{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T05:44:56Z","timestamp":1782798296497,"version":"3.54.5"},"reference-count":58,"publisher":"IEEE","license":[{"start":{"date-parts":[[2026,5,18]],"date-time":"2026-05-18T00:00:00Z","timestamp":1779062400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,5,18]],"date-time":"2026-05-18T00:00:00Z","timestamp":1779062400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/100006190","name":"Research and Development","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100006190","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,5,18]]},"DOI":"10.1109\/infocom59046.2026.11571517","type":"proceedings-article","created":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T19:38:15Z","timestamp":1782761895000},"page":"1-10","source":"Crossref","is-referenced-by-count":0,"title":["3D-INA: An Exploration of Integrating In-Network Aggregation into 3D Parallelism for LLM Training"],"prefix":"10.1109","author":[{"given":"Huifeng","family":"Xing","sequence":"first","affiliation":[{"name":"Fudan University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hao","family":"Wang","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yinfan","family":"Hu","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xin","family":"Ai","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zixuan","family":"Chen","sequence":"additional","affiliation":[{"name":"China Telecom"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yang","family":"Chen","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wanxin","family":"Shi","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sen","family":"Liu","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yang","family":"Xu","sequence":"additional","affiliation":[{"name":"Fudan University"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"MegaScale: Scaling large language model training to more than 10,000 GPUs","author":"Jiang","year":"2024","journal-title":"NSDI\u201924"},{"key":"ref2","article-title":"Alibaba hpn: A data center network for large language model training","author":"Kun","year":"2024","journal-title":"SIGCOMM\u201924"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1145\/3267809.3267840"},{"key":"ref4","article-title":"Megatron-lm: Training multi-billion parameter language models using model parallelism","author":"Shoeybi","year":"2020"},{"key":"ref5","article-title":"Gpipe: Efficient training of giant neural networks using pipeline parallelism","author":"Huang","year":"2018"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1145\/3579371.3589350"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575712"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/2640087.2644155"},{"key":"ref9","article-title":"Nvidia collective communication library"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2022.3213237"},{"key":"ref11","article-title":"Scaling distributed machine learning with In-Network aggregation","author":"Sapio","year":"2021","journal-title":"NSDI\u201921"},{"key":"ref12","article-title":"Atp: In-network aggregation for multi-tenant learning","author":"Lao","year":"2021","journal-title":"NSDI\u201921"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM53939.2023.10228956"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/COMHPC.2016.006"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-50743-5_3"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/3575693.3575708"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1016\/j.jpdc.2008.09.002"},{"key":"ref18","article-title":"ns-3 network simulator","year":"2011-2022"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1147\/rd.492.0265"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1145\/1402946.1402967"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1145\/3127479.3132018"},{"key":"ref22","article-title":"Improving language understanding by generative pretraining","author":"Radford","year":"2018"},{"key":"ref23","article-title":"Optimization with pulp"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2008.19"},{"key":"ref25","article-title":"Intel\u00ae tofino\u2122 programmable ethernet switch asic"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1147\/JRD.2019.2947013"},{"key":"ref27","article-title":"Swing: Short-cutting rings for higher bandwidth allreduce","author":"Sensi","year":"2024","journal-title":"NSDI 24"},{"key":"ref28","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2019"},{"key":"ref29","article-title":"Llama 2: Open foundation and fine-tuned chat models","author":"Touvron","year":"2023"},{"key":"ref30","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/D16-1264","article-title":"Squad: 100,000+ questions for machine comprehension of text","author":"Rajpurkar","year":"2016"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS57955.2024.00073"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/2785956.2787508"},{"key":"ref33","first-page":"739","article-title":"TopoOpt: Co-optimizing network topology and parallelization strategy for distributed training jobs","volume-title":"NSDI\u201923","author":"Wang"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1145\/3651890.3672228"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1145\/3552326.3587436"},{"key":"ref36","article-title":"A unified architecture for accelerating distributed dnn training in heterogeneous gpu\/cpu clusters","author":"Jiang","year":"2020","journal-title":"OSDI 20"},{"key":"ref37","article-title":"Horovod: fast and easy distributed deep learning in tensorflow","author":"Sergeev","year":"2018"},{"key":"ref38","article-title":"Massively distributed sgd: Imagenet\/resnet-50training in a flash","author":"Mikami","year":"2018"},{"key":"ref39","article-title":"Plink: Discovering and exploiting locality for accelerated distributed training on the public cloud","author":"Luo","year":"2020","journal-title":"MLSys 20"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1145\/1594977.1592577"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1145\/1402946.1402968"},{"key":"ref42","article-title":"Jellyfish: Networking data centers randomly","author":"Singla","year":"2012","journal-title":"NSDI\u201912"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/SC41404.2022.00016"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/ICNP59255.2023.10355574"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS63083.2025.00040"},{"key":"ref46","article-title":"Netreduce: Rdma-compatible in-network reduction for distributed dnn training acceleration","author":"Liu","year":"2020"},{"key":"ref47","article-title":"Panama: In-network aggregation for shared machine learning clusters","author":"Gebara","year":"2021","journal-title":"MLSys \u201921"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476178"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1145\/3307650.3322259"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1145\/3289602.3293963"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/HOTI51249.2020.00023"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2019.10.033"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2023.3244794"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1145\/3589334.3645394"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2024.3414853"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1109\/ICNP61940.2024.10858570"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/IWQoS57198.2023.10188783"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1145\/3617232.3624863"}],"event":{"name":"IEEE INFOCOM 2026 - IEEE Conference on Computer Communications","location":"Tokyo, Japan","start":{"date-parts":[[2026,5,18]]},"end":{"date-parts":[[2026,5,21]]}},"container-title":["IEEE INFOCOM 2026 - IEEE Conference on Computer Communications"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11571071\/11571169\/11571517.pdf?arnumber=11571517","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T05:17:21Z","timestamp":1782796641000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11571517\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,18]]},"references-count":58,"URL":"https:\/\/doi.org\/10.1109\/infocom59046.2026.11571517","relation":{},"subject":[],"published":{"date-parts":[[2026,5,18]]}}}