{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,4]],"date-time":"2026-06-04T20:07:12Z","timestamp":1780603632027,"version":"3.54.1"},"reference-count":81,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Guangdong Provincial Young Innovative Talent Program","award":["2025KQNCX077"],"award-info":[{"award-number":["2025KQNCX077"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Netw. Sci. Eng."],"published-print":{"date-parts":[[2026]]},"DOI":"10.1109\/tnse.2026.3695132","type":"journal-article","created":{"date-parts":[[2026,5,20]],"date-time":"2026-05-20T19:53:48Z","timestamp":1779306828000},"page":"9572-9590","source":"Crossref","is-referenced-by-count":0,"title":["Gsched: Coordinated Flow-Control and Priority Scheduling for DNN Training in AI Cluster"],"prefix":"10.1109","volume":"13","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-1599-8200","authenticated-orcid":false,"given":"Heng","family":"Xu","sequence":"first","affiliation":[{"name":"Faculty of Computer Science, Artificial Intelligence, Shenzhen University of Advanced Technology, Shenzhen, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-5313-7750","authenticated-orcid":false,"given":"Chengze","family":"Du","sequence":"additional","affiliation":[{"name":"Faculty of Computer Science, Artificial Intelligence, Shenzhen University of Advanced Technology, Shenzhen, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-7776-4985","authenticated-orcid":false,"given":"Zhiwei","family":"Yu","sequence":"additional","affiliation":[{"name":"Institute for Network Sciences, Cyberspace, Tsinghua University, Beijing, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6856-1785","authenticated-orcid":false,"given":"Letian","family":"Li","sequence":"additional","affiliation":[{"name":"Department of Information Engineering, Chinese University of Hong Kong, Shatin, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-9803-262X","authenticated-orcid":false,"given":"Ying","family":"Zhou","sequence":"additional","affiliation":[{"name":"Faculty of Computer Science, Artificial Intelligence, Shenzhen University of Advanced Technology, Shenzhen, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8695-7342","authenticated-orcid":false,"given":"Bo","family":"Liu","sequence":"additional","affiliation":[{"name":"Faculty of Computer Science, Artificial Intelligence, Shenzhen University of Advanced Technology, Shenzhen, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3416-5551","authenticated-orcid":false,"given":"Jialong","family":"Li","sequence":"additional","affiliation":[{"name":"Faculty of Computer Science, Artificial Intelligence, Shenzhen University of Advanced Technology, Shenzhen, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1145\/3544216.3544252"},{"key":"ref2","first-page":"51","article-title":"{PowerTCP}: Pushing the performance limits of datacenter networks","volume-title":"Proc. 19th USENIX Symp. Netw. Syst. Des. Implementation","author":"Addanki","year":"2022"},{"key":"ref3","first-page":"89","article-title":"Hedera: Dynamic flow scheduling for data center networks","volume-title":"Proc. 7th USENIX Symp. Netw. Syst. Des. Implementation","author":"Al-Fares","year":"2010"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1145\/2486001.2486031"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1145\/2619239.2626316"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1145\/1851182.1851192"},{"key":"ref7","first-page":"219","article-title":"Bolt:{Sub-RTT} congestion control for {ultra-low} latency","volume-title":"Proc. 20th USENIX Symp. Netw. Syst. Des. Implementation","author":"Arslan","year":"2023"},{"key":"ref8","first-page":"537","article-title":"Enabling {ECN} in {multi-service} {multi-queue} data centers","volume-title":"Proc. 13th USENIX Symp. Netw. Syst. Des. Implementation","author":"Bai","year":"2016"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2017.2669216"},{"key":"ref10","first-page":"37","article-title":"Yeah-TCP: Yet another highspeed TCP","volume-title":"Proc. PFLDnet","volume":"7","author":"Baiocchi","year":"2007"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/49.464716"},{"key":"ref12","first-page":"1877","article-title":"Language models are few-shot learners","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Brown","year":"2020"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3651890.3672239"},{"key":"ref14","first-page":"407","article-title":"{eTran}: Extensible kernel transport with {eBPF}","volume-title":"Proc. 22nd USENIX Symp. Netw. Syst. Des. Implementation","author":"Chen","year":"2025"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/2740070.2626315"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/2829988.2787480"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1810.04805"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1145\/2740070.2626322"},{"key":"ref19","article-title":"Cross-timeslot optimization for distributed GPU inference using reinforcement learning","author":"Du","year":"2025"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM48880.2022.9796820"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1007\/11499169_22"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/GLOCOM.2001.965869"},{"key":"ref23","first-page":"485","article-title":"Tiresias: A GPU cluster manager for distributed deep learning","volume-title":"Proc. 16th USENIX Symp. Netw. Syst. Des. Implementation","author":"Gu","year":"2019"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1145\/1400097.1400105"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1145\/3098822.3098825"},{"key":"ref26","first-page":"418","article-title":"TicTac: Accelerating distributed deep learning with communication scheduling","volume-title":"Proc. Mach. Learn. Syst.","author":"Sayed","year":"2019"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1145\/2829988.2787507"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1145\/2377677.2377710"},{"key":"ref30","first-page":"709","article-title":"Characterization of large language model development in the datacenter","volume-title":"Proc. 21st USENIX Symp. Netw. Syst. Des. Implementation","author":"Hu","year":"2024"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1145\/52324.52356"},{"key":"ref32","first-page":"833","article-title":"Your coflow has many flows: Sampling them for fun and speed","volume-title":"Proc. USENIX Annu. Tech. Conf.","author":"Jajoo"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1145\/3600006.3613175"},{"key":"ref34","first-page":"463","article-title":"A unified architecture for accelerating distributed DNN training in heterogeneous GPU\/CPU clusters","volume-title":"Proc. 14th USENIX Symp. Operating Syst. Des. Implementation","author":"Jiang","year":"2020"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1145\/964725.633035"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1145\/3387514.3406591"},{"key":"ref37","first-page":"2515","article-title":"Fast: An efficient scheduler for all-to-all GPU communication","volume-title":"Proc. 23rd USENIX Symp. Netw. Syst. Des. Implementation (NSDI 26)","author":"Lei","year":"2026"},{"key":"ref38","doi-asserted-by":"crossref","first-page":"2478","DOI":"10.1109\/TON.2025.3645806","volume":"34","author":"Li","year":"2026","journal-title":"IEEE Trans. Netw."},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1145\/3663408.3663409"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1145\/3341302.3342085"},{"key":"ref41","article-title":"RoBERTa: A robustly optimized bert pretraining approach","author":"Liu","year":"2019"},{"key":"ref42","first-page":"289","article-title":"Themis: Fair and efficient GPU cluster scheduling","volume-title":"Proc. 17th USENIX Symp. Netw. Syst. Des. Implementation","author":"Mahajan","year":"2020"},{"key":"ref43","first-page":"809","article-title":"Better Together: Jointly optimizing ML collective scheduling and execution planning using SYNDICATE","volume-title":"Proc. 20th USENIX Symp. Netw. Syst. Des. Implementation","author":"Mahajan","year":"2023"},{"key":"ref44","article-title":"Optimizing network performance in distributed machine learning","volume-title":"Proc. 7th USENIX Workshop Hot Topics Cloud Comput.","author":"Mai","year":"2015"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.acl-main.645"},{"key":"ref46","first-page":"1","article-title":"Expanding across time to deliver bandwidth efficiency and low latency","volume-title":"Proc. 17th USENIX Symp. Netw. Syst. Des. Implementation","author":"Mellette","year":"2020"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1145\/2829988.2787510"},{"key":"ref48","first-page":"579","article-title":"Looking beyond {GPUs} for {DNN} scheduling on {multi-tenant} clusters","volume-title":"Proc. 16th USENIX Symp. Operating Syst. Des. Implementation","author":"Mohan","year":"2022"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1145\/3230543.3230564"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1145\/3563766.3564096"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1145\/3190508.3190517"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359642"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1145\/2619239.2626309"},{"key":"ref54","first-page":"1","article-title":"Pollux: Co-adaptive cluster scheduling for goodput-optimized deep learning","volume-title":"Proc. 15th USENIX Symp. Operating Syst. Des. Implementation","author":"Qiao","year":"2021"},{"issue":"8","key":"ref55","article-title":"Language models are unsupervised multitask learners","volume":"1","author":"Radford","year":"2019","journal-title":"OpenAI Blog"},{"key":"ref56","first-page":"1403","article-title":"CASSINI: Network-aware job scheduling in machine learning clusters","volume-title":"Proc. 21st USENIX Symp. Networked Syst. Des. Implementation","author":"Rajasekaran","year":"2024"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1145\/3696348.3696878"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1145\/2785956.2787472"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1287\/opre.16.3.687"},{"key":"ref60","first-page":"593","article-title":"TACCL: Guiding collective algorithm synthesis using communication sketches","volume-title":"Proc. 20th USENIX Symp. Netw. Syst. Des. Implementation","author":"Shah"},{"key":"ref61","article-title":"Very deep convolutional networks for large-scale image recognition","author":"Simonyan","year":"2014"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.1145\/3603269.3604849"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS57955.2024.00066"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.1145\/3386367.3431316"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2023.3331372"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1109\/TNSE.2024.3523320"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1145\/3563766.3564096"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.1145\/2342356.2342388"},{"key":"ref69","first-page":"407","article-title":"Let it flow: Resilient asymmetric load balancing with flowlet switching","volume-title":"Proc. 14th USENIX Symp. Netw. Syst. Des. Implementation","author":"Vanini","year":"2017"},{"key":"ref70","first-page":"595","article-title":"Gandiva: Introspective cluster scheduling for deep learning","volume-title":"Proc. 13th USENIX Symp. Operating Syst. Des. Implementation","author":"Xiao","year":"2018"},{"key":"ref71","first-page":"533","article-title":"{AntMan}: Dynamic scaling on {GPU} clusters for deep learning","volume-title":"Proc. 14th USENIX Symp. Operating Syst. Des. Implementation","author":"Xiao","year":"2020"},{"key":"ref72","doi-asserted-by":"publisher","DOI":"10.1109\/TON.2026.3676382"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1109\/INFCOM.2004.1354672"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1145\/3452296.3472927"},{"key":"ref75","doi-asserted-by":"publisher","DOI":"10.1109\/TNSE.2023.3271869"},{"key":"ref76","doi-asserted-by":"publisher","DOI":"10.1145\/3131365.3131375"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1145\/3544216.3544224"},{"key":"ref78","first-page":"703","article-title":"Shockwave: Fair and efficient cluster scheduling for dynamic adaptation in machine learning","volume-title":"Proc. 20th USENIX Symp. Netw. Syst. Des. Implementation","author":"Zheng","year":"2023"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.1109\/TNSE.2025.3626049"},{"key":"ref80","article-title":"UCCL-Tran: An extensible software transport layer for GPU networking","volume-title":"Proc. USENIX OSDI","author":"Zhou","year":"2026"},{"key":"ref81","doi-asserted-by":"publisher","DOI":"10.1145\/2785956.2787484"}],"container-title":["IEEE Transactions on Network Science and Engineering"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/6488902\/11264281\/11527076.pdf?arnumber=11527076","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,4]],"date-time":"2026-06-04T19:57:02Z","timestamp":1780603022000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11527076\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":81,"URL":"https:\/\/doi.org\/10.1109\/tnse.2026.3695132","relation":{},"ISSN":["2327-4697","2334-329X"],"issn-type":[{"value":"2327-4697","type":"electronic"},{"value":"2334-329X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026]]}}}