0 views collected around this technical thread.
This article explains the fundamentals of Layer‑4 and Layer‑7 load balancing, compares their advantages and disadvantages, and describes how a hybrid approach can combine high‑performance traffic handling with flexible application‑level routing for large‑scale systems.
The article explains how distributed AI training performance depends on reducing inter‑card communication latency, introduces RDMA technology and its implementations (InfiniBand, RoCEv2, iWARP), compares their latency and scalability against traditional TCP/IP, and outlines the hardware components and trade‑offs of InfiniBand and RoCEv2 networks.
This article explains the concepts of Nginx forward and reverse proxy, compares their roles, client and server perceptions, typical usage scenarios, and highlights why reverse proxy is a core Nginx feature while forward proxy is rarely used without third‑party modules.
iQIYI replaced its limited private‑cloud DNS by a hybrid‑cloud Anycast solution that uses an Enterprise Router with BGP‑advertised Anycast IPs and GRE tunnels to integrate public‑cloud DNS nodes, delivering unified, elastic domain resolution, rapid failover, and centralized management across IDC and cloud environments.
The article explains the emerging “new‑quality internet” (Net5.5G) concept, its AI‑driven motivations, the four‑link scenarios, four‑new technologies, and recent industry progress, illustrating how this next‑generation network architecture will support the AI era.
At the 2024 Wuzhen Summit, Alibaba Cloud R&D Vice President Cai Dezhi discussed the convergence of AI and next‑generation internet, outlining the “Network for AI” and “AI for Network” concepts, the HPN7.0 high‑performance network, AI‑driven operations, and the importance of open standards and protocol innovation to lower costs and enable widespread AI adoption.
This article provides an in‑depth explanation of computing power networks, covering their definition, key characteristics such as elasticity, agility and security, underlying technologies like SRv6 and network slicing, architectural components including the brain, and emerging trends spanning AI, big data, optical networking and blockchain.
At the CCF ChinaNet conference, Alibaba Cloud’s VP of R&D presented a vision of AI compute scaling to ten‑fold larger clusters, highlighting the shift from InfiniBand to high‑throughput Ethernet, the HPN7.0 architecture, emerging Scale‑up challenges, and the roadmap for high‑throughput Ethernet and the ENode+ super‑node system.
Although IPv6, standardized in 1998, offers a 128‑bit address space, its global rollout remains sluggish—only about 40 % of users have IPv6, early tunneling methods proved unreliable, and NAT and CDNs lessen urgency—so experts predict a dual‑stack Internet persisting until roughly 2045, with a gradual shift toward a name‑centric architecture rather than a rapid replacement of IPv4.
This article explains the growing demand for high‑performance computing, introduces NVIDIA's Quantum‑2 InfiniBand platform with its high‑speed, low‑latency capabilities, provides a curated list of related technical articles, and offers an extensive Q&A covering compatibility, cabling, UFM, PCIe limits, and best‑practice deployment for AI and HPC workloads.
The article introduces the concept of an egress gateway, explains why iQIYI’s member‑cooperation services require a unified outbound gateway to handle IP‑whitelisting and external access, critiques the limitations of an Nginx proxy, and describes adopting APISIX with security, traffic‑control plugins and a UI‑driven URL translation layer.
This article examines the two dominant AI compute network architectures, InfiniBand and RoCEv2, detailing their designs, flow‑control mechanisms, performance, cost and scalability characteristics, and evaluates their respective advantages and limitations to guide network selection for AI data centers.
The article explains how RDMA technologies—including InfiniBand, RoCE, and iWARP—provide high‑throughput, low‑latency, CPU‑free data transfer for massive generative AI model training, compares their architectures, and discusses modern network designs and load‑balancing strategies to optimize AI‑focused data‑center networks.
This article provides a detailed comparison of InfiniBand and RoCEv2 network architectures, examining their technical features, flow‑control mechanisms, performance, cost, and suitability for AI compute environments to guide designers in selecting the optimal solution.
The article analyzes traditional cloud data‑center networking challenges for AI workloads and compares two‑layer and three‑layer fat‑tree architectures, presenting high‑bandwidth, non‑blocking, and low‑latency designs such as AI‑Pool networks and offering practical deployment scales from hundreds to tens of thousands of GPUs.
Bilibili’s hybrid‑cloud architecture combines multiple public clouds and its own data centers using a star‑centered topology, employing VPCs, load balancers, NAT gateways, and dedicated lines managed via a multi‑cloud platform, while outlining project‑network setup, security rules, and proxy‑forwarding solutions for reliable, cost‑effective cloud networking.
Baidu built a high‑availability, low‑latency distributed transmission system that connects directly to global securities exchanges via dedicated optical lines and multicast support, synchronizes data across multiple cloud regions, provides Internet fallback for disaster recovery, and delivers sub‑60 ms market data to billions of daily financial service requests with five‑nine reliability.
This article provides a comprehensive overview of network switches, covering global market trends, definitions, OSI‑based classifications, hardware components, various switching architectures such as CLOS and leaf‑spine, performance metrics, differences from hubs and routers, industrial switch requirements, and the overall industry chain.
In a detailed interview at the 2023 Wuzhen Internet Forum, Alibaba Cloud’s infrastructure lead Cai Dezhong outlines the three‑phase IPv6 rollout, highlights organizational and technical innovations, stresses the need for global cooperation, and explains how IPv6 underpins the next generation AI infrastructure and predictable high‑performance networking.
This article provides a comprehensive overview of Kubernetes network architecture, including CNI plugins, network policies, service discovery, and ingress controllers, with practical configuration examples.