The article presents JD's self‑developed JMQ middleware, detailing its 2018 performance improvements, comparative benchmarks with Kafka, storage and index architecture evolution from JMQ2 to JMQ4, and the high‑performance I/O and caching strategies that enable millions of TPS on modern hardware.
This guide explains how to monitor critical Redis health indicators—including ping response, client connections, blocked clients, memory usage, fragmentation, cache hit rate, OPS, persistence status, expired keys, and slow logs—to ensure optimal performance and prevent failures.
This article explains the concept of cache breakdown, illustrates why malicious requests can bypass caches and overload databases, and presents three practical mitigation techniques—using a Redis SETNX‑based mutex lock, asynchronous cache rebuilding, and a Bloom filter—along with their advantages, drawbacks, and performance testing results.
This guide introduces Resilience4j, a lightweight Java library for implementing fault‑tolerance patterns such as circuit breakers, rate limiters, bulkheads, retries, caches, and time limiters, and shows how to configure each module with Maven and example code.
This article explains how Linux memory management differs from Windows, clarifies the output of the free command—including buffers and cache—and demonstrates the performance benefits of caching through a simple 1 GB file read experiment.
Youzan’s Transparent Multilevel Cache (TMC) adds automatic hotspot detection and a 64 MB local cache to existing distributed caches via a Hermes‑SDK‑augmented Jedis client, delivering transparent Java integration, strong consistency, up to 80 % local‑hit rates, and improved QPS during high‑traffic events.
This article provides a comprehensive introduction to Redis, covering its core concepts, advantages, data structures, persistence mechanisms, eviction policies, clustering, common usage scenarios, and typical interview questions for developers working with this high‑performance in‑memory key‑value store.
Linux appears to consume most of a system’s RAM, but the free command’s output, including buffers and cache, actually reflects memory used for performance optimization; this article explains the distinction, how to interpret free’s columns, and demonstrates the impact with a 1 GB file read experiment.
This article explains Linux kernel memory management from a global perspective, covering process address space layout, memory allocation mechanisms, OOM handling, the location of allocated memory, cache behavior, tmpfs usage, and the kernel's automatic memory reclamation strategies.
This article explores Linux kernel memory management, detailing process address space layout, allocation mechanisms, OOM handling, various mmap mappings, cache behavior, tmpfs usage, and the kernel's automatic memory reclamation strategies.
The article explains various cache problems—including consistency, concurrency, penetration, avalanche, and the bottomless‑pit phenomenon—and presents practical mitigation techniques such as active updates, locking, empty‑object caching, request filtering, consistent hashing, and multi‑level caching to ensure reliable high‑performance systems.
Through detailed debugging and source code analysis, the author discovers that MyBatis's SqlSession.clearCache() does clear the first‑level cache, but the observed stale query results are caused by MySQL's default REPEATABLE‑READ isolation level, which can be resolved by setting the transaction isolation to READ COMMITTED.
This article details Alibaba’s first production use of non‑volatile memory in the Tair MDB cache service, covering deployment, performance challenges such as write imbalance and lock overhead, the optimizations applied, and design guidelines for building NVM‑based caching systems.
This article examines three common cache‑update strategies, explains why some lead to data inconsistency, and proposes practical solutions such as delayed double‑delete, asynchronous retries, and message‑queue based recovery to maintain cache‑database consistency in high‑concurrency systems.
This article explains how Weibo handles massive daily traffic of over a hundred billion requests by employing a five‑layer feed system and a six‑layer cache architecture that evolved from simple KV storage to sophisticated counter and existence‑check services, highlighting design choices, performance optimizations, and future directions.
This article examines three common cache‑update strategies, explains their concurrency and performance drawbacks, and proposes practical improvement techniques such as delayed double‑delete, asynchronous retries, and binlog‑based messaging to maintain consistency between cache and database.
This article examines three common cache‑update approaches—updating the cache after the database, deleting the cache before updating the database, and updating the database then deleting the cache—analyzes their drawbacks, and presents practical solutions such as delayed double‑delete and retry mechanisms to ensure data consistency.
This article explains why distributed caching is essential for high‑traffic services, presents three programming approaches (direct code, Spring‑Data‑Redis injection, and annotation‑based), compares three cache‑access patterns, details sharding models, outlines smooth and offline migration steps, and addresses cache penetration, concurrency, avalanche, and transaction support with concrete code examples.
This article explains Linux kernel memory management, covering process address space layout, memory allocation, OOM handling, where allocated memory resides, and both manual and automatic memory reclamation techniques, with practical examples and visual illustrations.
This article translates hardware latencies—from CPU cache and memory to disk and network—into human‑scale time equivalents, using vivid analogies and benchmark data to illustrate why storage and networking appear extremely slow from the processor's viewpoint.
Linux’s free command often shows high memory usage, leading many to think the OS consumes RAM, but understanding buffers, cache, and the distinction between used and free memory reveals that Linux efficiently utilizes idle memory to improve I/O performance, as demonstrated by simple read‑speed experiments.
The article explains Chrome's guiding principles, multi‑process architecture, network request lifecycle, and a suite of performance optimizations—including DNS pre‑fetch, TCP pre‑connect, caching strategies, and predictive pre‑rendering—that together reduce latency and make the browser faster as users continue to browse.
The article thoroughly examines MyBatis’s first‑level (session) and second‑level (namespace) caches, detailing their configurations, internal workflows, source‑code implementations, experimental behaviors, common pitfalls such as stale or dirty data, and advises disabling built‑in caching in favor of dedicated solutions for production.
This article explains why Linux appears to consume most of the memory, clarifies the meaning of the free command’s buffers/cache fields, and demonstrates through a simple experiment how Linux’s caching dramatically speeds up repeated file reads.
An in‑depth comparison of Memcached and Redis reveals how both key‑value cache servers operate, covering their service models, event loops, memory allocation strategies, database implementations, persistence mechanisms, transaction support, and publish/subscribe features, highlighting the trade‑offs between simplicity and rich functionality.
This article explains the fundamentals of CDN, introduces QunarCache (CDN + Cache) architecture, compares cache solutions such as Squid, Varnish, ATS and Nginx, describes deployment patterns, operational monitoring, and future improvements for scalable, resilient content delivery.
This article explains various cache eviction strategies, compares heap, off‑heap, disk and distributed cache types, and provides concrete Java implementations using Guava Cache, EhCache 3.x and MapDB with code examples and usage patterns such as Cache‑Aside and Cache‑As‑SoR.
The article explains why Linux appears to consume most memory when using the free command, clarifies the roles of buffers and cache, shows how to interpret free’s output, and demonstrates through experiments how caching dramatically speeds up repeated file reads.
This article examines why Redis data can disappear, walks through a real incident where 90,000 keys vanished, and provides concrete monitoring, diagnostic, and preventive measures to avoid cache‑only and storage‑critical data loss scenarios.
This guide explains Linux memory reporting with the free command, the roles of buffer and page caches, and provides step‑by‑step instructions to manually release cached memory using /proc/sys/vm/drop_caches, including practical command examples and clarifications of key concepts.
This article explains Linux kernel memory management, covering process address space layout, memory allocation methods, OOM killer behavior, where different types of memory reside, and both manual and automatic memory reclamation techniques, illustrated with diagrams and command examples.
This article explains why caching is essential, defines core concepts such as hits, misses, and costs, compares major replacement policies (LRU, LFU, FIFO, ARC, etc.), and provides Java code examples for implementing these algorithms in a backend system.
The article analyzes the massive scale of Chinese New Year red‑packet activity on WeChat, presents usage statistics, and explains the backend architecture—including distributed KV storage, cache‑layer atomic operations, and database transaction handling—that enables high‑throughput red‑packet distribution.
JIMDB is JD’s in‑house distributed cache platform that combines automatic fault detection, seamless online scaling, multi‑language support, and containerized deployment to replace traditional Memcached/Redis solutions, offering features such as one‑click cluster creation, elastic expansion, lossless scaling, and comprehensive monitoring for high‑traffic e‑commerce services.
This article summarizes common Redis cache update problems, compares four update approaches—including proactive, passive, and pre‑loading methods—and recommends a pre‑load strategy with key versioning to avoid data loss, improve performance, and reduce memory waste.
This article reviews the evolution of distributed locking—from simple MySQL table locks to Redis cache locks and the Redlock algorithm—examines expert criticisms of Redlock’s correctness, presents the Redis author’s rebuttal, and ultimately recommends Zookeeper as a more reliable solution for high‑availability distributed locks.
A multi‑person technical discussion explores the reasons behind low cache hit rates, examines memcached LRU mechanisms, proposes AOP‑based caching strategies, and shares practical solutions such as proactive cache invalidation, lock‑based stampede protection, and workload‑aware configuration for backend systems.
This article explains how traditional ACID transactions fall short for modern large‑scale systems and presents lock‑free transaction methods, atomic and partial cache‑switch techniques, and a PostgreSQL‑style MVCC model for key‑value databases, including practical rules and trade‑offs.
The article details how Weibo’s cache infrastructure progressed from simple Memcache deployments to a sophisticated, service‑oriented architecture featuring multi‑layer caching, proxy layers, dynamic configuration, monitoring, and automated scaling to meet massive read‑write demands and high availability requirements.
The article explains how Redis replicas can return keys that should have expired due to the master‑only expiration process, describes the three expiration strategies (lazy, periodic, active), and offers practical solutions such as using SCAN or upgrading to Redis 3.2 to avoid stale reads.
The article analyzes write‑mostly, read‑light cache scenarios such as driver location updates and counter increments, discusses lock bottlenecks, proposes horizontal sharding and per‑record locking, explores lock‑free approaches with data‑integrity signatures, and summarizes practical optimization techniques for high‑concurrency back‑end systems.
This article presents a systematic overview of web caching, detailing a seven‑layer cache hierarchy, clarifying the differences between buffer and cache, and explaining key cache metrics and design considerations for modern web architectures.
This article analyzes how introducing caches can cause service overload, examines five cache get patterns, and proposes prevention, recovery, and flow‑control strategies for both client and server sides to ensure system stability.
This article explores typical cache challenges in distributed systems—including data consistency, high availability, cache avalanche, and cache penetration—explaining their causes, real‑world scenarios, and practical mitigation strategies to ensure reliable and efficient caching.
Service overload caused by cache failures can cripple dependent systems, but by adopting smart cache get patterns, proactive client‑side checks, traffic throttling, service degradation, and dynamic scaling, developers can both prevent overload and recover gracefully when it occurs.
This article explains how the Linux free command reports memory usage, clarifies the roles of buffer and page cache, shows how to manually drop caches, and reveals scenarios—such as tmpfs, shared memory, and mmap—where cached memory cannot be reclaimed, helping readers achieve a deeper understanding of system memory behavior.
Through a whimsical first‑person narrative, the article walks readers through a CPU’s start‑up sequence, BIOS interrupt handling, loading the boot sector, memory access patterns, the principle of locality, cache usage, and the introduction of pipelining to illustrate fundamental computer architecture concepts.
This article explains how WeChat redesigned its red‑packet system for the 2016 Chinese New Year, detailing north‑south data separation, multi‑layer caching, asynchronous settlement, high‑concurrency controls, sharding strategies and graceful degradation to handle billions of red‑packet requests.
This article explains why cache and database can become inconsistent in distributed systems, why the "invalidate‑cache‑then‑write‑DB" rule is preferred, and presents practical serialization techniques—modifying DB and service connection pools—to guarantee per‑key sequential execution while maintaining availability and load balance.
When Redis reaches its maxmemory limit, it can use one of six eviction policies—volatile‑lru, allkeys‑lru, volatile‑random, allkeys‑random, volatile‑ttl, or noeviction—to decide which keys to remove, with defaults and configurable sampling influencing performance and accuracy.
This article explains the fundamentals of caching, why caches are needed, describes common replacement policies such as LRU, LFU, FIFO, ARC, and provides Java code examples for each algorithm, helping developers choose and implement the right cache strategy for their applications.
This article details the design and implementation of a high‑concurrency, easily extensible proxy built in Go for Weibo's recommendation system, covering background, challenges with twemproxy, technical research, architecture, configuration, logging, monitoring, module breakdown, business logic, performance testing, and future improvements.
This article explains how to enable Rails fragment caching, use HTML fragment cache helpers, understand cache digests, observe read/write fragment logs, and migrate the cache store from file system to Redis by adding redis-namespace and redis-rails gems, configuring cache_store, and handling cache invalidation on data changes.
This article explains the backend architecture behind WeChat’s red‑packet feature, detailing how the system stores packet data, uses cache‑level atomic operations for grabbing, separates claim and settlement steps, and ensures high‑throughput, fault‑tolerant processing during peak usage.
This guide explains the essential MySQL cache parameters—including key_buffer_size, table_cache, sort_buffer_size, read_buffer_size, join_buffer_size, thread_cache_size, query_cache_size, and tmp_table_size—detailing their impact on performance and providing recommended values based on server memory to help DBAs optimize MyISAM and InnoDB workloads.
This article explains how to implement a scalable timeline for a subscription feed by combining MongoDB storage with a Redis list cache, detailing the data model, query challenges, Redis operations for pushing and trimming statuses, rebuilding strategies, and handling data integrity trade‑offs.
This article examines how popular caching systems such as Redis, Memcached, Guava Cache, and Ehcache implement eviction policies like LRU, LFU, and FIFO, comparing their algorithmic approaches, code implementations, and expiration handling mechanisms.
