{"id":"https://openalex.org/W4416199140","doi":"https://doi.org/10.1145/3712285.3759823","title":"gLLM: Global Balanced Pipeline Parallelism Systems for Distributed LLMs Serving with Token Throttling","display_name":"gLLM: Global Balanced Pipeline Parallelism Systems for Distributed LLMs Serving with Token Throttling","publication_year":2025,"publication_date":"2025-11-12","ids":{"openalex":"https://openalex.org/W4416199140","doi":"https://doi.org/10.1145/3712285.3759823"},"language":null,"primary_location":{"id":"doi:10.1145/3712285.3759823","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3712285.3759823","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis","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/A5108251396","display_name":"Tianyu Guo","orcid":"https://orcid.org/0009-0005-2979-4486"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Tianyu Guo","raw_affiliation_strings":["Sun Yat-sen University, Guangzhou, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039843650","display_name":"Xianwei Zhang","orcid":"https://orcid.org/0000-0003-3507-4299"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xianwei Zhang","raw_affiliation_strings":["Sun Yat-sen University, Guangzhou, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024085713","display_name":"Jiangsu Du","orcid":"https://orcid.org/0000-0003-4707-9492"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiangsu Du","raw_affiliation_strings":["Sun Yat-sen University, Guangzhou, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101483479","display_name":"Zhiguang Chen","orcid":"https://orcid.org/0000-0002-9318-5715"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhiguang Chen","raw_affiliation_strings":["Sun Yat-sen University, Guangzhou, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023506057","display_name":"Nong Xiao","orcid":"https://orcid.org/0000-0002-2166-977X"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Nong Xiao","raw_affiliation_strings":["Sun Yat-sen University, Guangzhou, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101633465","display_name":"Yutong Lu","orcid":"https://orcid.org/0000-0001-5315-3375"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yutong Lu","raw_affiliation_strings":["Sun Yat-sen University, Guangzhou, China"],"affiliations":[{"raw_affiliation_string":"Sun Yat-sen University, Guangzhou, China","institution_ids":["https://openalex.org/I157773358"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5108251396"],"corresponding_institution_ids":["https://openalex.org/I157773358"],"apc_list":null,"apc_paid":null,"fwci":1.4888,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":{"value":0.8835264,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":91,"max":95},"biblio":{"volume":null,"issue":null,"first_page":"1725","last_page":"1741"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12127","display_name":"Software System Performance and Reliability","score":0.26429998874664307,"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"}},"topics":[{"id":"https://openalex.org/T12127","display_name":"Software System Performance and Reliability","score":0.26429998874664307,"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.20579999685287476,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10181","display_name":"Natural Language Processing Techniques","score":0.12929999828338623,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"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/pipeline","display_name":"Pipeline (software)","score":0.6675999760627747},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.625},{"id":"https://openalex.org/keywords/parallelism","display_name":"Parallelism (grammar)","score":0.5803999900817871},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.542900025844574},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4361000061035156},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.4205000102519989},{"id":"https://openalex.org/keywords/pipeline-transport","display_name":"Pipeline transport","score":0.40939998626708984},{"id":"https://openalex.org/keywords/bandwidth-throttling","display_name":"Bandwidth throttling","score":0.37540000677108765}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8199999928474426},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.6675999760627747},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.625},{"id":"https://openalex.org/C2781172179","wikidata":"https://www.wikidata.org/wiki/Q853109","display_name":"Parallelism (grammar)","level":2,"score":0.5803999900817871},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.542900025844574},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5198000073432922},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.49559998512268066},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4361000061035156},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.4205000102519989},{"id":"https://openalex.org/C175309249","wikidata":"https://www.wikidata.org/wiki/Q725864","display_name":"Pipeline transport","level":2,"score":0.40939998626708984},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.38440001010894775},{"id":"https://openalex.org/C173061102","wikidata":"https://www.wikidata.org/wiki/Q478819","display_name":"Bandwidth throttling","level":3,"score":0.37540000677108765},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.3727000057697296},{"id":"https://openalex.org/C526435321","wikidata":"https://www.wikidata.org/wiki/Q1303814","display_name":"Processor design","level":2,"score":0.31349998712539673},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.3127000033855438},{"id":"https://openalex.org/C150495011","wikidata":"https://www.wikidata.org/wiki/Q128392","display_name":"Concurrent computing","level":2,"score":0.28529998660087585},{"id":"https://openalex.org/C193702766","wikidata":"https://www.wikidata.org/wiki/Q1414548","display_name":"Concurrency","level":2,"score":0.28290000557899475},{"id":"https://openalex.org/C106515295","wikidata":"https://www.wikidata.org/wiki/Q26806595","display_name":"Parallel processing","level":2,"score":0.2815999984741211},{"id":"https://openalex.org/C72108876","wikidata":"https://www.wikidata.org/wiki/Q844565","display_name":"Transaction processing","level":3,"score":0.27619999647140503},{"id":"https://openalex.org/C61483411","wikidata":"https://www.wikidata.org/wiki/Q3124522","display_name":"Data parallelism","level":3,"score":0.2741999924182892},{"id":"https://openalex.org/C205711294","wikidata":"https://www.wikidata.org/wiki/Q176953","display_name":"Rendering (computer graphics)","level":2,"score":0.2533999979496002}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3712285.3759823","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3712285.3759823","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":36,"referenced_works":["https://openalex.org/W3193985311","https://openalex.org/W3204998121","https://openalex.org/W4297639955","https://openalex.org/W4312060029","https://openalex.org/W4327810158","https://openalex.org/W4386348101","https://openalex.org/W4387321091","https://openalex.org/W4389157038","https://openalex.org/W4389501005","https://openalex.org/W4391971727","https://openalex.org/W4392679398","https://openalex.org/W4393160097","https://openalex.org/W4394946008","https://openalex.org/W4395117348","https://openalex.org/W4395117922","https://openalex.org/W4398160786","https://openalex.org/W4398184813","https://openalex.org/W4399403248","https://openalex.org/W4401211704","https://openalex.org/W4401306886","https://openalex.org/W4401307699","https://openalex.org/W4402667043","https://openalex.org/W4402670901","https://openalex.org/W4402670978","https://openalex.org/W4402671526","https://openalex.org/W4402671709","https://openalex.org/W4402671761","https://openalex.org/W4402671905","https://openalex.org/W4402684157","https://openalex.org/W4404783031","https://openalex.org/W4404783127","https://openalex.org/W4405655184","https://openalex.org/W4405903187","https://openalex.org/W4408029418","https://openalex.org/W4408029480","https://openalex.org/W6929457143"],"related_works":[],"abstract_inverted_index":{"Pipeline":[0],"parallelism":[1,35],"has":[2],"emerged":[3],"as":[4],"a":[5,72],"predominant":[6],"approach":[7],"for":[8],"deploying":[9],"large":[10],"language":[11],"models":[12],"(LLMs)":[13],"across":[14,52],"distributed":[15],"nodes,":[16],"owing":[17],"to":[18,24,59,98],"its":[19],"lower":[20],"communication":[21],"overhead":[22],"compared":[23],"tensor":[25],"parallelism.":[26],"While":[27],"demonstrating":[28],"high":[29],"throughput":[30],"in":[31],"request":[32],"serving,":[33],"pipeline":[34,42],"often":[36],"faces":[37],"performance":[38],"limitations":[39],"caused":[40],"by":[41],"bubbles,":[43],"which":[44],"are":[45],"primarily":[46],"resulted":[47],"from":[48,86],"imbalanced":[49],"computation":[50],"delays":[51],"batches.":[53],"Existing":[54],"methods":[55,78],"like":[56],"Sarathi-Serve":[57],"attempt":[58],"address":[60],"this":[61],"through":[62],"hybrid":[63],"scheduling":[64],"of":[65,93],"chunked":[66],"prefill":[67,88],"and":[68],"decode":[69,94],"tokens":[70,89],"with":[71],"fixed":[73],"token":[74],"budget.":[75],"However,":[76],"such":[77],"may":[79],"still":[80],"experience":[81],"significant":[82],"fluctuations,":[83],"arising":[84],"either":[85],"insufficient":[87],"or":[90],"uneven":[91],"distribution":[92],"tokens,":[95],"ultimately":[96],"leading":[97],"computational":[99],"imbalance.":[100]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-03-27T05:58:40.876381","created_date":"2025-11-12T00:00:00"}
