This article provides a comprehensive, illustrated guide to TCP and UDP, covering the TCP header fields, flow‑control and congestion‑control mechanisms, the three‑way handshake and four‑way termination process, as well as a side‑by‑side comparison of the two protocols.
veRoCE is a ByteDance‑engineered RDMA protocol that retains RoCEv2 compatibility while adding multi‑path transmission, native out‑of‑order delivery, efficient loss detection and selective retransmission, allowing high‑performance clusters to operate over lossy Ethernet without PFC.
This article explains TCP and UDP fundamentals, detailing TCP header fields, flow‑control and congestion‑control mechanisms, the three‑way handshake and four‑way termination, UDP’s connectionless nature, and a side‑by‑side comparison of the two protocols.
This article traces the evolution from HTTP/1.1 to HTTP/3, explains how QUIC leverages UDP to achieve faster, secure connections, and details its handshake, connection migration, head‑of‑line blocking solutions, congestion and flow control mechanisms.
This article explains TCP’s flow control and congestion control mechanisms, detailing sliding window operation, zero‑window probing, SWS, window scaling, and the core congestion algorithms (slow start, congestion avoidance, fast retransmit/recovery), and clearly contrasts their purposes, scopes, and key variables.
This article explains the DCQCN end‑to‑end congestion control algorithm for RoCEv2, describes its three roles (RP, CP, NP), and provides detailed NS‑3 code implementations for ECN marking, CNP generation, and rate‑adjustment logic, including scheduling and parameter calculations.
This article provides a comprehensive English guide to the Transmission Control Protocol (TCP), covering its connection‑oriented design, reliability features, packet structure, three‑way handshake, data transfer process, flow and congestion control, four‑step termination, and example C++ socket code for establishing, sending, receiving, and closing TCP connections.
This article explores the hardware and networking foundations for training massive AI models, detailing the challenges of large‑scale RDMA deployment, the evolution of congestion‑control algorithms like DCQCN, TIMELY, HPCC, and AWS's SRD, and how hardware offload and programmable switches enable scalable, low‑latency AI infrastructure.
RoCE, originally built for InfiniBand, was adapted to Ethernet as RoCEv2, which uses IP/UDP headers to enable L3 routing but is highly sensitive to packet loss, requiring a lossless network and employing technologies such as PFC, ECN, DCQCN, and multi‑path transmission to maintain high RDMA performance.
This article analyzes the L4S (Low‑Latency, Low‑Loss, Scalable‑Throughput) internet service, its classic congestion‑control background, the L4S architecture—including host, network, and protocol support—and how 3GPP R18 integrates L4S into 5G systems for ultra‑responsive applications.
The article provides a comprehensive technical overview of HTTP's history, the design and advantages of the QUIC protocol—including connection establishment, migration, head‑of‑line blocking mitigation, congestion control, ACK handling, and stream/connection flow control—while contrasting it with TCP and HTTP/2 mechanisms.
This article explains the Simulcast standard in WebRTC, compares it with transcoding and SVC, describes how an SFU rewrites RTP headers for seamless layer switching, outlines congestion detection using TWCC, and presents automated bandwidth allocation strategies to optimize video quality and reduce bandwidth in large‑scale meetings.
This article explains how TCP ensures reliable data transfer through connection management, sequence and acknowledgment mechanisms, retransmission strategies, flow control, and congestion control, and includes C++ example code illustrating a simple client‑server file transfer implementation.
This tutorial series for newcomers explains TCP timeout retransmission, window management, and congestion control, combining theory with practical experiments using tcpdump, netstat, iptables, and Scapy, and demonstrates how Linux kernel parameters affect retransmission behavior and performance.
This article traces the evolution of HTTP from its 1991 origins through HTTP/2 and HTTP/3, explains how QUIC leverages UDP to reduce latency, and details QUIC's connection establishment, migration, head‑of‑line blocking mitigation, congestion control, and flow‑control mechanisms.
This article traces the evolution of HTTP from its 1991 inception through HTTP/1.1, HTTP/2, and HTTP/3, explains the QUIC protocol’s design and advantages, and details connection establishment, congestion control, flow control, and head‑of‑line blocking mitigation techniques used in modern web transport.
This article traces the evolution from HTTP/1.1 to HTTP/3, explains how QUIC built on UDP to reduce connection latency, avoid head‑of‑line blocking, support seamless connection migration, and introduce advanced congestion and flow‑control mechanisms that give HTTP/3 a performance edge over its predecessors.
This article explains the QoS mechanisms NACK and TWCC used in WebRTC, detailing their packet‑loss detection, retransmission strategies, congestion‑control algorithms, and how they improve real‑time audio/video communication while minimizing bandwidth waste.
This article provides a detailed technical analysis of Meta's SIGCOMM paper on RDMA over Ethernet for large‑scale AI training, examining the physical network deployment, congestion‑control mechanisms, topology choices, routing strategies, hardware design, and the practical challenges that remain.
This article details the deployment of QUIC in Trip.com’s mobile app, covering multi‑process architecture, containerized upgrades, service discovery, health monitoring, push‑pull resilience, full‑link tracing, congestion‑control algorithm redesign, and the resulting performance and reliability improvements achieved across global users.
The article thoroughly examines QUIC—the user‑space, UDP‑based transport underpinning HTTP/3—detailing its architecture, fast 0‑RTT handshakes, connection migration, stream‑level independence, congestion control, and flow‑control mechanisms, while also discussing implementation hurdles like 0‑RTT reliability, load‑balancer routing, UDP restrictions, CPU overhead, and real‑world performance gains reported by major Chinese tech firms.
This article explains the evolution from HTTP/1.1 to HTTP/3, detailing QUIC's UDP‑based design, its impact on connection establishment, migration, head‑of‑line blocking, multiplexing, congestion control, forward error correction, and flow‑control mechanisms for modern web services.
This article explains the evolution from HTTP/1.1 to HTTP/3, detailing QUIC’s UDP‑based design, its 1‑RTT handshake, connection migration, head‑of‑line blocking elimination, congestion‑control enhancements, forward error correction, and stream‑ and connection‑level flow control.
This comprehensive guide covers TCP fundamentals—including differences from UDP, the three‑way and four‑way handshakes, half‑open queues and SYN flood attacks, header fields, Fast Open, timestamps, RTO calculation, flow and congestion control, Nagle algorithm, delayed ACKs, and keep‑alive mechanisms—providing clear explanations and practical examples for engineers.
The article analyzes how CLOS spine‑leaf data‑center networks encounter congestion under many‑to‑one and all‑to‑all traffic patterns, explains the limitations of simply enlarging buffers, and details how ECN and PFC mechanisms can be tuned to achieve loss‑less, low‑latency operation.
The article provides a detailed technical analysis of RoCE (RDMA over Converged Ethernet), its two protocol versions, packet overhead, congestion‑control mechanisms, Soft‑RoCE implementation, and the challenges and performance implications of deploying RoCE in modern HPC environments compared to InfiniBand and traditional Ethernet solutions.
The article examines how InfiniBand’s specialized features—collective communication, in‑network computing, adaptive routing, congestion control, cut‑through forwarding, shallow buffering, and self‑healing—are optimized for large‑scale AI‑generated content (AIGC) training, delivering higher bandwidth, lower latency, and greater fault tolerance than Ethernet alternatives.
The talk examines the challenges of using standard Linux TCP for high‑performance data‑center workloads, explores how RDMA can provide zero‑copy and asynchronous kernel bypass, and presents experimental results from an FPGA‑based prototype that approaches 800 Gbps packet rates while highlighting congestion‑control and CPU‑utilization trade‑offs.
The article analyzes how TCP's three‑way handshake, slow‑start, and initial congestion window can introduce extra round‑trip times, causing service latency to far exceed network RTT, and shows practical packet captures, Linux defaults, and mitigation techniques.
The article analyzes how TCP's congestion control and slow‑start mechanisms introduce extra round‑trip times, causing service latency to far exceed network RTT, and explains how TCP connection setup, cwnd limits, and long‑lived connections affect overall response times.
This article explains the fundamentals of TCP and UDP, comparing their reliability, connection models, flow and congestion control, error detection, and detailing the three‑way handshake and four‑way termination processes to help readers choose the appropriate protocol for different applications.
This comprehensive guide explains TCP fundamentals—including the differences from UDP, the three‑way handshake, four‑way termination, half‑open queues, SYN‑Flood attacks, header fields, Fast Open, timestamps, retransmission timers, flow control, congestion control, Nagle’s algorithm, delayed ACKs, and keep‑alive—providing clear diagrams and practical insights for engineers.
This article explains the TCP/IP protocol suite, detailing its four layers—link, network, transport, and application—along with key protocols such as IP, TCP, UDP, ARP, ICMP, DNS, and the mechanisms of packet encapsulation, routing, flow and congestion control, and connection management.
This article provides a comprehensive overview of TCP and UDP, covering TCP header fields, flow and congestion control mechanisms, the three‑way handshake and four‑way termination, SYN‑flood attacks, UDP characteristics, and a detailed TCP/UDP feature comparison.
This article provides a comprehensive overview of TCP and UDP protocols, covering TCP header fields, flow and congestion control mechanisms, three-way handshake and four-way termination, reliability guarantees, as well as a detailed comparison with UDP, including connection orientation, broadcast support, and header overhead.
The paper introduces PACC, a perception‑aware congestion‑control algorithm that leverages a CNN‑based video‑quality sensor to adjust bitrate based on user‑perceived delay and quality trends, achieving 6.8%‑32.4% QoE improvements over existing model‑based, hybrid, and RL‑based schemes in diverse network conditions.
The article analyzes the SIGCOMM 2022 paper introducing Zhuge, a wireless‑AP‑only solution that separates congestion feedback from the data path to shrink the control loop, dramatically cutting tail latency and boosting real‑time communication performance.
This article examines the RoCE protocol—an RDMA‑enabled Ethernet technology—its evolution, technical details, congestion‑control mechanisms, performance comparisons with InfiniBand, practical deployment issues in HPC clusters, and real‑world case studies such as Slingshot and application benchmarks.
This comprehensive guide delves into TCP’s core mechanisms—including header fields, segmentation and reassembly, the three‑way handshake and four‑step termination, sliding‑window flow and congestion control, retransmission strategies, and connection design patterns—providing a solid foundation for network engineers and developers.
This comprehensive guide walks you through the TCP/IP protocol suite, covering the four-layer model, data link, network, transport, and application layers, and explains key protocols such as IP, ARP, ICMP, TCP, UDP, DNS, as well as connection management, flow control, and congestion control techniques.
This article explains the essential concepts of TCP and UDP, covering the OSI model layers, socket communication, reliable transmission mechanisms such as stop‑and‑wait and sliding‑window, congestion control, connection establishment, and practical differences between TCP and UDP, providing a comprehensive networking fundamentals overview.
HARP is Tencent Cloud's high-performance, highly available network transmission protocol that quickly reroutes around switch failures within 100 µs, offering zero packet loss, low latency, high bandwidth, scalable connections, and custom congestion control for storage, HPC, AI, and big data workloads.
This comprehensive guide explains the TCP/IP model’s four layers, key protocols such as IP, ARP, ICMP, DNS, and the differences between TCP and UDP, and details connection setup, termination, flow and congestion control mechanisms like slow start, fast retransmit, and fast recovery.
BBRv2 improves upon the original BBR algorithm by more accurately measuring delivery rate, constraining inflight data with loss and ECN signals, and introducing a dynamic state machine that enhances fairness and coexistence with Reno/CUBIC, addressing the multi‑flow fairness and queueing issues that plagued BBRv1.
This article provides a comprehensive, step‑by‑step explanation of TCP and related transport‑layer concepts, covering the OSI model, physical to application layers, socket and port mechanics, TCP header fields, reliable transmission techniques, flow and congestion control, connection setup/teardown, and a brief comparison with UDP.
This article explains TCP fundamentals—including connection-oriented design, byte‑stream transmission, reliability, header structure, flag meanings, the three‑way handshake, four‑way termination, SYN‑Flood attacks, time‑wait handling, retransmission strategies, kernel tuning parameters, Nagle’s algorithm, and congestion control mechanisms such as slow start and congestion avoidance.
While UDP is often assumed to be faster than TCP, this article explains socket basics, the reliability mechanisms of TCP, scenarios where UDP can be slower, and how application-layer solutions like KCP or QUIC add reliability, helping readers understand when each protocol truly excels.
The article surveys the evolution of live streaming toward low‑latency, interactive scenarios and details WebRTC‑based real‑time audio/video system design, covering RTP/UDP transport, FEC and ARQ loss recovery, congestion control, jitter buffering, echo cancellation, edge‑node path optimization, and a multi‑layer architecture with signaling, routing, mixing services for scalable, high‑availability PK deployments.
At the 2022 OCP Global Summit in San Jose, Alibaba and Broadcom presented two technical talks covering in‑band telemetry‑based network performance anomaly detection and the HPCC++ congestion‑control algorithm, highlighting deployment challenges, resource trade‑offs, and real‑world data‑center use cases.
This article explains TCP's retransmission mechanisms—including timeout retransmission, fast retransmit, SACK and D‑SACK—along with sliding window operation, window sizing, and congestion control algorithms such as slow start, congestion avoidance, and fast recovery, providing a comprehensive network fundamentals guide.
This article examines the RoCE protocol and its use in high‑performance computing, describing its low‑latency advantages, congestion‑control mechanisms, performance comparisons with InfiniBand, practical deployment issues, and real‑world case studies such as Slingshot and CESM/GROMACS benchmarks.
This article explains the evolution of HTTP from 1.1 to HTTP/3, how QUIC leverages UDP and TLS to achieve low‑latency 0‑RTT connections, solves head‑of‑line blocking, improves congestion and flow control, and why these advances matter for modern web applications.
This article explains the evolution of HTTP, the release of the HTTP/3 RFC, how QUIC leverages UDP to provide faster, secure connections with 0‑RTT handshakes, solves head‑of‑line blocking, improves congestion and flow control, and details its core mechanisms such as connection migration and ACK handling.
This comprehensive guide walks through the fundamentals of computer networking, detailing the OSI layers, the inner workings of TCP and UDP, reliable transmission techniques such as stop‑and‑wait, sliding windows, congestion control, connection setup and teardown, as well as security concerns, performance trade‑offs, and practical use‑case scenarios.
This article explains the design and operation of TCP BBR (Bottleneck Bandwidth and Round‑trip propagation time), its relationship to bufferbloat, the shortcomings of loss‑based congestion algorithms, and details the BBR phases, gain cycle, source code, advantages, disadvantages, and practical considerations for network engineers.
This article continues the previous discussion on TCP handshakes and termination, then thoroughly explains nine essential TCP mechanisms—including sliding window, flow control, congestion control, delayed and piggyback ACKs, sticky packet handling, and keep‑alive—illustrated with diagrams and practical examples.
The article analyzes how hyper‑converged data‑center networking unifies compute, storage, and HPC on a lossless Ethernet fabric, addresses AI‑era performance bottlenecks, compares RDMA over Ethernet with InfiniBand, and outlines the core metrics, value, and key technologies that enable zero‑loss, low‑latency, high‑throughput operation.
To meet 5G remote‑control demands of sub‑200 ms latency, sub‑0.2 % jitter and 20‑30 % loss, the article proposes jointly optimizing capture, transport, a reduced jitter buffer, a hybrid GCC‑plus‑SNR congestion‑control algorithm, and adaptive interleaved FEC to approach the 100 ms target.
The DAP system, developed by Kuaishou’s audio‑video network transmission team, won the ACM MM 2021 Grand Challenge by introducing deadline‑aware scheduling and a novel paired‑sending congestion control algorithm that dramatically improves QoE for real‑time video streams, outperforming Copa, BBR and DRL by up to 43.5%.
This article explains how QUIC solves TCP's head‑of‑line blocking, reduces connection‑setup overhead, enables seamless connection migration, lowers retransmission probability, and improves congestion control, offering a more efficient alternative to TCP for modern network environments.
This article provides a comprehensive overview of computer networking fundamentals, covering OSI and TCP/IP layer models, the functions of each layer, detailed comparisons between TCP and UDP, the three‑way handshake, four‑way termination, SYN‑flood attacks, congestion control, and practical mitigation techniques.
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.
This article provides a comprehensive overview of computer networking fundamentals, covering the essence of network communication, distributed packet‑switching, layered protocol design, congestion‑control mechanisms, the evolution of network technologies, and security considerations, all aimed at helping engineers and students build a solid networking foundation.
QUIC, a UDP‑based transport protocol that merges TCP reliability, UDP speed, and TLS security, underpins HTTP/3 by eliminating TCP’s three‑way handshake and head‑of‑line blocking, offering connection migration, 0‑RTT latency, customizable congestion control, fully encrypted packets, and is already being deployed in services such as Tencent Cloud CDN.
This comprehensive guide explains TCP fundamentals—including the differences between TCP and UDP, the three‑way handshake, four‑way termination, half‑open queues, SYN‑Flood attacks, header fields, Fast Open, timestamps, retransmission timeout calculation, flow and congestion control, Nagle's algorithm, delayed ACKs, and keep‑alive mechanisms—providing clear examples and diagrams for each concept.
This article provides a comprehensive overview of TCP fundamentals, comparing it with UDP, detailing the three‑way handshake and four‑way termination, explaining SYN flood attacks, describing the TCP header fields, fast open, timestamps, retransmission timeout calculations, flow and congestion control mechanisms, as well as Nagle's algorithm, delayed ACKs, and keep‑alive behavior.
This comprehensive guide explains TCP vs UDP, the three‑way handshake, four‑way termination, half‑open queues, SYN‑Flood attacks, header fields, timestamps, Fast Open, retransmission timeout calculations, flow control, congestion control, Nagle’s algorithm, delayed ACKs and keep‑alive mechanisms, providing essential knowledge for networking interviews and system design.
This comprehensive guide walks you through the TCP/IP protocol suite—covering the four‑layer model, key protocols like IP, ARP, ICMP, DNS, the mechanics of ping and traceroute, TCP/UDP differences, connection establishment and termination, flow control, and congestion control—providing the core knowledge needed for networking interviews and solidifying fundamental networking concepts.
This article provides a comprehensive introduction to the transport layer of computer networks, explaining the roles of TCP and UDP, socket APIs, port numbers, multiplexing, packet structures, connection management, three‑way handshake, four‑way termination, and congestion‑control mechanisms.
This article provides a detailed introduction to the TCP/IP model, explaining each layer—from the physical and data link layers up to the application layer—along with key protocols such as IP, ARP, ICMP, TCP, UDP, DNS, and the mechanisms of connection establishment, termination, flow control, and congestion control.
This comprehensive guide explains TCP fundamentals—including its differences from UDP, the three‑way handshake, four‑way termination, half‑open queues, SYN flood attacks, header fields, Fast Open, timestamps, retransmission timers, flow and congestion control, Nagle's algorithm, delayed ACKs, and keep‑alive mechanisms—providing clear diagrams and practical insights for developers.
This article provides an in‑depth overview of TCP, explaining its core differences from UDP, the three‑way handshake and four‑way termination processes, half‑open queues and SYN‑Flood attacks, Fast Open, timestamp usage, retransmission timeout calculation, flow and congestion control mechanisms, as well as Nagle’s algorithm, delayed acknowledgments and TCP keep‑alive.
This comprehensive guide explains the TCP/IP protocol suite, detailing each layer from the physical link to the application, and covers essential protocols such as IP, ARP, ICMP, DNS, as well as connection management techniques like three‑way handshake, four‑way termination, flow control, and congestion control mechanisms.
The article explains why TCP performance degrades in high‑loss, high‑latency environments, analyzing the impact of congestion‑control algorithms, the three‑way handshake overhead, and the retransmission strategy, and then outlines modern alternatives and optimizations to mitigate these issues.
This article explains how to design a low‑overhead, high‑performance lossless RoCE network for data centers, covering RDMA basics, mainstream network options, QoS, lossless and congestion‑control designs, buffer management, deadlock analysis, and practical tuning to achieve sub‑100 µs latency and near‑full bandwidth utilization.
This article provides an in‑depth overview of TCP congestion control, covering the sliding‑window flow control, the congestion control state machine with its five states, and the four core algorithms—slow start, congestion avoidance, fast retransmit, and fast recovery—plus practical insights for middleware design.
Tencent Cloud’s H5 voice‑communication solution improves QoE by applying a comprehensive WebRTC QoS framework—bandwidth estimation, adaptive congestion control, loss‑resilient FEC/NACK, latency and jitter mitigation, and global node optimization—while addressing interoperability challenges and continuously refining performance through monitoring, AB‑testing, and extensive bug fixes.
This article explains how the BBR congestion control algorithm, designed for long‑fat networks, overcomes the limitations of traditional TCP to dramatically improve cloud storage download speeds, detailing its principles, implementation steps, and real‑world performance gains.
This article explores TCP/IP fundamentals—including three‑way handshake, receive windows, congestion control mechanisms, and HTTP header limits—while demonstrating practical packet captures with tcpdump and building a simple WebSocket chat demo to illustrate protocol behavior and differences from WebRTC.
This article provides a comprehensive introduction to TCP, covering its header structure, optional fields, the role of MSS and TSO, mechanisms for reliability and retransmission, as well as detailed explanations of flow control, sliding windows, and congestion control algorithms.
