as blocking commands, big keys, database flushes, synchronous AOF writes, and the SAVE command, and offers guidance on preventing performance stalls.
This article explains how RocketMQ’s CommitLog architecture—sequential writes, mmap zero‑copy, PageCache acceleration, fixed‑size log files, flexible flushing strategies, and efficient ConsumeQueue indexing—enables the system to sustain million‑level QPS with high reliability and low latency.
This article explains the main conditions that cause Java objects to be promoted from the young generation to the old generation, covering age thresholds, space guarantees, dynamic age determination, and large‑object allocation rules in HotSpot JVM.
JVM’s generational garbage collection divides the heap into young and old generations, using two Survivor spaces to incrementally age objects, apply a copying algorithm, reduce fragmentation, improve memory utilization, and accelerate short‑lived object reclamation, ultimately enhancing overall GC performance.
This article explains a metadata‑driven approach for extending massive tables—using a separate extension table, sharding, and Elasticsearch sync—to add new fields to billion‑row databases without locking the primary table or disrupting online services.
This article explains how ThreadLocal creates per‑thread variable copies, demonstrates its usage with sample Java code, illustrates the internal storage structure, and reveals why improper handling can lead to memory leaks when thread‑local values persist in thread pools.
