Sharing insights on technical analysis and exchange, making life better through technology.
Even when scanning tables with tens of millions of rows, MySQL avoids out‑of‑memory crashes by streaming data in small 16 KB net buffers, using socket buffers, and employing an improved LRU algorithm that isolates cold data in the buffer pool’s old generation.
This article explains how Java creates objects for various String assignments—using new String("long"), a literal "long", or concatenated literals "lo" + "ng"—detailing when one or two objects are allocated based on the presence of the literal in the string constant pool.
This article explains why storing daily sign‑in data in a traditional database can cause performance issues at scale and demonstrates how Redis bitmap, a memory‑efficient bit‑level structure, can record and compute continuous user sign‑ins both by date and by user.
This article explains how Java classes are loaded into the JVM, details the class loader’s parent‑delegation mechanism, and walks through the execution flow of a sample program, covering stack frames, heap allocation, dynamic linking, method area, program counter, native method stack, and garbage collection.
This article explores the core functions of e‑commerce shopping carts and compares client‑side storage methods such as Cookies and LocalStorage with server‑side solutions like Redis and MySQL, offering guidance on choosing the appropriate approach based on business scale and reliability requirements.
This article explains the Paxos consensus algorithm, detailing its roles (proposer, acceptor, learner), the two-phase prepare and accept process, handling of proposal numbers, and how it ensures strong consistency across distributed nodes through examples and diagrams.
This article compares Dubbo and Feign, outlining their architectures, features, and suitability for different microservice scenarios, and provides practical guidance on selecting the appropriate RPC framework based on performance, reliability, and development efficiency requirements.
This article explains how to design a scalable leaderboard using Redis Zset for moderate traffic and a sharded‑key plus local cache architecture for ultra‑high concurrency, detailing synchronization strategies, data flow, and practical implementation steps.
This article explains why using Docker to host MySQL is generally discouraged, highlighting challenges in database scaling due to container‑exclusive storage, difficulties sharing data files, and memory contention among containers, while also outlining possible synchronization solutions and scenarios where containerized MySQL may still be viable.
Redis can become a bottleneck when a hot key concentrates traffic on a single shard; this article explains how to split such hot keys into multiple sub‑keys, distribute them across cluster nodes using CRC16 slot mapping, and implement the approach in high‑concurrency scenarios like coupon grabs or real‑time leaderboards.
