This article explains why Nginx HTTPS latency matters for instant search, describes how TLS handshake round‑trips affect response time, and provides step‑by‑step Nginx TLS configuration tweaks—including HTTP/2, cipher ordering, OCSP stapling, buffer sizing, and session caching—that together reduced real‑world latency by about 30%.
Redis can experience latency spikes due to factors such as complex commands, big keys, expiration patterns, memory limits, fork overhead, AOF persistence, CPU binding, swap usage, memory fragmentation, and network saturation, and this article provides a systematic troubleshooting methodology and practical optimization solutions for each scenario.
This article explains how to use Redis INFO and related commands to monitor critical performance metrics such as memory usage, memory fragmentation, total commands processed, latency, and key eviction, and provides practical tips for interpreting these metrics and optimizing Redis deployments.
This article explains the concept of latency, compares human and machine reaction times, breaks down cloud network latency components, quantifies the latency requirements of various telecom and cloud‑based applications—including gaming, video, industrial IoT and 5G—and outlines how different deployment architectures and measurement tools can be used to meet those requirements.
This article explains why Redis latency can suddenly increase, covering high‑complexity commands, large keys, concentrated expirations, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network overload, and provides concrete diagnostic commands and mitigation strategies.
This article examines common causes of increased latency in Redis—including high‑complexity commands, large keys, concentrated expirations, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network saturation—and provides practical monitoring and configuration techniques to diagnose and reduce delays.
This article explains the typical reasons behind Redis latency spikes—such as complex commands, big keys, concentrated expirations, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network overload—and provides practical steps and monitoring techniques to identify and resolve each problem.
This article traces the history of computer networking from early batch processing and time‑sharing systems through the emergence of LAN, WAN, TCP/IP, OSI model, and modern protocols, while explaining key concepts such as packet switching, routing, delay types, and physical media.
This article explains common causes of Redis latency spikes, such as high‑complexity commands, large keys, concentrated expirations, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network saturation, and provides step‑by‑step troubleshooting commands and best‑practice recommendations.
This article examines the typical reasons Redis latency spikes—such as high‑complexity commands, big keys, concentrated expirations, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network saturation—and provides practical diagnostics and mitigation steps to keep Redis responsive.
This article explains the typical reasons for Redis latency spikes—including complex commands, large keys, expiration bursts, memory limits, fork overhead, AOF settings, swap usage, and network saturation—and provides practical steps and commands to identify and mitigate each issue.
This article explains why Redis can become slow, covering typical latency causes such as high‑complexity commands, large keys, concentrated expirations, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network saturation, and provides practical troubleshooting steps and best‑practice recommendations.
This article explains why Redis latency can suddenly increase—covering high‑complexity commands, large keys, concentrated expirations, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network saturation—and provides practical diagnostic steps and mitigation techniques.
This article explains why a Redis instance may experience sudden latency spikes, walks through diagnosing the problem with slow‑log and big‑key scans, and provides concrete configuration tweaks, command‑level best practices, and operational guidelines to keep Redis performing at peak speed.
This article examines typical Redis latency spikes—such as complex O(n) commands, large keys, concentrated expirations, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network saturation—explaining how to detect, monitor, and mitigate each issue to maintain high‑performance operation.
This article explains why a high‑performance Redis instance can suddenly experience high latency, covering complex commands, large keys, expiration bursts, memory limits, fork overhead, CPU binding, AOF settings, swap usage, and network saturation, and provides practical troubleshooting and mitigation steps.
This article examines the challenges of existing stream‑processing benchmarks, introduces the open‑source Nexmark framework designed for reproducible, comprehensive performance testing, describes its metrics, query set, workload configurability, and presents experimental results on Flink, highlighting its role in advancing big‑data stream benchmarking.
This article explains how latency is used as a key indicator for application risk identification, defines slow interfaces, describes why percentile‑based thresholds are preferred over averages, and outlines the architecture, task workflow, and practical optimization strategies for a full‑chain monitoring system in a microservice environment.
This article compares monolithic applications and microservice architectures across dimensions such as network latency, development and operational complexity, reliability, resource consumption, scaling precision, throughput, deployment speed, and team communication, highlighting where each approach wins and offering guidance on when to adopt microservices.
This article explains the evolution of mobile network generations, compares their theoretical and typical download speeds, examines the impact on streaming and latency, and provides practical guidance on how these speeds affect everyday smartphone usage.
This article explains why MySQL master‑slave replication can suffer from latency, examines the root causes on both master and slave sides, and provides practical architectural, hardware, and configuration solutions—including semi‑sync and parallel replication—to improve data consistency and performance.
This article explains how to use HBase's built‑in PerformanceEvaluation tool to run baseline read/write latency tests (P99 and P999), describes key command‑line parameters, presents benchmark results for random and sequential operations, and discusses the implications for HBase performance tuning.
This article presents a series of C‑based experiments that measure how sequential and random memory accesses behave across different array sizes and strides, revealing that random accesses can be up to four times slower than sequential ones due to cache and memory I/O effects.
The article explains how the rapid growth of IoT data drives enterprises to combine edge computing with cloud services to reduce latency, improve scalability, and enable smarter factories and buildings, while outlining the challenges, use cases, and practical steps for designing effective edge‑cloud solutions.
This article explains how MongoDB's nearest mode achieves proximity‑aware reads across multiple data centers by analyzing the internal mongos and driver code, detailing latency collection, smoothing algorithms, node selection logic, and providing configuration recommendations for latency‑sensitive workloads.
Ren Zhengfei reveals Huawei’s HarmonyOS aims beyond smartphones, targeting IoT, autonomous driving and data centers with sub‑5 ms latency, claiming up to 60 % speed gains over Android, while acknowledging ecosystem challenges and outlining 5G rollout and US sanction updates.
This article explains the concepts, costs, architectural patterns, and practical design techniques for building multi‑active systems across different geographic locations while managing latency, data consistency, and business continuity.
Modern cloud outages reveal that cold‑standby or simple multi‑cloud promises often provide only psychological comfort; achieving true high availability requires active‑active designs with local traffic handling, data partitioning, and low‑latency synchronization, while balancing cost, complexity, and physical distance constraints.
The article recounts a pre‑promotion MySQL replication incident at JD.com, explains the master‑slave architecture and thread roles, identifies slow‑SQL and missing indexes as root causes of replication lag, and proposes practical measures to reduce latency and improve system stability.
A colleague noticed a 15‑second pause between the first two ping replies to pop3.sina.net, prompting a deep dive with tcpdump, strace, and DNS tests that revealed the delay was caused by slow local DNS resolution rather than network latency.
This article explains the main factors that cause Redis latency—including network bottlenecks, Redis’s single‑threaded design, deployment choices, OS settings, and slow commands—and provides concrete commands and best‑practice solutions to diagnose and reduce the delay.
This article examines the five critical demands—latency/determinism, local interactivity, data/bandwidth, privacy/security, and limited autonomy—that will dictate the evolution of edge computing, offering strategic recommendations for I&O leaders and outlining how these forces will shape future edge solutions.
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.
The article examines blockchain's future by highlighting three core technical constraints—high latency, costly computation, and limited storage—that must be overcome for the technology to move beyond speculation and achieve widespread practical use.
The study benchmarks Apache Flink against Apache Storm on a shared cluster, showing Flink delivering three‑to‑five times higher throughput and roughly half the latency across simple, sleep‑induced, and windowed workloads, with modest throughput loss for exactly‑once semantics, leading to a recommendation of Flink for high‑performance, stateful real‑time stream processing.
This article explains how to evaluate SSD performance by processing raw fio benchmark logs, interpreting bandwidth, IOPS, and latency metrics, converting logs to CSV, visualizing results with charts, and drawing practical conclusions for storage selection and reliability.
A detailed benchmark compares Redis 3.0.7 and Hazelcast 3.6.1 on a four‑node HP ProLiant cluster, showing Redis excels with few clients while Hazelcast scales better under high concurrency, highlighting trade‑offs for developers and architects.
This article presents a comprehensive Kafka benchmark using six machines to evaluate producer and consumer throughput, replication effects, message size impact, and end‑to‑end latency, providing detailed results, analysis, and reproducible test commands.
This report ranks Beijing data centers by network quality, using over 27 million ping samples to evaluate packet loss and latency, and explains the testing methodology, key terminology, and the resulting comprehensive ranking of 40 facilities.
This article explains the core concepts of system performance, defines throughput and latency, describes how to measure them, and provides detailed, code‑level techniques for locating and eliminating common performance bottlenecks across operating systems, networks, databases, and application code.
This interview with Alibaba's Lin Hao (aka Bi Xuan) explains why the company pursued a multi‑active, geographically distributed data‑center architecture, how it was deployed for the 2014 Double‑11 event, and the technical challenges—including latency, routing consistency, and data synchronization—that had to be solved to achieve high availability and low‑cost disaster recovery.
This article explains what Traceroute is, its limitations, the underlying packet‑TTL mechanism, how DNS reverse‑lookup enriches hop information, and how to interpret serial, queue and propagation delays to diagnose network performance problems more accurately.
