{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,12]],"date-time":"2026-03-12T12:17:46Z","timestamp":1773317866546,"version":"3.50.1"},"publisher-location":"New York, NY, USA","reference-count":53,"publisher":"ACM","funder":[{"name":"ERC PSAP","award":["101002047"],"award-info":[{"award-number":["101002047"]}]},{"name":"Swiss State Secreteriat for Education, Research and Innnovation","award":["SwissTwins"],"award-info":[{"award-number":["SwissTwins"]}]},{"name":"WeatherGenerator","award":["101187947"],"award-info":[{"award-number":["101187947"]}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2025,11,16]]},"DOI":"10.1145\/3712285.3759902","type":"proceedings-article","created":{"date-parts":[[2025,11,12]],"date-time":"2025-11-12T16:04:47Z","timestamp":1762963487000},"page":"1223-1239","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["SDR-RDMA: Software-Defined Reliability Architecture for Planetary Scale RDMA Communication"],"prefix":"10.1145","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0862-4662","authenticated-orcid":false,"given":"Mikhail","family":"Khalilov","sequence":"first","affiliation":[{"name":"ETH Z\u00fcrich, Zurich, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-0061-5072","authenticated-orcid":false,"given":"Siyuan","family":"Shen","sequence":"additional","affiliation":[{"name":"ETH Z\u00fcrich, Zurich, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-7654-6038","authenticated-orcid":false,"given":"Marcin","family":"Chrapek","sequence":"additional","affiliation":[{"name":"ETH Z\u00fcrich, Zurich, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-8071-2552","authenticated-orcid":false,"given":"Tiancheng","family":"Chen","sequence":"additional","affiliation":[{"name":"ETH Z\u00fcrich, Zurich, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-7701-8258","authenticated-orcid":false,"given":"Kenji","family":"Nakano","sequence":"additional","affiliation":[{"name":"ETH Z\u00fcrich, Zurich, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-6881-8261","authenticated-orcid":false,"given":"Nicola","family":"Mazzoletti","sequence":"additional","affiliation":[{"name":"Swiss National Supercomputing Centre (CSCS), Lugano, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-8809-254X","authenticated-orcid":false,"given":"Peter-Jan","family":"Gootzen","sequence":"additional","affiliation":[{"name":"NVIDIA Corporation, Zwolle, Netherlands"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2197-8860","authenticated-orcid":false,"given":"Salvatore","family":"Di Girolamo","sequence":"additional","affiliation":[{"name":"NVIDIA Corporation, Zurich, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-3601-2786","authenticated-orcid":false,"given":"Rami","family":"Nudelman","sequence":"additional","affiliation":[{"name":"NVIDIA Corporation, Santa Clara, USA"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-6224-9802","authenticated-orcid":false,"given":"Gil","family":"Bloch","sequence":"additional","affiliation":[{"name":"NVIDIA Corporation, Yokne'am Illit, Israel"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9549-7918","authenticated-orcid":false,"given":"Jithin","family":"Jose","sequence":"additional","affiliation":[{"name":"Microsoft Corporation, Redmond, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7106-3824","authenticated-orcid":false,"given":"Abdul","family":"Kabbani","sequence":"additional","affiliation":[{"name":"Microsoft Corporation, Redmond, USA"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-0414-2200","authenticated-orcid":false,"given":"Sreevatsa","family":"Anantharamu","sequence":"additional","affiliation":[{"name":"Microsoft Corporation, Redmond, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5223-8446","authenticated-orcid":false,"given":"Jie","family":"Zhang","sequence":"additional","affiliation":[{"name":"Microsoft Corporation, Redmond, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1356-5951","authenticated-orcid":false,"given":"Konstantin","family":"Taranov","sequence":"additional","affiliation":[{"name":"Microsoft Corporation, Zurich, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8846-5229","authenticated-orcid":false,"given":"Zhuolong","family":"Yu","sequence":"additional","affiliation":[{"name":"Microsoft Corporation, Redmond, USA"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-8500-5810","authenticated-orcid":false,"given":"Scott","family":"Moe","sequence":"additional","affiliation":[{"name":"Microsoft Corporation, Redmond, USA"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-0335-7898","authenticated-orcid":false,"given":"Mahmoud","family":"Elhaddad","sequence":"additional","affiliation":[{"name":"Microsoft Corporation, Redmond, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1333-9797","authenticated-orcid":false,"given":"Torsten","family":"Hoefler","sequence":"additional","affiliation":[{"name":"ETH Z\u00fcrich, Zurich, Switzerland"}]}],"member":"320","published-online":{"date-parts":[[2025,11,15]]},"reference":[{"key":"e_1_3_3_3_2_2","doi-asserted-by":"publisher","DOI":"10.1145\/215399.215427"},{"key":"e_1_3_3_3_3_2","unstructured":"OpenFabrics Alliance. 2025. Open Fabrics Enterprise Distribution (OFED) Performance Tests. https:\/\/github.com\/linux-rdma\/perftest."},{"key":"e_1_3_3_3_4_2","doi-asserted-by":"publisher","DOI":"10.1109\/SC41406.2024.00089"},{"key":"e_1_3_3_3_5_2","first-page":"263","volume-title":"NSDI","author":"Balakrishnan Mahesh","year":"2008","unstructured":"Mahesh Balakrishnan, Tudor Marian, Ken Birman, Hakim Weatherspoon, and Einar Vollset. 2008. Maelstrom: Transparent Error Correction for Lambda Networks.. In NSDI. 263\u2013278."},{"key":"e_1_3_3_3_6_2","first-page":"49","volume-title":"11th USENIX Symposium on Operating Systems Design and Implementation (OSDI 14)","author":"Belay Adam","year":"2014","unstructured":"Adam Belay, George Prekas, Ana Klimovic, Samuel Grossman, Christos Kozyrakis, and Edouard Bugnion. 2014. IX: a protected dataplane operating system for high throughput and low latency. In 11th USENIX Symposium on Operating Systems Design and Implementation (OSDI 14). 49\u201365."},{"key":"e_1_3_3_3_7_2","doi-asserted-by":"crossref","unstructured":"Tal Ben-Nun and Torsten Hoefler. 2019. Demystifying parallel and distributed deep learning: An in-depth concurrency analysis. ACM Computing Surveys (CSUR) 52 4 (2019) 1\u201343.","DOI":"10.1145\/3320060"},{"key":"e_1_3_3_3_8_2","volume-title":"Data networks","author":"Bertsekas Dimitri","year":"2021","unstructured":"Dimitri Bertsekas and Robert Gallager. 2021. Data networks. Athena Scientific."},{"key":"e_1_3_3_3_9_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICNP61940.2024.10858560"},{"key":"e_1_3_3_3_10_2","doi-asserted-by":"publisher","DOI":"10.1145\/3581784.3607099"},{"key":"e_1_3_3_3_11_2","unstructured":"Jo\u00e3o da Silva. 2024. Google turns to nuclear to power AI data centres. BBC News (15 October 2024). https:\/\/www.bbc.com\/news\/articles\/c748gn94k95o Business."},{"key":"e_1_3_3_3_12_2","doi-asserted-by":"publisher","DOI":"10.5555\/3433701.3433747"},{"key":"e_1_3_3_3_13_2","doi-asserted-by":"crossref","unstructured":"Jeffrey Dean and Luiz\u00a0Andr\u00e9 Barroso. 2013. The tail at scale. Commun. ACM 56 2 (2013) 74\u201380.","DOI":"10.1145\/2408776.2408794"},{"key":"e_1_3_3_3_14_2","doi-asserted-by":"crossref","unstructured":"Haotian Dong Jingyan Jiang Rongwei Lu Jiajun Luo Jiajun Song Bowen Li Ying Shen and Zhi Wang. 2025. Beyond A Single AI Cluster: A Survey of Decentralized LLM Training. arXiv preprint arXiv:https:\/\/arXiv.org\/abs\/2503.11023 (2025).","DOI":"10.18653\/v1\/2025.emnlp-main.1486"},{"key":"e_1_3_3_3_15_2","doi-asserted-by":"publisher","DOI":"10.1145\/3651890.3672233"},{"key":"e_1_3_3_3_16_2","unstructured":"Manjunath GorentlaVenkata Valentine Petrov Sergey Lebedev Devendar Bureddy Ferrol Aderholdt Joshua Ladd Gil Bloch Mike Dubman and Gilad Shainer. 2025. Unified Collective Communication (UCC): A Unified Library for CPU GPU and DPU Collectives. IEEE Micro (2025)."},{"key":"e_1_3_3_3_17_2","doi-asserted-by":"publisher","DOI":"10.1145\/3126908.3126970"},{"key":"e_1_3_3_3_18_2","doi-asserted-by":"crossref","unstructured":"Torsten Hoefler Duncan Roweth Keith Underwood Robert Alverson Mark Griswold Vahid Tabatabaee Mohan Kalkunte Surendra Anubolu Siyuan Shen Moray McLaren et\u00a0al. 2023. Data center ethernet and remote direct memory access: Issues at hyperscale. Computer 56 7 (2023) 67\u201377.","DOI":"10.1109\/MC.2023.3261184"},{"key":"e_1_3_3_3_19_2","unstructured":"C. Hopps. 2009. Analysis of an Equal-Cost Multi-Path Algorithm. RFC 2992. https:\/\/www.ietf.org\/rfc\/rfc2992.txt"},{"key":"e_1_3_3_3_20_2","unstructured":"InfiniBand Trade Association. 2024. InfiniBand Specification. https:\/\/www.infinibandta.org."},{"key":"e_1_3_3_3_21_2","unstructured":"Intel. 2025. Intel Intelligent Storage Acceleration Library. https:\/\/www.intel.com\/content\/www\/us\/en\/developer\/tools\/isa-l\/overview.html."},{"key":"e_1_3_3_3_22_2","doi-asserted-by":"publisher","DOI":"10.1109\/SC41406.2024.00109"},{"key":"e_1_3_3_3_23_2","doi-asserted-by":"publisher","DOI":"10.1145\/3098822.3098842"},{"key":"e_1_3_3_3_24_2","first-page":"931","volume-title":"17th USENIX Symposium on Operating Systems Design and Implementation (OSDI 23)","author":"Li Qiang","year":"2023","unstructured":"Qiang Li, Yixiao Gao, Xiaoliang Wang, Haonan Qiu, Yanfang Le, Derui Liu, Qiao Xiang, Fei Feng, Peng Zhang, Bo Li, et\u00a0al. 2023. Flor: An open high performance RDMA framework over heterogeneous NICs. In 17th USENIX Symposium on Operating Systems Design and Implementation (OSDI 23). 931\u2013948."},{"key":"e_1_3_3_3_25_2","unstructured":"Shen Li Yanli Zhao Rohan Varma Omkar Salpekar Pieter Noordhuis Teng Li Adam Paszke Jeff Smith Brian Vaughan Pritam Damania et\u00a0al. 2020. Pytorch distributed: Experiences on accelerating data parallel training. arXiv preprint arXiv:https:\/\/arXiv.org\/abs\/2006.15704 (2020)."},{"key":"e_1_3_3_3_26_2","first-page":"357","volume-title":"15th USENIX symposium on networked systems design and implementation (NSDI 18)","author":"Lu Yuanwei","year":"2018","unstructured":"Yuanwei Lu, Guo Chen, Bojie Li, Kun Tan, Yongqiang Xiong, Peng Cheng, Jiansong Zhang, Enhong Chen, and Thomas Moscibroda. 2018. Multi-Path transport for RDMA in datacenters. In 15th USENIX symposium on networked systems design and implementation (NSDI 18). 357\u2013371."},{"key":"e_1_3_3_3_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/IZS.2004.1287411"},{"key":"e_1_3_3_3_28_2","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359657"},{"key":"e_1_3_3_3_29_2","doi-asserted-by":"crossref","unstructured":"Matt Mathis Jamshid Mahdavi Sally Floyd and Allyn Romanow. 1996. RFC2018: TCP selective acknowledgement options.","DOI":"10.17487\/rfc2018"},{"key":"e_1_3_3_3_30_2","doi-asserted-by":"publisher","DOI":"10.1145\/3230543.3230557"},{"key":"e_1_3_3_3_31_2","unstructured":"Andrew Moseman. 2024. Amazon Vies for Nuclear-Powered Data Center: The deal has become a flash point over energy fairness. IEEE Spectrum (12 August 2024). https:\/\/spectrum.ieee.org\/amazon-data-center-nuclear-power"},{"key":"e_1_3_3_3_32_2","unstructured":"NVIDIA. 2023. NVIDIA BlueField-3 Datasheet. https:\/\/resources.nvidia.com\/en-us-accelerated-networking-resource-library\/datasheet-nvidia-bluefield."},{"key":"e_1_3_3_3_33_2","unstructured":"NVIDIA. 2023. Spectrum-X Datasheet. https:\/\/resources.nvidia.com\/en-us-networking-ai\/networking-ethernet-1."},{"key":"e_1_3_3_3_34_2","unstructured":"NVIDIA. 2025. ConnectX-8 SuperNIC Datasheet. https:\/\/resources.nvidia.com\/en-us-accelerated-networking-resource-library\/connectx-datasheet-c."},{"key":"e_1_3_3_3_35_2","unstructured":"NVIDIA. 2025. DPA Subsystem. https:\/\/docs.nvidia.com\/doca\/sdk\/DPA+Subsystem\/index.html."},{"key":"e_1_3_3_3_36_2","unstructured":"NVIDIA. 2025. NVIDIA Collective Communications Library (NCCL). https:\/\/developer.nvidia.com\/nccl."},{"key":"e_1_3_3_3_37_2","unstructured":"OpenFabrics Alliance. 2025. Libfabric OpenFabrics. https:\/\/ofiwg.github.io\/libfabric\/."},{"key":"e_1_3_3_3_38_2","doi-asserted-by":"publisher","DOI":"10.1145\/50202.50214"},{"key":"e_1_3_3_3_39_2","doi-asserted-by":"crossref","unstructured":"Peter Sanders Jochen Speck and Jesper\u00a0Larsson Tr\u00e4ff. 2009. Two-tree algorithms for full bandwidth broadcast reduction and scan. Parallel Comput. 35 12 (2009) 581\u2013594.","DOI":"10.1016\/j.parco.2009.09.001"},{"key":"e_1_3_3_3_40_2","doi-asserted-by":"crossref","unstructured":"Pasi Sarolahti Markku Kojo and Kimmo Raatikainen. 2003. F-RTO: an enhanced recovery algorithm for TCP retransmission timeouts. ACM SIGCOMM Computer Communication Review 33 2 (2003) 51\u201363.","DOI":"10.1145\/956981.956987"},{"key":"e_1_3_3_3_41_2","unstructured":"SemiAnalysis. 2024. Multi-Datacenter Training: OpenAI\u2019s Ambitious Plan To Beat Google\u2019s Infrastructure. https:\/\/semianalysis.com\/2024\/09\/04\/multi-datacenter-training-openais\/."},{"key":"e_1_3_3_3_42_2","doi-asserted-by":"publisher","DOI":"10.1145\/3642970.3655843"},{"key":"e_1_3_3_3_43_2","doi-asserted-by":"publisher","DOI":"10.1145\/1122971.1122978"},{"key":"e_1_3_3_3_44_2","unstructured":"Gemini Team Rohan Anil Sebastian Borgeaud Jean-Baptiste Alayrac Jiahui Yu Radu Soricut Johan Schalkwyk Andrew\u00a0M Dai Anja Hauth Katie Millican et\u00a0al. 2023. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:https:\/\/arXiv.org\/abs\/2312.11805 (2023)."},{"key":"e_1_3_3_3_45_2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-540-39924-7_38"},{"key":"e_1_3_3_3_46_2","doi-asserted-by":"crossref","unstructured":"Hao Wang Sreeram Potluri Devendar Bureddy Carlos Rosales and Dhabaleswar\u00a0K Panda. 2013. GPU-aware MPI on RDMA-enabled clusters: Design implementation and evaluation. IEEE Transactions on Parallel and Distributed Systems 25 10 (2013) 2595\u20132605.","DOI":"10.1109\/TPDS.2013.222"},{"key":"e_1_3_3_3_47_2","first-page":"1","volume-title":"20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23)","author":"Wang Zilong","year":"2023","unstructured":"Zilong Wang, Layong Luo, Qingsong Ning, Chaoliang Zeng, Wenxue Li, Xinchen Wan, Peng Xie, Tao Feng, Ke Cheng, Xiongfei Geng, et\u00a0al. 2023. SRNIC: A scalable architecture for RDMA NICs. In 20th USENIX Symposium on Networked Systems Design and Implementation (NSDI 23). 1\u201314."},{"key":"e_1_3_3_3_48_2","doi-asserted-by":"publisher","DOI":"10.5555\/554634"},{"key":"e_1_3_3_3_49_2","doi-asserted-by":"publisher","DOI":"10.1145\/2830318.2830321"},{"key":"e_1_3_3_3_50_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCD.2003.1240929"},{"key":"e_1_3_3_3_51_2","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM48880.2022.9796898"},{"key":"e_1_3_3_3_52_2","unstructured":"Yanli Zhao Andrew Gu Rohan Varma Liang Luo Chien-Chin Huang Min Xu Less Wright Hamid Shojanazeri Myle Ott Sam Shleifer et\u00a0al. 2023. Pytorch fsdp: experiences on scaling fully sharded data parallel. arXiv preprint arXiv:https:\/\/arXiv.org\/abs\/2304.11277 (2023)."},{"key":"e_1_3_3_3_53_2","doi-asserted-by":"publisher","DOI":"10.1145\/3098822.3098849"},{"key":"e_1_3_3_3_54_2","doi-asserted-by":"publisher","DOI":"10.1109\/IWQoS61813.2024.10682853"}],"event":{"name":"SC '25: The International Conference for High Performance Computing, Networking, Storage and Analysis","location":"St. Louis MO USA","acronym":"SC '25","sponsor":["SIGHPC ACM Special Interest Group on High Performance Computing, Special Interest Group on High Performance Computing"]},"container-title":["Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis"],"original-title":[],"link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3712285.3759902","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,11]],"date-time":"2026-03-11T18:29:11Z","timestamp":1773253751000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3712285.3759902"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,15]]},"references-count":53,"alternative-id":["10.1145\/3712285.3759902","10.1145\/3712285"],"URL":"https:\/\/doi.org\/10.1145\/3712285.3759902","relation":{},"subject":[],"published":{"date-parts":[[2025,11,15]]},"assertion":[{"value":"2025-11-15","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}