This article explains Intel Omni-Path Architecture and InfiniBand, describing their protocol stack, network topology, RDMA technology, and related protocols such as RoCE and iWARP, while also comparing their design goals and performance characteristics.
This article traces the history from isolated computers to modern networked systems, explains LAN, MAN, and WAN classifications, details the birth of ARPANET and TCP/IP, compares OSI and TCP/IP models, and discusses core concepts such as switching, delays, throughput, multicast, and physical media.
The article analyzes how the AI era’s demand for ultra‑low latency and high throughput exposes fundamental limits of the traditional TCP/IP stack, and explains why RDMA’s kernel‑bypass, zero‑copy design, and emerging congestion‑control algorithms are becoming the preferred network fabric for modern data‑center workloads.
By detailing Dewu App’s end‑to‑end IPv6 overhaul—including dual‑stack client integration, server‑side gateway adjustments, CDN and live‑streaming reconfiguration, gray‑scale rollout, and real‑time monitoring—the case study shows how the app reached a 94.6% IPv6 density, topping China’s 2021 app survey.
The article outlines the three major visions for future networks—Universal Connectivity, Compute‑Network Convergence, and Precise Networking—detailing their technical requirements, security and mobility challenges, and the design principles and architectural models needed to achieve them by 2030.
RegNet introduces a scalable design space defined by quantized linear functions, enabling flexible trade‑offs between accuracy, efficiency, and mobile deployment, and demonstrates superior performance compared with ResNet, EfficientNet, and other mobile‑optimized networks.
This article explains what load balancing is, why it’s needed, outlines its classifications—including DNS, hardware, and software methods—introduces Linux Virtual Server (LVS) and its four operating modes, and reviews common static and dynamic scheduling algorithms used to distribute traffic efficiently.
This article explains the fundamentals of load balancing, covering its purpose, vertical and horizontal scaling, various classifications such as DNS, IP, link‑layer and hybrid methods, common algorithms like round‑robin and weighted, as well as hardware solutions, providing a comprehensive guide for building scalable, high‑availability systems.
This article provides a comprehensive analysis of Kafka Broker's network and request‑handling architecture, tracing its evolution from a simple sequential model through multithreaded and event‑driven designs to a Reactor‑based NIO implementation, and offers concrete tuning recommendations for high‑concurrency deployments.
This article provides a comprehensive overview of load balancing technologies—including LVS, Nginx, and HAProxy—detailing their architectures, operating modes, advantages, and drawbacks, and explains how to choose and configure them for high‑performance, scalable backend systems.
The article examines the core requirements of future networks—enhanced deterministic capabilities, end‑to‑end extensibility, and a streamlined control plane—introducing precise connection, computing‑network integration, and inherent security as key technologies to meet diverse QoS demands.
This article provides a comprehensive overview of Linux Virtual Server (LVS), covering its basic concepts, three‑tier architecture, load‑balancing techniques, scheduling algorithms, high availability, reliability, and suitable deployment environments for building high‑performance, scalable server clusters.
This article explains how Software-Defined Networking (SDN) decouples control and forwarding planes, overcomes the limitations of traditional networks, outlines its layered architecture, compares classic and hardware‑centric SDN approaches, and discusses its advantages, relationship with NFV, future prospects, and key challenges.
The article explains what data centers are, why interconnecting multiple data centers is essential for modern cloud, AI, and big‑data workloads, and compares two practical DCI architectures—end‑to‑end VXLAN and Segment VXLAN—detailing their underlying VXLAN, EVPN, and underlay requirements.
The article explains how 5G mobile edge computing (MEC) can be deployed in wide‑area and local‑area scenarios, outlines the relevant ETSI and 3GPP standards, and discusses the security threats and protection measures associated with edge nodes and data handling.
The article analyzes how AI‑driven workloads, exploding storage and compute capabilities, and distributed architectures expose the limits of traditional three‑network data‑center designs, and explains why a lossless, hyper‑converged Ethernet network with zero‑loss, low‑latency, high‑throughput characteristics is becoming essential.
The article provides a systematic analysis of why modern data centers must evolve, detailing the rise of DPU and IPU technologies, their technical stages, advantages over traditional NICs, current challenges such as power and cost, and their pivotal role in edge and cloud computing futures.
The article examines the transition of Chinese government external networks from IPv4 to IPv6, highlighting address scarcity, operational complexities, and security concerns of IPv4, while detailing IPv6's abundant addressing, extensibility, security advantages, industry readiness, and the emerging IPv6+ innovations that support modern digital government services.
This article explains the fundamentals of Content Delivery Networks (CDN), their architecture, key functions, and principles, and then details Global Server Load Balancing (GSLB) techniques—including DNS‑based, HTTP‑redirect, and IP‑spoofing methods—along with server‑selection strategies and their advantages and drawbacks.
ByteDance’s data centers, including Douyin’s, operate with terabit‑level outbound bandwidth, housing up to hundreds of thousands of servers across China and abroad; the article breaks down server counts, total export bandwidth, CDN acceleration, and the network designs that enable billions of daily video streams.
The article provides a comprehensive technical overview of InfiniBand, covering its history, standards, architecture layers, packet format, performance advantages, and a detailed comparison with Ethernet, highlighting why it has become the preferred high‑speed interconnect for HPC workloads.
This article provides a comprehensive overview of data centers, covering definitions, classifications, cloud‑compatible designs, scale and availability tiers, lifecycle stages, energy consumption metrics, power and cooling technologies, network architectures, and emerging industry trends.
This article examines how UCloud’s rapidly expanding cloud services drove a series of backbone network upgrades, detailing the transition from the original MPLS‑based design to a Segment Routing (SR‑TE) architecture that delivers intelligent, reliable and programmable traffic engineering across global data‑center regions.
This article examines the rapid evolution of UCloud's data‑center network, detailing the challenges of scaling MAN and backbone infrastructures, the transition from Backbone 1.0 to 2.0, and how Segment Routing and SR‑TE enable intelligent, reliable, and programmable cloud connectivity.
Enterprise A consolidated its numerous VPC peering links and dedicated‑line gateways into a single Tencent Cloud Connect Network instance, enabling automatic route learning, ECMP load‑balancing across four 10 Gbps lines, boosting bandwidth utilization from about 50 % to 75 % while cutting operational complexity and enhancing cross‑region high availability.
This article provides a comprehensive overview of Multiprotocol Label Switching (MPLS), covering its origins, definition, working principles, architectural components, label operations, LSP creation, and deployment methods, illustrating how MPLS enhances routing efficiency, VPN services, traffic engineering, and QoS in modern networks.
This article explains how to partition data‑center networks, apply security‑level (等级保护) controls with firewalls and VRF, compares common three‑layer and flat topologies, and outlines future directions such as elasticity, simplicity, and openness for next‑generation data‑center networking.
5G's low latency, high bandwidth, and massive connectivity drive mobile edge computing, while merging IT and communication domains demands architecture, deployment, and scheduling integration; the article explains how cloud‑edge‑center convergence and unified QoS management unlock enhanced services and new business scenarios.
This article examines the rise of web‑application firewalls, outlines common deployment challenges, compares several WAF operating modes—including bypass, layer‑2 transparent, and proxy architectures—and proposes load‑balancing strategies to achieve secure, high‑availability web services.
Xianyu’s IPv6 migration, driven by China’s carrier upgrades, implements a dual‑stack architecture that uses NAT at the LVS layer to translate IPv6 to IPv4 while embedding the client’s IPv6 address in HTTP headers, enabling gradual gray‑release, comprehensive testing, and monitoring, resulting in about 55 % IPv6 daily active users.
This article provides a comprehensive overview of InfiniBand, covering its history, the InfiniBand Trade Association, architecture layers, packet format, switching mechanisms, performance advantages, and a detailed comparison with Ethernet for high‑performance computing environments.
This guide explains why Tencent Cloud’s Basic Network must be replaced by VPC for security isolation and details migration methods—including interconnection, CVM conversion, IP changes, image rebuilding, database VPC conversion, DTS, and load‑balancer adjustments—illustrated by a case study moving 600+ CVMs across 80 projects.
This article explains Cisco's evolution from traditional three‑tier data center designs to modern spine‑and‑leaf and Clos network architectures, highlighting the limitations of STP, the introduction of vPC, and the benefits of virtualization for high‑bandwidth, low‑latency server connectivity.
This article explains how InfiniBand’s RDMA‑based architecture, layered protocol stack, and Mellanox Socket Direct technology deliver far higher bandwidth, lower latency, and better CPU efficiency than traditional TCP/IP networks, and it presents performance test results that show up to an 80% latency reduction.
This article provides a comprehensive overview of InfiniBand technology, its high‑bandwidth low‑latency architecture, dominant market players such as Mellanox, and its critical impact on the performance and evolution of global supercomputing systems as reflected in TOP500 rankings.
The article analyzes the rapid evolution of data‑center networking from early gigabit links to the current 25G access + 100G interconnect architecture, examines IEEE 200G/400G standards, server and optical‑module trends, and concludes that 400G is poised to overtake 200G as the next dominant bandwidth solution.
The membership service now employs a multi‑region, multi‑data‑center high‑availability architecture that integrates DNS‑based load balancing, automatic failover for MySQL, Redis, and RocketMQ, cross‑region network backup, unified monitoring, and an operations platform, ensuring seamless traffic switching and fault recovery without user impact.
This article examines the evolution of network architectures from traditional tree‑topology back‑ends to peer‑to‑peer designs, explores pure and hybrid P2P models, and provides practical guidance for deploying and managing multi‑cloud P2P networks using Kubernetes, Flannel, DNS and platform‑based automation.
The 2019 Yunqi Conference Cloud Network Forum gathered over two hundred network enthusiasts to review a decade of Alibaba data‑center networking evolution, explore emerging technologies such as AI, big data, and programmable chips, and outline the next ten years of high‑performance, data‑centric cloud networking.
This article explains the principles of LVS load balancing, its benefits, the Hulk LVS cluster architecture, request hashing mechanisms, system‑level high availability, and the HULK LVS forwarding panel, providing a comprehensive guide for building resilient high‑traffic web services.
This article details the two‑year evolution of 58 Group’s container cloud platform, describing its background, overall architecture, network and service‑discovery advancements, operational challenges, and lessons learned from migrating over 90% of the company’s traffic to a Docker‑Kubernetes‑based cloud environment.
This article introduces Linux Virtual Server (LVS), explains its three‑tier architecture, details its main components and three load‑balancing methods (NAT, IP Tunneling, Direct Routing), compares their performance, and reviews the eight scheduling algorithms used for traffic distribution.
This presentation details NetEase Games' journey of moving its mobile titles overseas onto AWS, covering the challenges of performance, elasticity, security and global connectivity, the systematic cloud evaluation framework they built, and the architectural and operational practices they adopted to achieve reliable, cost‑effective worldwide game services.
This article details Ctrip’s multi‑generation cloud networking solutions, tracing the progression from early VLAN‑based Layer‑2 designs through SDN‑enabled large‑scale networks to container‑centric and cloud‑native architectures, highlighting hardware topologies, software integrations, and operational lessons for large‑scale data‑center environments.
This article outlines the three main load‑balancing methods—DNS, hardware, and software—detailing their advantages and drawbacks, then explains common algorithms such as round‑robin, weighted round‑robin, least‑connections, performance‑based, and hash, and provides guidance on combining them for optimal architecture.
Load balancing distributes incoming traffic across multiple servers to improve performance, ensure high availability, and enable scalability, and this article explains its core principles, various classifications such as DNS, IP, layer‑2 and hybrid methods, and common algorithms like round‑robin, least connections, hash and weighted distribution.
This article explains server clustering and load‑balancing concepts, detailing the architecture and operation of LVS, Nginx, and HAProxy, and compares their advantages, disadvantages, and typical deployment scenarios; it also discusses NAT and DR modes, load‑balancer placement, and best‑practice recommendations for different traffic volumes.
This article provides a comprehensive overview of Remote Direct Memory Access (RDMA), covering its background, limitations of traditional TCP/IP, related technologies such as TOE, U-Net, VIA, and detailed explanations of RDMA concepts, hardware implementations, verbs, and communication workflows.
This article provides a detailed overview of Internet Data Center (IDC) design, covering classification levels, functional zones, network architecture, server and storage systems, power and cooling infrastructure, fire protection, security, and green building practices for reliable, uninterrupted operation.
InfiniBand is a high‑performance network protocol that uses credit‑based flow control and switched fabric architecture to provide low latency, high bandwidth, and reliable data transfer, offering advantages over TCP/IP such as reduced packet loss, efficient RDMA, and support for various upper‑layer protocols.
This article provides a comprehensive overview of OpenFabrics Enterprise Distribution (OFED), its history, component stack, and the layered InfiniBand software architecture, explaining how various protocols such as IPoIB, SDP, and iSER enable high‑performance, low‑latency networking for Linux and Windows applications.
This article explains the fundamentals of load balancing, compares DNS‑based, hardware, and software solutions, and details common balancing algorithms such as round‑robin, load‑aware, response‑time, and hash strategies, helping architects choose the right approach for high‑performance systems.
The article outlines Qunar's data‑center network evolution, describing the limitations of traditional STP‑based designs, the adoption of vPC for active‑active redundancy, the transition to leaf‑spine topology for scalability, and the implementation of VXLAN to support large‑scale multi‑tenant cloud environments.
The article explains how Kafka's network layer uses the Reactor pattern with Java NIO, describes the single‑threaded selector model, its limitations, and how Kafka introduces multiple selectors, acceptor, processor, and handler thread pools to achieve high‑concurrency network I/O.
This article uses a vivid village allegory to illustrate how Software Defined Networking (SDN) separates control and forwarding, centralizes management, and improves efficiency, while also explaining SDN's core concepts, benefits, and its emerging role in future network architectures.
This article chronicles Tencent's network journey from its modest 2004 infrastructure through rapid expansion, critical incidents, and architectural breakthroughs like SET zones, SDN, and MPLS VPN, illustrating how the company transformed its data‑center operations to support massive user growth.
This article chronicles Tencent’s network infrastructure development over the past two decades, detailing early IDC setups, major outages, rapid expansion during the gaming boom, the introduction of SET architecture, large‑scale data centers, SDN and DCI implementations, and ongoing challenges as the company embraces mobile and cloud services.
This article explains the concept of load balancing, compares HTTP redirection, DNS load balancing, reverse proxy, and various LVS methods (NAT, Direct Routing, IP Tunnel), discusses their advantages, limitations, configuration steps, and practical considerations for building scalable server architectures.
This article analyzes the shortcomings of traditional TCP/IP networking, explains the principles of Software‑Defined Networking and OpenFlow, examines flow‑table evolution, and explores overlay technologies like VXLAN, while discussing the ongoing industry debate between revolutionary and incremental approaches.
This article details JD.com's multi-active distributed architecture, covering its evolution from single‑data‑center to multi‑region deployments, network design, leaf‑spine topology, data consistency mechanisms, application scheduling, monitoring, and disaster recovery strategies that enhance high availability and user experience.
FCoE (Fibre Channel over Ethernet) encapsulates Fibre Channel frames within enhanced Ethernet, enabling storage networking over a unified physical medium while preserving FC’s upper‑layer features, and the article details its protocol stack, differences from FC, VLAN discovery, virtual link establishment, and required Ethernet enhancements.
The article provides a detailed analysis of Intel’s Omni-Path and InfiniBand technologies, covering their histories, architectural differences, product lineups, performance benchmarks, and market positioning within high‑performance computing; it also examines the role of the InfiniBand Trade Association, the impact of acquisitions by Intel and Mellanox, and the future prospects of both interconnect solutions.
The article explains how hybrid cloud adoption demands on‑demand, elastic networking and argues that Software‑Defined Networking (SDN) provides the scalability, multi‑vendor interoperability, reliability and intelligence needed to seamlessly connect private and public clouds.
An in‑depth look at Vancl’s massive IT infrastructure—covering network layers, DNS/DHCP design, Windows file services, email systems, instant messaging, AD account management, SCCM automation, and virtualization—offering practical insights for building reliable, scalable operations in large enterprises.
This article provides a detailed overview of China Telecom’s extensive IDC infrastructure, explaining the hierarchy of its backbone and metropolitan networks, the roles of ChinaNet and CN2, BGP routing practices, inter‑carrier connections, and practical traffic‑optimization techniques used by both major and smaller operators.
The article explains Doom 3's deterministic client‑server network architecture, describing how player inputs are sent to the server, how compressed state snapshots are returned at 10‑20 Hz, and the UDP‑based reliable/unreliable messaging and compression methods used to keep the game synchronized.
The article provides a comprehensive overview of various backup software networking models—including LAN Base, LAN Free, Server Free, NDMP, and the emerging Server‑Less approach—detailing their deployment options, performance impacts, and suitability for different data volumes and infrastructure requirements.
This article explains how load balancing solves high‑traffic, high‑concurrency challenges by distributing requests across multiple servers, covering its core principles, vertical and horizontal scaling, various classification methods such as DNS, IP, link‑layer, and hybrid approaches, along with their advantages, drawbacks, and deployment recommendations.
This article explains the fundamentals of Linux Virtual Server (LVS), covering its architecture, kernel model, packet‑forwarding modes (NAT, DR, TUN), and both static and dynamic scheduling algorithms to help readers design scalable load‑balancing solutions.
This article explains Software Defined Networking (SDN) as a software‑driven overhaul of traditional IP networks, highlighting its ability to separate control and forwarding, simplify management, accelerate service deployment, and transform networks into programmable platforms akin to an Android system for networking.
This article explains how to redesign an iOS app's network layer to fully exploit HTTP/2 by switching from NSURLConnection to NSURLSession, discusses centralized versus distributed API designs, data handling strategies, security with SSL pinning, and provides concrete code examples and batch request techniques.
The article chronicles UPYUN's cloud CDN development from its modest 30‑40 node deployment in 2010, through multiple architectural revisions, scaling to over 130 edge nodes by 2014, and outlines future plans for DNS‑based load balancing, international backbone expansion, and video/live‑stream acceleration up to 2016.
This article details Meituan Cloud's evolution from a private, self‑developed cloud platform to a high‑performance public cloud, covering the motivations, architectural upgrades across physical links, host networking, gateways, controllers, and the adoption of overlay technologies like VXLAN to achieve industry‑leading scalability and reliability.
This article explains data‑center wiring zones, LAN, SAN and KVM network architectures, common distributed designs, modular POD concepts, and Top‑of‑Rack (ToR) switching, providing detailed guidance on cabling, equipment placement, and scalability for modern data‑center operations.
The article outlines a real‑world scenario where a small company's database was directly reachable from the Internet, then presents a redesign that inserts an intermediate database, restricts firewall access, and enforces strict credential management to protect sensitive data.
This article examines the recent price cuts in major CDN services, debates whether Xunlei’s unlimited‑node CDN represents a genuine technological breakthrough or merely market disruption, and offers practical guidance on selecting and deploying CDN solutions for different business needs.
