This article presents a detailed performance comparison of ClickHouse using two real‑world cases, analyzes query timings, explores parameter tuning, schema and SQL adjustments, and summarizes practical conclusions for achieving sub‑second analytics on billion‑row datasets.
This article explains a practical approach to performance‑testing Dubbo add and delete methods of a message queue by pre‑generating payloads, using a thread‑safe LinkedBlockingQueue, and driving the calls with JMeter‑compatible Java code.
The article introduces a Java performance testing framework that shifts from request‑centric to method‑centric design, provides a reusable ThreadBase class, and demonstrates concrete implementations for HTTP requests, database queries, and a concurrent executor to collect timing metrics, while also discussing the merits of coding versus tooling and Groovy advantages.
This article introduces a lightweight Groovy TimeWatch utility that lets developers record execution times with customizable markers, offering both millisecond and nanosecond precision, and explains its implementation, usage, and future multithreading considerations.
This article details a step‑by‑step optimization of a Java search API—covering OpenSearch configuration, batch caching, thread‑pool tuning, and Redis pipelining—that reduced average response time from 150 ms to 32 ms and met a sub‑100 ms SLA.
This article outlines the CoolJia search service quality system, detailing its architecture, the three‑pronged business testing approach, performance and effect testing, data synchronization tests, and the dedicated debug platform to ensure reliable, high‑performance search operations.
This article explains that high concurrency is fundamentally about efficiently squeezing CPU resources, discusses the evolution of concurrency models, presents a control‑variable analysis of HTTP request flow, and compares PHP‑Swoole with Java‑Netty through detailed benchmark results and code examples.
This article debunks the myth that larger database connection pools always improve performance, explains how CPU, disk, and network I/O affect optimal pool size, presents real‑world test results, and provides a practical formula for calculating the right number of connections.
This article explains the challenges of handling Double‑11 traffic spikes, introduces adaptive load‑balancing concepts, analyzes the 5th Middleware Performance Challenge scenario, and outlines algorithm design considerations and evaluation steps for building a robust, self‑adjusting load‑balancing solution.
This article explains the evolution and advantages of video SDKs, details the core playback pipeline—including decoding, rendering, and audio‑video synchronization—covers cross‑platform implementation on Android, iOS and PC, and shares practical testing experiences and performance‑optimization techniques.
An in‑depth performance benchmark compares MySQL on CentOS 6 and CentOS 7 using sysbench with 10 tables of 5 million rows, revealing up to 40% higher QPS on CentOS 7, detailing test setup, results, and key configuration tweaks that dramatically improve throughput.
Based on data transmission and reliability metrics, this article compares Apache Storm and Apache Flink in stream processing, presenting benchmark designs, test environments, results for synthetic and Kafka data, and offers practical recommendations such as operator chaining, object reuse, and checkpoint strategies to maximize Flink performance.
The article details how a Tencent front‑end engineer migrated a Node.js service to Tencent Cloud Serverless, used WeTest load‑testing to expose cold‑start and network latency issues, applied code and connection optimizations plus regional deployment, achieving sub‑200 ms response for 200 concurrent users while outlining further roadmap improvements.
The case study details how Pandora’s marketing solution for Tencent’s games was migrated to a managed Tencent Cloud Kubernetes cluster, compares container‑service, serverless, and self‑built options, outlines the new image‑based deployment workflow, autoscaling, monitoring practices, and demonstrates the operational benefits of tighter dev‑ops collaboration.
This article introduces HttpRunner, covering supported platforms, installation commands, key features such as parameter extraction, global configuration, variable handling, data-driven testing, report generation, and practical steps for converting HAR files to test cases and executing them with hrun.
This article examines the limitations of HTTP/1.x, explains how HTTP/2 addresses them, details the NDSF framework's implementation choices, presents comprehensive JMeter‑based performance tests, and analyzes the results to show traffic reduction, latency stability, and overall business benefits.
This guide walks through integrating Codecentric's Chaos Monkey into a Spring Boot microservice ecosystem, covering dependency setup, configuration, actuator endpoints, Dockerized MySQL, performance testing with Gatling, and timeout tuning for Feign and Ribbon clients to simulate real‑world failures.
To meet Xianyu’s rapid‑release targets, the team built a CI‑integrated automation suite that gathers Flutter FPS, CPU, memory and network data via extended Android tools, drives mixed Flutter/Native UI tests with Appium and integration‑test, and generates lane‑by‑lane performance reports that quickly flag regressions across dozens of feature branches.
This article presents a comprehensive methodology for performance testing and optimization of a new corporate banking system, covering evaluation metrics, monitoring tools, test environment setup, common bottlenecks, detailed troubleshooting steps, and real‑world case studies to achieve higher throughput, lower latency, and better scalability.
This presentation details how to correctly benchmark MySQL middleware performance, exposing common pitfalls, describing practical testing methodologies, emphasizing the need to observe both middleware and actual database pressure, and discussing distributed transaction considerations and metric selection for reliable results.
MAXIM is Youzan’s distributed full‑link load‑testing engine built on Gatling, featuring a central control center, multiple load injectors, a GUI for test orchestration, data‑parameter binding, real‑time injector monitoring, automated reporting with historical retention, and extensible architecture supporting Dubbo and centralized InfluxDB logging.
This article explains how to conduct performance testing for long‑connection services such as MQTT using JMeter, highlights common factors that restrict connection counts like system port limits and file descriptor caps, and provides step‑by‑step instructions to adjust sysctl and ulimit settings to obtain accurate results.
Youzan’s full‑link load‑testing architecture combines a traffic generator, a data‑factory pipeline, and the Maxim platform to replay realistic e‑commerce user actions, tag and isolate test traffic via unified headers, route reads/writes to shadow storage, and integrate Gatling for capacity planning, degradation, alarm, disaster‑recovery and throttling drills.
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.
This article introduces thirteen practical Linux operation tools—including Nethogs, IOZone, IOTop, IPtraf, IFTop, Fail2ban, and more—providing concise descriptions, download links, and step‑by‑step installation commands to help system administrators monitor performance, network traffic, and protect against attacks.
This article describes a real‑world test of loading over 2 billion records from a big‑data platform into MySQL using XeLabs TokuDB, showing configuration details, performance results, and practical tips for achieving up to 570 k rows per second on a cloud instance.
At Didi, the team replaced a chaotic mix of Kafka, Redis, and other queues with a custom, RocketMQ‑based service, using dual‑write and dual‑read migration, extensive performance testing, custom failover, batch extensions, and operational tweaks to achieve stable high‑throughput, low‑latency messaging at massive scale.
This article chronicles Alibaba's decade‑long journey of capacity planning for Double 11, detailing five evolutionary phases—from manual estimates to full‑link testing ecosystems—while balancing cost, stability, and efficiency in massive distributed systems.
This article explains Android's VSync mechanism, defines smoothness metrics such as FPS, skipped frames and smoothness, describes how VSync timing affects rendering, and provides practical testing commands like `adb shell dumpsys gfxinfo` and `adb shell dumpsys SurfaceFlinger` for measuring frame performance.
Over ten years of Alibaba's Double 11, the company transformed a modest marketing event into a global e‑commerce platform by continuously improving backend architecture, scaling strategies, full‑link stress testing, multi‑active data centers, cloud migration, and real‑time incident response, offering valuable operational insights.
The JD Finance technical teams—including Wealth R&D, Consumer Finance, Payment, Middle‑Platform, and Crowdfunding—conducted comprehensive system reviews, performance stress tests, capacity expansions, monitoring enhancements, and emergency downgrade plans to ensure stable, high‑throughput service during the 11.11 shopping festival.
From the humble 2009 launch of Double 11 to the massive, cloud-native, multi-region architecture of 2018, Alibaba’s engineers chronicle yearly technical hurdles—traffic spikes, system crashes, CDN limits, over-selling, and the evolution of stress-testing, capacity planning, and operational safeguards that turned the shopping festival into a global engineering showcase.
This article presents Lest, a persistent KV‑store cache designed to overcome the limitations of traditional in‑memory caches by combining disk‑based persistence, lock‑free multithreading, custom protocols, and advanced load‑balancing, and it evaluates its performance on petabyte‑scale workloads.
The second Hangzhou Testing Salon, organized by Youzan with partners KuJiaLe, Mogujie and 51 Credit Card, gathered over 300 attendees from 130 companies to share practical insights on full‑chain performance testing, a distributed pressure‑testing engine, rule‑based quality platforms, front‑end coverage tools, search‑service reliability and data‑quality monitoring.
This article outlines common challenges managers face with performance testing, explains when testing should be introduced, describes its process, types, metrics, and quality checks, and provides practical guidance for analyzing requirements and ensuring effective, repeatable performance test execution.
The article details the structure, daily responsibilities, quality‑assurance processes, automation strategies, tool development, performance testing, and quality‑culture initiatives of the Shanghai testing team at Dada‑JD.com, illustrating how systematic testing and DevOps practices enhance product stability and development efficiency.
This article explores PostgreSQL’s NoSQL capabilities by comparing JSON and JSONB data types, demonstrating their storage differences, read/write performance tests on millions of rows, index strategies, and the new full‑text search support for JSON/JSONB introduced in version 10.
This article explains how to reduce HTTP requests and page load time for H5 web pages by leveraging Chrome DevTools and its WebSocket‑based protocol to capture network metrics, render trees, and screenshots on both PC and Android, providing step‑by‑step debugging and performance‑testing guidance.
This article explains how to integrate Microsoft Azure DevOps Server (formerly TFS) with the open‑source load‑testing tool JMeter, describing the background of DevOps testing challenges, JMeter features, a demo PartsUnlimited project, pipeline design, environment setup, script execution, and performance‑testing options.
This article provides a comprehensive overview of the daily work of the Shanghai R&D Center Test Team at Dada-Jingdong to Home, covering project quality assurance, automation, tool construction, and quality culture building.
This article explains how to automatically generate and parse Android Traceview files to track SDK API response times, thread activity, and memory usage, detailing trace file structures, generation methods, and a workflow for collecting and analyzing performance data.
This article introduces Java Microbenchmark Harness (JMH), explains why precise benchmarking matters, and walks through twelve typical testing pitfalls—such as dead‑code elimination, constant folding, loop misuse, fork isolation, method inlining, false sharing, branch prediction, and multithreading—showing how JMH helps developers obtain reliable performance measurements.
This article provides a comprehensive overview of performance testing, explaining its definition, key metrics such as throughput, response time and concurrent users, the relationship between these indicators, optimal testing timing, a step‑by‑step workflow from requirement analysis to reporting, and practical tips for effective load testing and system optimization.
This article introduces Badboy as a free web‑script recorder, explains how to add assertions and parameterize variables, shows how to export the recorded script to JMeter, and outlines the subsequent steps to run a performance test with JMeter.
This article explains the principle of HTTP pipelining, presents detailed performance tests comparing single‑thread, thread‑pool, and pipelined request methods, and provides a simple .NET implementation that demonstrates why pipelining can achieve ten‑fold speed gains over conventional HTTP keep‑alive.
This article explains the concept of HTTP pipelining, presents a series of performance tests comparing pipe, thread‑group, and asynchronous request methods on a typical PC, analyzes the results and underlying TCP behavior, and provides a simple .NET implementation for practical use.
This article chronicles JD.com's journey of evolving mobile backend API testing—from manual proxy checks to script‑based JMeter suites, then to integrated testing and performance platforms, and finally to a continuous DevOps pipeline that automates verification, monitoring, and delivery of high‑traffic services.
A Chinese company, StarRing Technology, became the first ever to pass the notoriously difficult TPC‑DS benchmark, revealing the test’s history, methodology, and why this achievement marks a major milestone for database performance in the big‑data era.
This article walks through building a Java‑based flash‑sale (秒杀) service, diagnosing overselling issues, and progressively enhancing it with optimistic locking, distributed rate limiting, Redis caching, and asynchronous Kafka processing to achieve higher throughput and data consistency under heavy concurrency.
The article details Minsheng Bank’s shift to distributed architecture, outlining regulatory drivers, business requirements, the adoption of sharding, cross‑center high‑availability, and new‑type distributed databases, and showcases performance results of SequoiaDB 3.0 across multiple high‑throughput banking scenarios.
This article explains why and how to parameterize LoadRunner scripts, covering the benefits, core challenges, step‑by‑step configuration, data construction methods, and the various selection and update options that control parameter value retrieval during test execution.
Alibaba’s fourth middleware performance challenge invites developers to create a high‑performance Service Mesh Agent for Dubbo, leveraging Docker containers, etcd registration, load balancing, protocol translation, and offering optimization tips such as coroutines, async communication, and caching to achieve superior QPS scores.
The article details Alibaba's fourth Middleware Performance Challenge, explaining the Dubbo‑based Service Mesh problem, the Docker‑driven test scenario, scoring methodology, and optimization suggestions such as coroutines, async communication, and caching, offering participants a comprehensive guide to excel in the competition.
This article explains the differences between stack and heap memory on Android devices, outlines typical memory problems such as high usage, leaks, and overflow, and demonstrates practical testing methods using the dumpsys meminfo command and the LeakCanary library with code examples.
This article explains the TCP TIME_WAIT state, its role in connection termination, the performance impact of large numbers of TIME_WAIT sockets, and practical optimization methods such as using keep‑alive and tuning Linux kernel parameters to mitigate the issue.
This article details the end‑to‑end testing strategy—including data consistency, interface logic, functional, data‑sync, and performance testing—used during the migration of product, order, and RMA services, highlighting automation scripts, Python decorators, and ThreadPool‑based load testing to ensure a smooth, zero‑downtime rollout.
The article reports on the inaugural JD Quality Testing Technical Forum, where experts shared advances in performance testing platforms, test anti‑patterns, container‑based test environments, system‑graded testing strategies, and intelligent tools for handling massive data.
This article outlines a comprehensive, step‑by‑step approach to performance testing and root‑cause analysis for backend services, covering environment validation, tool selection, Linux system limits, dependency checks, empty‑endpoint verification, throughput calculation, log monitoring, and essential Linux commands such as netstat, vmstat, mpstat, iostat, top and free.
This article explains how the Qudian team built a lightweight Koa‑based front‑back end separation framework called koa‑grace, covering its architecture, routing, data proxying, deployment strategies, performance benchmarks, and open‑source resources for developers seeking efficient web application development.
The article details how to identify and resolve memory leaks and performance bottlenecks in high‑concurrency Nginx workloads using tools such as Valgrind, AddressSanitizer, perf and flame‑graphs, while also sharing practical tips and personal reflections on debugging under pressure.
The article provides a comprehensive overview of Ceph, covering its object storage concepts, architecture, core technologies, performance testing results versus Swift, and concludes with a summary of findings and future outlook for this distributed storage solution.
This article provides a detailed overview of software testing fundamentals, covering testing stages across project lifecycles, classifications such as black‑box, white‑box, and gray‑box testing, functional and performance testing, manual versus automated testing, and specialized tests like smoke, regression, random, exploratory, and security testing.
The author shares a detailed personal journey from a novice software tester to a testing leader, describing how self‑learning automation tools like Selenium, Appium, and RobotFramework, adopting performance testing, establishing quality processes, and leading a team transformed both personal growth and product quality.
This guide explains how to set up JMeter’s PerfMon plugins for both GUI and non-GUI server performance monitoring, including required plugin downloads, agent installation, JAR placement, test plan configuration, command-line execution, and result analysis.
This guide outlines comprehensive compatibility, performance, and UI testing procedures for live‑streaming applications, covering device and network variations, CPU, memory, bandwidth, battery, temperature, latency, resolution, UI layout, navigation, and content validation to ensure a stable and user‑friendly streaming experience.
The article compares Flash and HTML5, explains Canvas 2D and WebGL rendering modes, and presents performance tests showing WebGL’s superior speed across browsers, while also discussing practical implications for web game development.
This article explains the background, installation, certificate generation, Elasticsearch and Kibana configuration, role mapping, performance testing, and final recommendations for securing an Elasticsearch 5.1.1 cluster with the open‑source Search Guard plugin, focusing on index‑level access control and RPC traffic encryption.
This guide explains how to use JMeter's command‑line mode and the JMeter Plugins CMD tool to convert JTL files into HTML reports and various chart types, detailing required parameters, plugin installation, and example commands for comprehensive performance‑testing reporting.
This article introduces the Chrome DevTools Protocol API, provides practical coding tips for using the chrome-remote-interface library, and demonstrates complete Node.js examples for gathering performance metrics, automating Baidu searches, and extracting first‑page results with headless Chrome.
This article traces the evolution of cryptography from ancient substitution ciphers to modern standards like DES, RSA, and ECC, explains symmetric and asymmetric encryption, compares ECC and RSA in security and efficiency, and presents cloud‑server performance tests showing ECC‑256’s superiority in speed and resource usage.
This article explains how JMeter implements its various distributed testing modes, examines the SampleSender interface and its overloads for standard, batch, and asynchronous modes, and presents optimization techniques that reduce network overhead and improve overall load‑testing performance.
A backend process that reports CPU and memory every 10 seconds missed occasional reports on Windows because the Sleep function oversleeps by up to one scheduler time slice, a problem solved by adjusting timing or using higher‑precision wait mechanisms such as Event.wait on Linux.
This article explains JMeter's distributed testing capabilities, compares standard, batch, and asynchronous modes, and shows performance results using a visual platform, highlighting how controller bottlenecks can affect load testing at scale.
The Fault‑Drill platform for hotel back‑end services combines real‑time traffic replication to shadow clusters with UI‑driven fault injection via a java‑agent, enabling developers to validate incident‑response plans, measure latency impacts, and reduce MTTR by testing normal and abnormal conditions on live traffic.
The 2017 Direct Banking App Evaluation, conducted by Sina Finance and Baidu MTC, objectively assesses 60 Chinese banks' mobile banking applications across technical performance, user experience, functionality, security, and growth dimensions using a detailed 58‑item scoring rubric.
The article presents a comprehensive evaluation of common image formats and compression algorithms—including Guetzli, WebP, MozJPEG, and Libjpeg‑Turbo—by comparing their compression ratios, quality metrics, encoding speed, and practical deployment considerations for large‑scale mobile services.
This article details the design, technology selection, architecture, encryption, compression, and performance evaluation of Ctrip's real-time user data collection system, which leverages Java, Netty, Kafka, and Avro to achieve high throughput, low latency, and robust fault tolerance for mobile and web applications.
The PCC (Performance Challenge Championship) was a one‑day offline competition where engineers built a high‑throughput “like” service, explored architectures such as OpenResty‑Lua, micro‑services with Go, caching strategies, and used Tsung for stress testing, with winners showcased and code released on GitHub.
This article explains how to design performance testing scenarios, covering test models, metrics, script preparation, concurrency calculations, pressure strategies, run times, delay settings, user termination, monitoring methods, and various typical scenario types such as baseline, load, mixed, capacity, large‑concurrency, stability and scalability tests.
This article details how to design, implement, and benchmark a single‑machine backend capable of handling up to 1 million concurrent connections and 60 000 queries per second while simulating the shake‑and‑send red‑packet workflow of a large‑scale messaging app, including capacity calculations, architecture choices, Go‑based implementation, and multi‑stage performance testing.
This article describes how to design, implement, and evaluate a backend system that simulates the Spring Festival red‑packet service, achieving up to 6 × 10⁴ QPS on a single server while handling one million concurrent connections, and discusses the hardware, software, architecture, and performance results.
This article chronicles Alibaba's evolution of the full‑link pressure testing platform—from its 2013 inception tackling massive Double 11 traffic, through data construction, isolation, traffic generation, and platform upgrades—to a mature, automated, cloud‑native solution that safeguards large‑scale e‑commerce stability.
This article provides a detailed tutorial on measuring and optimizing iOS app performance, covering startup and load times, memory and CPU usage, smoothness metrics, and practical code snippets using Apple Instruments and other tools to obtain reliable data across devices.
The article discusses practical solutions for long C++ build times, introduces a company‑wide compile cluster with a localbuild.sh script, outlines how developers can run automated tests, defines testing strategies, sets criteria for feature readiness, and presents approaches to performance testing and estimation adjustments within a CI‑driven development workflow.
This article explores Nginx’s built‑in and extended load‑balancing algorithms—including weighted round‑robin, IP‑hash, fair, generic‑hash and consistent‑hash—by dissecting source code, explaining implementation details, and presenting comprehensive performance tests that compare their balance, consistency, and fault‑tolerance across varied scenarios.
The article explains how LinkedIn’s Redliner system automatically measures service capacity and performs low‑impact, production‑traffic stress tests to identify bottlenecks, guide resource allocation, and support proactive capacity planning and performance regression testing.
The article presents a comprehensive, application‑centric approach to automated capacity management that analyzes why server utilization is low, defines safe usage thresholds, describes a load‑balancer‑driven stress‑testing workflow with regression modeling, and explains how this practice improves resource efficiency, cost savings, and developer‑ops collaboration.
This article explains how to automate application‑centric capacity assessment, identify the safe utilization thresholds, use load‑balancer‑driven stress testing and regression modeling to pinpoint resource bottlenecks, and improve server usage while maintaining service reliability through close DevOps collaboration.
Tencent Cloud’s migration team successfully moved over 130 on‑premise MySQL instances totaling 20 TB to CDB by rigorously testing the migration tool, configuring master‑slave replication, handling high‑concurrency binlog sync, and resolving timestamp, stored‑procedure, and DDL issues, ensuring seamless production deployment.
This article explains how to automate capacity management focused on application performance, identifies common causes of low resource utilization, proposes safe utilization thresholds, describes a testing framework that uses load‑balancer weighting and real‑time monitoring to pinpoint bottlenecks, and outlines how ops and developers can collaborate to improve efficiency.
This article details a performance comparison of MySQL 5.6.19 and Percona 5.6.25 with and without InnoDB deadlock detection, using sysbench to simulate flash‑sale stock‑decrement scenarios, and shows how turning off deadlock checks and enabling thread‑pool can boost throughput by over a hundred times.
This article demonstrates how to replace the classic linear search loops for arrays and strings with a sentinel‑based approach, reducing boundary checks and improving performance, especially on large datasets, and includes a concrete performance test.
This guide defines key IT terminology and presents a comprehensive methodology for estimating hardware platforms, storage devices, database and application servers, load‑balancing, and resource allocation using performance benchmarks such as TPC‑C, SPECweb2005, and SPECjbb2005, while outlining basic architectural principles and best‑practice deployment patterns.
This article outlines mobile performance testing methods—including startup time, page response time, basic metrics, Monkey testing, and image‑recognition techniques—reviews third‑party tools, and introduces the in‑house ETest solution with its features and future development plans.
This article provides a detailed, step‑by‑step guide for iOS developers on measuring and optimizing app performance across CPU usage, GPU‑related stutter, network conditions, and energy consumption using built‑in Instruments, Activity Monitor, Network Link Conditioner, and Charles.
This article explains key performance testing concepts such as system throughput, QPS/TPS, concurrency, response time, and their interrelationships, provides practical calculation examples, shows how daily PV relates to TPS for different platforms, and outlines essential metrics and formulas for evaluating software performance.
This article explains how to assess and improve iOS app performance by using Xcode's Instruments tool to measure resource consumption, startup time, and other metrics, and provides practical steps, UI commands, and configuration tips for accurate time profiling and code‑level optimization.
This article introduces JMeter as an open‑source Java‑based performance testing tool, explains its directory and code structure, describes how to compile and run it with Ant, and provides a step‑by‑step guide to extending JMeter with a custom JSON assertion and GUI component.
This article explains MySQL replication fundamentals, explores multi‑source master‑slave architectures, details step‑by‑step configuration for both basic 1‑master‑1‑slave and multi‑source setups, introduces parallel replication to boost performance, and presents Sysbench‑based testing results that demonstrate the efficiency gains.
This article introduces the Linux sar command, explains its syntax and common options, and demonstrates how to sample and interpret CPU, memory, I/O, process queue, and swap metrics to identify system bottlenecks during performance testing.
