{"id":"https://openalex.org/W4415125436","doi":"https://doi.org/10.1109/icnp65844.2025.11192450","title":"Symphony: Collective Coordination in Multi-Tenant GPU Clusters","display_name":"Symphony: Collective Coordination in Multi-Tenant GPU Clusters","publication_year":2025,"publication_date":"2025-09-22","ids":{"openalex":"https://openalex.org/W4415125436","doi":"https://doi.org/10.1109/icnp65844.2025.11192450"},"language":"en","primary_location":{"id":"doi:10.1109/icnp65844.2025.11192450","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icnp65844.2025.11192450","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 33rd International Conference on Network Protocols (ICNP)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5044077636","display_name":"Manaf Bin-Yahya","orcid":"https://orcid.org/0000-0003-3688-1594"},"institutions":[{"id":"https://openalex.org/I4210160618","display_name":"Huawei Technologies (United Kingdom)","ror":"https://ror.org/056gzgs71","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250955327","https://openalex.org/I4210160618"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Manaf Bin-Yahya","raw_affiliation_strings":["Huawei Technologies Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd","institution_ids":["https://openalex.org/I4210160618"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073778515","display_name":"Amir Shani","orcid":"https://orcid.org/0000-0002-4614-5315"},"institutions":[{"id":"https://openalex.org/I4210160618","display_name":"Huawei Technologies (United Kingdom)","ror":"https://ror.org/056gzgs71","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250955327","https://openalex.org/I4210160618"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Amir Shani","raw_affiliation_strings":["Huawei Technologies Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd","institution_ids":["https://openalex.org/I4210160618"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076940594","display_name":"Hossein Shafieirad","orcid":"https://orcid.org/0000-0002-0633-404X"},"institutions":[{"id":"https://openalex.org/I4210160618","display_name":"Huawei Technologies (United Kingdom)","ror":"https://ror.org/056gzgs71","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250955327","https://openalex.org/I4210160618"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Hossein Shafieirad","raw_affiliation_strings":["Huawei Technologies Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd","institution_ids":["https://openalex.org/I4210160618"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032889517","display_name":"Seyed Hossein Mortazavi","orcid":"https://orcid.org/0000-0002-0451-1911"},"institutions":[{"id":"https://openalex.org/I4210160618","display_name":"Huawei Technologies (United Kingdom)","ror":"https://ror.org/056gzgs71","country_code":"GB","type":"company","lineage":["https://openalex.org/I2250955327","https://openalex.org/I4210160618"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Seyed Hossein Mortazavi","raw_affiliation_strings":["Huawei Technologies Co., Ltd"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Huawei Technologies Co., Ltd","institution_ids":["https://openalex.org/I4210160618"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100700515","display_name":"Chen Ying","orcid":"https://orcid.org/0000-0001-7713-2691"},"institutions":[{"id":"https://openalex.org/I185261750","display_name":"University of Toronto","ror":"https://ror.org/03dbr7087","country_code":"CA","type":"education","lineage":["https://openalex.org/I185261750"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Chen Ying","raw_affiliation_strings":["University of Toronto"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Toronto","institution_ids":["https://openalex.org/I185261750"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119975725","display_name":"Aaron Wang","orcid":null},"institutions":[{"id":"https://openalex.org/I151746483","display_name":"University of Waterloo","ror":"https://ror.org/01aff2v68","country_code":"CA","type":"education","lineage":["https://openalex.org/I151746483"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Aaron Wang","raw_affiliation_strings":["University of Waterloo"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Waterloo","institution_ids":["https://openalex.org/I151746483"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119975726","display_name":"Geng Li","orcid":null},"institutions":[{"id":"https://openalex.org/I4210144487","display_name":"Cloud Computing Center","ror":"https://ror.org/04aa0zm65","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210144487"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Geng Li","raw_affiliation_strings":["China Telecom Cloud Computing Research Institute"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"China Telecom Cloud Computing Research Institute","institution_ids":["https://openalex.org/I4210144487"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5081502954","display_name":"Majid Ghaderi","orcid":"https://orcid.org/0000-0002-3783-4346"},"institutions":[{"id":"https://openalex.org/I168635309","display_name":"University of Calgary","ror":"https://ror.org/03yjb2x39","country_code":"CA","type":"education","lineage":["https://openalex.org/I168635309"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Majid Ghaderi","raw_affiliation_strings":["University of Calgary"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Calgary","institution_ids":["https://openalex.org/I168635309"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":8,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.25500209,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"13"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12784","display_name":"Modular Robots and Swarm Intelligence","score":0.9905999898910522,"subfield":{"id":"https://openalex.org/subfields/2210","display_name":"Mechanical Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12784","display_name":"Modular Robots and Swarm Intelligence","score":0.9905999898910522,"subfield":{"id":"https://openalex.org/subfields/2210","display_name":"Mechanical Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12288","display_name":"Optimization and Search Problems","score":0.9675999879837036,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10101","display_name":"Cloud Computing and Resource Management","score":0.9663000106811523,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/testbed","display_name":"Testbed","score":0.7240999937057495},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.6190000176429749},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.5491999983787537},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.30469998717308044},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.29280000925064087},{"id":"https://openalex.org/keywords/execution-time","display_name":"Execution time","score":0.27970001101493835}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.828000009059906},{"id":"https://openalex.org/C31395832","wikidata":"https://www.wikidata.org/wiki/Q1318674","display_name":"Testbed","level":2,"score":0.7240999937057495},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.6190000176429749},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.604200005531311},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.5491999983787537},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.3828999996185303},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.35440000891685486},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.30469998717308044},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.2939000129699707},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.29280000925064087},{"id":"https://openalex.org/C2989134064","wikidata":"https://www.wikidata.org/wiki/Q288510","display_name":"Execution time","level":2,"score":0.27970001101493835},{"id":"https://openalex.org/C162319229","wikidata":"https://www.wikidata.org/wiki/Q175263","display_name":"Data structure","level":2,"score":0.27810001373291016},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2743000090122223},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.27250000834465027},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2612999975681305}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icnp65844.2025.11192450","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icnp65844.2025.11192450","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE 33rd International Conference on Network Protocols (ICNP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":19,"referenced_works":["https://openalex.org/W2057332538","https://openalex.org/W2131613942","https://openalex.org/W2149939304","https://openalex.org/W2159158649","https://openalex.org/W2919897868","https://openalex.org/W3006487403","https://openalex.org/W4235732943","https://openalex.org/W4255119735","https://openalex.org/W4281790033","https://openalex.org/W4293248052","https://openalex.org/W4391136507","https://openalex.org/W4392699261","https://openalex.org/W4401175629","https://openalex.org/W4401176521","https://openalex.org/W4401176799","https://openalex.org/W4401176820","https://openalex.org/W4401176822","https://openalex.org/W4401352466","https://openalex.org/W4404955085"],"related_works":[],"abstract_inverted_index":{"Multi-tenant":[0],"GPU":[1,127],"clusters":[2,93],"are":[3],"designed":[4],"to":[5,43,58,97,146,158],"concurrently":[6],"run":[7],"multiple":[8],"distributed":[9],"ML":[10,136],"training":[11,155],"workloads.":[12],"However,":[13],"frequent":[14],"data":[15],"transfers":[16],"among":[17],"GPUs":[18],"via":[19],"collective":[20,41,82,99],"operations":[21,83],"can":[22],"slow":[23],"down":[24],"training,":[25],"as":[26],"collectives":[27],"from":[28],"different":[29],"tenants":[30,85],"compete":[31],"for":[32],"network":[33,45,60],"bandwidth.":[34],"Recent":[35],"work":[36],"(e.g.,":[37],"CASSINI)":[38],"has":[39],"considered":[40],"coordination":[42],"prevent":[44],"contention,":[46],"but":[47],"primarily":[48],"focused":[49],"on":[50],"static":[51],"job-level":[52],"optimizations":[53],"at":[54,86],"deployment":[55],"time,":[56],"oblivious":[57],"runtime":[59],"conditions":[61],"and":[62,102,130,152],"the":[63,98,115],"specific":[64],"traffic":[65],"pattern":[66],"of":[67,144],"each":[68],"workload.":[69],"In":[70],"this":[71],"paper,":[72],"we":[73,141],"present":[74],"Symphony,":[75],"an":[76],"application-layer":[77],"solution":[78],"that":[79,109],"dynamically":[80],"coordinates":[81],"across":[84],"runtime.":[87],"Symphony":[88,122],"integrates":[89],"seamlessly":[90],"with":[91,94],"existing":[92],"minimal":[95],"modifications":[96],"communication":[100,150],"library":[101],"includes":[103],"a":[104,125],"lightweight":[105],"online":[106],"scheduling":[107],"mechanism":[108],"requires":[110],"no":[111],"advance":[112],"information":[113],"about":[114],"workloads":[116,137],"or":[117],"their":[118],"collectives.":[119],"We":[120],"evaluate":[121],"using":[123,134],"both":[124],"real":[126],"testbed":[128],"implementation":[129],"trace-driven":[131],"simulations.":[132],"Specifically,":[133],"realistic":[135],"in":[138,148,154],"our":[139],"testbed,":[140],"observe":[142],"improvements":[143],"up":[145],"13.2%":[147],"average":[149],"time":[151,156],"9.6%":[153],"compared":[157],"state-of-the-art":[159],"solutions.":[160]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-14T00:00:00"}
