The Evolution of High Performance Networks: A Storage Perspective and Practices

on YouTube ↗18m

" summary

Storage network stack evolution at Alibaba Cloud moves from kernel TCP to Luna user-space TCP for front-end and RDMA for backend, then to a UDP-based Solar protocol plus hardware offload on bare-metal cipu. Key claims: RDMA cuts one-way latency to ~2 ms but degrades past 10,000 connections; Luna cuts front-end 95p latency by ~80%; Solar reduces storage-agent latency by ~40% and breaks cipu PCIe bottlenecks measured in production at thousands of hosts. Deployment takes years and extensive grayscale testing.

AI-written summary · about

" chapters (6)

1. 0:00Introduction
2. 0:50Product Portfolio
3. 1:51Storage Services
4. 3:05Gen 1 to Gen 2
5. 9:12Gen 2 to Gen 3
6. 13:26Evaluation Lessons

" related talks

• 15mMore Than Capacity: Performance-oriented Evolution of Pangu in AlibabaChao Han, Qiang Li, Qiao Xiang, Jiesheng Wu, Jiwu Shu, Jiaji Zhu, Yuanyuan Dong, Zhaosheng Zhu, Yaohui Wu, Shaozong Liu, Zhongjie Wu, Junping Wu, Zheng Cao, Jinbo Wu, Lulu Chen, Shuo Huang, Wenhui Yao, Shanyang Liu, Huayong Wang, Haonan Qiu, Derui Liu, Zhiwu Wu, Zicheng Luo, Yuchao Shao, Gexiao Tian, Yuxin Wang, Haohao Song, Ridi Wen, Shuqi Zhao · FAST '23: 21st USENIX Conference on File and Storage Technologies
• 19mDeploying User-space TCP at Cloud Scale with LUNAWei Lin, Yifan Shen, Shu Ma, Erci Xu, Lingjun Zhu, Jiesheng Wu, Haonan Wu, Bo Shi, Ting Fu, Shuguang Chen, Zhongyu Wang, Xingyu Liao, Zhendan Yang, Zhongqing Chen, Yijun Hou, Rong Liu, Chao Shi, Jiaji Zhu · USENIX ATC '23
• 05mKeynote: Innovation of the Network Infrastructure at AlibabaHaiyong Wang · Open Networking Summit North America 2019
• 27mPredictable NetworkDennis Cai · Networking @Scale Fall 2022
• 20mEmpowering Azure Storage with RDMAJeff Cohen, Xin Liu, Ankit Kumar, Albert Greenberg, Ivan Lee, Jitendra Padhye, Aaron Ogus, Manish Gupta, Rohan Mehta, Yuemin Lu, Wei Bai, Mahmoud Elhaddad, Shachar Raindel, Shanim Sainul Abdeen, Ankit Agrawal, Krishan Kumar Attre, Paramvir Bahl, Ameya Bhagat, Gowri Bhaskara, Tanya Brokhman, Lei Cao, Ahmad Cheema, Rebecca Chow, Vivek Ette, Igal Figlin, Daniel Firestone, Mathew George, Ilya German, Lakhmeet Ghai, Eric Green, Randy Haagens, Matthew Hendel, Ridwan Howlader, Neetha John, Julia Johnstone, Tom Jolly, Greg Kramer, David Kruse, Erica Lan, Avi Levy, Marina Lipshteyn, Chen Liu, Guohan Lu, Xiakun Lu, Vadim Makhervaks, Ulad Malashanka, David A. Maltz, Ilias Marinos, Sharda Murthi, Anup Namdhari, Madhav Pandya, Douglas Phillips, Adrian Power, Suraj Puri, Jordan Rhee, Anthony Russo, Maneesh Sah, Ali Sheriff, Chris Sparacino, Ashutosh Srivastava, Weixiang Sun, Nick Swanson, Fuhou Tian, Lukasz Tomczyk, Vamsi Vadlamuri, Alec Wolman, Ying Xie, Joyce Yom, Lihua Yuan, Yanzhao Zhang, Brian Zill · NSDI 2023
• 11mWhen Cloud Storage Meets RDMAHaiyong Wang, Dennis Cai, Bo Li, Qiang Li, Chen Tian, Pengcheng Zhang, Jiesheng Wu, Fei Feng, Jiaji Zhu, Yixiao Gao, Lingbo Tang, Yongqing Xi, Wenwen Peng, Yaohui Wu, Shaozong Liu, Lei Yan, Yan Zhuang, Fan Liu, Pan Liu, Xingkui Liu, Zhongjie Wu, Junping Wu, Zheng Cao, Jinbo Wu · NSDI ’21
• 19mEnhancement and Optimization of computing, storage & network for bare metal with hardware offloadingYajun Yang, Yunxiang Tao, yin rong · OpenInfra Summit 2020
• 31mLive Q+A 1Ying Zhang · Networking @Scale 2024