FedSA: A Semi-Asynchronous Federated Learning Mechanism in Heterogeneous Edge Computing

Federated learning (FL) involves training machine learning models over distributed edge nodes ( i.e. , workers) while facing three critical challenges, edge heterogeneity, Non-IID data and communication resource constraint. In the synchronous FL, the parameter server has to wait for the slowest work...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE journal on selected areas in communications Vol. 39; no. 12; pp. 3654 - 3672
Main Authors:	Ma, Qianpiao, Xu, Yang, Xu, Hongli, Jiang, Zhida, Huang, Liusheng, Huang, He
Format:	Journal Article
Language:	English
Published:	New York IEEE 01.12.2021 The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:	Accuracy Adaptive learning Algorithms Analytical models Cognitive tasks Collaborative work Completion time Computational modeling Convergence Data models Edge computing Federated learning Heterogeneity Machine learning Mathematical models non-IID Parameters semi-asynchronous mechanism Servers Staff participation Training data
ISSN:	0733-8716, 1558-0008
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Abstract	Federated learning (FL) involves training machine learning models over distributed edge nodes ( i.e. , workers) while facing three critical challenges, edge heterogeneity, Non-IID data and communication resource constraint. In the synchronous FL, the parameter server has to wait for the slowest workers, leading to significant waiting time due to edge heterogeneity. Though asynchronous FL can well tackle the edge heterogeneity, it requires frequent model transfers, resulting in massive communication resource consumption. Moreover, the different relative frequency of workers participating in asynchronous updating may seriously hurt training accuracy, especially on Non-IID data. In this paper, we propose a semi-asynchronous federated learning mechanism (FedSA), where the parameter server aggregates a certain number of local models by their arrival order in each round. We theoretically analyze the quantitative relationship between the convergence bound of FedSA and different factors, e.g. , the number of participating workers in each round, the degree of data Non-IID and edge heterogeneity. Based on the convergence bound, we present an efficient algorithm to determine the number of participating workers to minimize the training completion time. To further improve the training accuracy on Non-IID data, FedSA deploys adaptive learning rates for workers by their relative participation frequency. We extend our proposed mechanism to the dynamic and multiple learning tasks scenarios. Experimental results on the testbed show that our proposed mechanism and algorithms address the three challenges more effectively than the state-of-the-art solutions.
AbstractList	Federated learning (FL) involves training machine learning models over distributed edge nodes ( i.e. , workers) while facing three critical challenges, edge heterogeneity, Non-IID data and communication resource constraint. In the synchronous FL, the parameter server has to wait for the slowest workers, leading to significant waiting time due to edge heterogeneity. Though asynchronous FL can well tackle the edge heterogeneity, it requires frequent model transfers, resulting in massive communication resource consumption. Moreover, the different relative frequency of workers participating in asynchronous updating may seriously hurt training accuracy, especially on Non-IID data. In this paper, we propose a semi-asynchronous federated learning mechanism (FedSA), where the parameter server aggregates a certain number of local models by their arrival order in each round. We theoretically analyze the quantitative relationship between the convergence bound of FedSA and different factors, e.g. , the number of participating workers in each round, the degree of data Non-IID and edge heterogeneity. Based on the convergence bound, we present an efficient algorithm to determine the number of participating workers to minimize the training completion time. To further improve the training accuracy on Non-IID data, FedSA deploys adaptive learning rates for workers by their relative participation frequency. We extend our proposed mechanism to the dynamic and multiple learning tasks scenarios. Experimental results on the testbed show that our proposed mechanism and algorithms address the three challenges more effectively than the state-of-the-art solutions.
Author	Xu, Yang Huang, Liusheng Xu, Hongli Huang, He Ma, Qianpiao Jiang, Zhida
Author_xml	– sequence: 1 givenname: Qianpiao orcidid: 0000-0001-8684-3495 surname: Ma fullname: Ma, Qianpiao email: maqiu@mail.ustc.edu.cn organization: School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China – sequence: 2 givenname: Yang orcidid: 0000-0003-0839-3892 surname: Xu fullname: Xu, Yang email: xuyangcs@ustc.edu.cn organization: School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China – sequence: 3 givenname: Hongli surname: Xu fullname: Xu, Hongli email: xuhongli@ustc.edu.cn organization: School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China – sequence: 4 givenname: Zhida surname: Jiang fullname: Jiang, Zhida email: zdjiang@mail.ustc.edu.cn organization: School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China – sequence: 5 givenname: Liusheng surname: Huang fullname: Huang, Liusheng email: lshuang@ustc.edu.cn organization: School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui, China – sequence: 6 givenname: He orcidid: 0000-0003-2768-6607 surname: Huang fullname: Huang, He email: huangh@suda.edu.cn organization: School of Computer Science and Technology, Soochow University, Suzhou, China
BookMark	eNp9kLtOwzAUQC1UJErhAxBLJOYUP_Kw2aKopaAihoIYI9e5aV01drGToX-Po1YMDCy-g8-5VzrXaGSsAYTuCJ4SgsXj66oopxRTMmWE8ISlF2hM0pTHGGM-QmOcMxbznGRX6Nr7HcYkSTgdo6851KviKSqiFbQ6LvzRqK2zxvY-Cl_gZAd1tATpjDab6A3UVhrt20ibaAEdOLsBAwM9qzcQlbY99F0gb9BlI_cebs9zgj7ns49yES_fn1_KYhkrKlgX51LUognvmqk1zxhuFNR5k4gsqyFb5wBNIkkuoeYUM8FVwiFtiBKM5kEENkEPp70HZ7978F21s70z4WRFUyFoiEOTQOUnSjnrvYOmUrqTnbamc1LvK4KroWI1VKyGitW5YjDJH_PgdCvd8V_n_uRoAPjlRZrRlHH2Axdof48
CODEN	ISACEM
CitedBy_id	crossref_primary_10_1109_JSYST_2022_3206404 crossref_primary_10_1109_TCAD_2024_3446881 crossref_primary_10_1109_JSAC_2023_3242710 crossref_primary_10_1109_TMC_2024_3402080 crossref_primary_10_1109_TMC_2024_3461852 crossref_primary_10_1016_j_future_2023_11_001 crossref_primary_10_1109_TMC_2023_3307610 crossref_primary_10_1016_j_patrec_2025_08_009 crossref_primary_10_1016_j_ins_2024_121458 crossref_primary_10_1109_ACCESS_2025_3573806 crossref_primary_10_1002_spe_3265 crossref_primary_10_1109_TNNLS_2024_3354924 crossref_primary_10_1109_TCE_2025_3534468 crossref_primary_10_1109_TWC_2023_3303492 crossref_primary_10_1109_TPDS_2024_3460185 crossref_primary_10_1145_3639825 crossref_primary_10_1145_3690639 crossref_primary_10_1109_TMC_2024_3487294 crossref_primary_10_1016_j_future_2024_107551 crossref_primary_10_1109_JIOT_2024_3406634 crossref_primary_10_1016_j_future_2022_05_003 crossref_primary_10_1109_TCOMM_2024_3425635 crossref_primary_10_1109_TPDS_2024_3379933 crossref_primary_10_1109_TWC_2022_3210434 crossref_primary_10_1109_TCE_2024_3385440 crossref_primary_10_1109_TMC_2022_3186936 crossref_primary_10_1016_j_comnet_2025_111223 crossref_primary_10_1109_TCOMM_2024_3486979 crossref_primary_10_1109_JIOT_2024_3486121 crossref_primary_10_1109_TNSM_2023_3252818 crossref_primary_10_1016_j_comnet_2025_111346 crossref_primary_10_1109_JIOT_2023_3281909 crossref_primary_10_1109_TNSM_2025_3560657 crossref_primary_10_1016_j_future_2024_107528 crossref_primary_10_1109_ACCESS_2023_3335603 crossref_primary_10_1109_TCAD_2023_3346274 crossref_primary_10_1109_TITS_2024_3357138 crossref_primary_10_1109_TVT_2023_3287355 crossref_primary_10_1016_j_adhoc_2023_103153 crossref_primary_10_1109_TWC_2022_3211998 crossref_primary_10_1016_j_future_2024_07_045 crossref_primary_10_1109_ACCESS_2024_3511430 crossref_primary_10_1145_3659205 crossref_primary_10_1007_s12083_023_01616_4 crossref_primary_10_1109_ACCESS_2023_3284976 crossref_primary_10_1109_JSAC_2022_3213344 crossref_primary_10_1109_TWC_2023_3257132 crossref_primary_10_1016_j_ymssp_2024_111837 crossref_primary_10_1109_TMC_2024_3507286 crossref_primary_10_1109_JIOT_2022_3162322 crossref_primary_10_1109_JSAC_2022_3211564 crossref_primary_10_1109_TNET_2024_3424444 crossref_primary_10_1088_1361_6501_acf7da crossref_primary_10_1109_TMC_2024_3366947 crossref_primary_10_1109_TPAMI_2023_3332428 crossref_primary_10_1109_TMC_2024_3416216 crossref_primary_10_1109_TNET_2023_3299851 crossref_primary_10_1016_j_future_2024_07_053 crossref_primary_10_1109_TBDATA_2024_3484651 crossref_primary_10_1109_COMST_2023_3329027 crossref_primary_10_1109_TVT_2023_3331344 crossref_primary_10_1109_TKDE_2023_3332770 crossref_primary_10_1007_s13042_024_02238_9 crossref_primary_10_3390_fi17080375 crossref_primary_10_1016_j_dcan_2024_09_005 crossref_primary_10_1109_TWC_2022_3153495 crossref_primary_10_1109_JIOT_2024_3399404 crossref_primary_10_1109_TIFS_2025_3559411 crossref_primary_10_1109_TITS_2024_3523488 crossref_primary_10_1016_j_engappai_2024_107927 crossref_primary_10_1109_TMC_2024_3418613 crossref_primary_10_1016_j_future_2023_09_008 crossref_primary_10_1109_TWC_2024_3366393 crossref_primary_10_1109_TCE_2023_3330501 crossref_primary_10_1109_ACCESS_2023_3234761 crossref_primary_10_1109_JIOT_2024_3354869 crossref_primary_10_1109_JIOT_2025_3572121 crossref_primary_10_1109_TBDATA_2024_3403381 crossref_primary_10_1007_s13042_023_01956_w crossref_primary_10_1109_TMC_2023_3247798 crossref_primary_10_1109_TMC_2024_3446271 crossref_primary_10_1145_3643560 crossref_primary_10_1109_TWC_2023_3286990 crossref_primary_10_1109_TNET_2023_3329005 crossref_primary_10_1109_ACCESS_2024_3418900 crossref_primary_10_1109_TCCN_2025_3531464 crossref_primary_10_1109_TMC_2025_3551368 crossref_primary_10_1109_TMC_2023_3311188 crossref_primary_10_1109_TMC_2023_3337016 crossref_primary_10_1016_j_future_2024_107683 crossref_primary_10_1038_s41598_024_78239_z crossref_primary_10_1145_3652613 crossref_primary_10_1109_ACCESS_2024_3487073 crossref_primary_10_1109_JSAC_2023_3240716 crossref_primary_10_1109_COMST_2024_3399612 crossref_primary_10_1109_JIOT_2023_3290818 crossref_primary_10_1109_TII_2025_3567377 crossref_primary_10_1109_TWC_2023_3337773 crossref_primary_10_3390_s22041672 crossref_primary_10_1109_TMC_2023_3335403 crossref_primary_10_1109_TCCN_2025_3527711 crossref_primary_10_1109_TCCN_2024_3391329 crossref_primary_10_1016_j_jpdc_2024_104950 crossref_primary_10_1109_TCOMM_2023_3261383 crossref_primary_10_1109_TMC_2024_3492140 crossref_primary_10_1109_TWC_2023_3281765 crossref_primary_10_1016_j_asoc_2025_112925 crossref_primary_10_3390_electronics12204214
Cites_doi	10.1109/MCOM.001.1900103 10.1109/COMST.2017.2745201 10.1109/TSC.2017.2753775 10.1145/3065386 10.1109/ACCESS.2017.2685434 10.24963/ijcai.2019/960 10.1109/JSAC.2019.2904348 10.1109/5.726791 10.1109/ICPADS51040.2020.00030 10.1109/IJCNN52387.2021.9533794 10.1145/2640087.2644155 10.1109/TNNLS.2019.2944481 10.1109/ICC.2019.8761315 10.1109/MLSP.2014.6958872 10.1017/CBO9781107298019 10.1109/TII.2019.2942179 10.1109/INFOCOM41043.2020.9155268 10.1109/INFOCOM.2018.8486403 10.1109/ICDM.2016.0012 10.1109/TSMCC.2009.2032660 10.1109/INFOCOM41043.2020.9155494 10.1109/TC.2020.2994391 10.1109/TWC.2020.3008091 10.1002/9781118548387
ContentType	Journal Article
Copyright	Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2021
Copyright_xml	– notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2021
DBID	97E RIA RIE AAYXX CITATION 7SP 8FD L7M
DOI	10.1109/JSAC.2021.3118435
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef Electronics & Communications Abstracts Technology Research Database Advanced Technologies Database with Aerospace
DatabaseTitle	CrossRef Technology Research Database Advanced Technologies Database with Aerospace Electronics & Communications Abstracts
DatabaseTitleList	Technology Research Database
Database_xml	– sequence: 1 dbid: RIE name: IEEE/IET Electronic Library (IEL) (UW System Shared) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering
EISSN	1558-0008
EndPage	3672
ExternalDocumentID	10_1109_JSAC_2021_3118435 9562538
Genre	orig-research
GrantInformation_xml	– fundername: National Science Foundation of China (NSFC) grantid: 62132019; 61936015; 62102391; U1709217 funderid: 10.13039/501100001809
GroupedDBID	-~X .DC 0R~ 29I 3EH 4.4 41~ 5GY 5VS 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACGFO ACGFS ACIWK ACNCT ADRHT AENEX AETIX AGQYO AGSQL AHBIQ AI. AIBXA AKJIK AKQYR ALLEH ALMA_UNASSIGNED_HOLDINGS ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 EBS EJD HZ~ H~9 IBMZZ ICLAB IES IFIPE IFJZH IPLJI JAVBF LAI M43 O9- OCL P2P RIA RIE RNS TN5 VH1 AAYXX CITATION 7SP 8FD L7M
ID	FETCH-LOGICAL-c293t-7a9d9f7a9b3cb8630fced7f4966de6b7eef4a17aed820398c48e5f1c9327d9fe3
IEDL.DBID	RIE
ISICitedReferencesCount	142
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000720517900009&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	0733-8716
IngestDate	Sat Nov 01 15:08:19 EDT 2025 Sat Nov 29 03:23:04 EST 2025 Tue Nov 18 20:44:37 EST 2025 Wed Aug 27 02:26:04 EDT 2025
IsPeerReviewed	true
IsScholarly	true
Issue	12
Language	English
License	https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c293t-7a9d9f7a9b3cb8630fced7f4966de6b7eef4a17aed820398c48e5f1c9327d9fe3
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
ORCID	0000-0003-0839-3892 0000-0003-2768-6607 0000-0001-8684-3495
PQID	2599211024
PQPubID	85481
PageCount	19
ParticipantIDs	crossref_citationtrail_10_1109_JSAC_2021_3118435 proquest_journals_2599211024 crossref_primary_10_1109_JSAC_2021_3118435 ieee_primary_9562538
PublicationCentury	2000
PublicationDate	2021-12-01
PublicationDateYYYYMMDD	2021-12-01
PublicationDate_xml	– month: 12 year: 2021 text: 2021-12-01 day: 01
PublicationDecade	2020
PublicationPlace	New York
PublicationPlace_xml	– name: New York
PublicationTitle	IEEE journal on selected areas in communications
PublicationTitleAbbrev	J-SAC
PublicationYear	2021
Publisher	IEEE The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml	– name: IEEE – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
References	ref35 ref13 ref37 mohri (ref12) 2019 ref36 ref14 ref30 wang (ref23) 2018 ref33 ref11 krizhevsky (ref38) 2009 ref32 chen (ref20) 2019 kone?n? (ref8) 2016 ref1 ref17 mcmahan (ref9) 2017 ref16 hosmer (ref34) 2013; 398 ref18 anguita (ref2) 2013 mcmahan (ref4) 2017 xie (ref19) 2019 zhao (ref10) 2018 li (ref25) 2018 li (ref15) 2014 ref24 ref26 ref22 ref21 ref28 ref29 ref7 nesterov (ref31) 2003; 87 ref3 ref6 ref5 chai (ref27) 2020
References_xml	– year: 2019 ident: ref19 article-title: Asynchronous federated optimization publication-title: arXiv 1903 03934 – ident: ref3 doi: 10.1109/MCOM.001.1900103 – ident: ref13 doi: 10.1109/COMST.2017.2745201 – ident: ref16 doi: 10.1109/TSC.2017.2753775 – ident: ref36 doi: 10.1145/3065386 – ident: ref14 doi: 10.1109/ACCESS.2017.2685434 – ident: ref5 doi: 10.24963/ijcai.2019/960 – ident: ref17 doi: 10.1109/JSAC.2019.2904348 – ident: ref37 doi: 10.1109/5.726791 – year: 2017 ident: ref4 publication-title: Google – start-page: 3 year: 2013 ident: ref2 article-title: A public domain dataset for human activity recognition using smartphones publication-title: Proc ESANN – ident: ref29 doi: 10.1109/ICPADS51040.2020.00030 – volume: 87 year: 2003 ident: ref31 publication-title: Introductory Lectures on Convex Optimization A Basic Course – ident: ref21 doi: 10.1109/IJCNN52387.2021.9533794 – year: 2009 ident: ref38 publication-title: Learning multiple layers of features from tiny images – ident: ref7 doi: 10.1145/2640087.2644155 – year: 2018 ident: ref25 article-title: Federated optimization in heterogeneous networks publication-title: arXiv 1812 06127 – ident: ref11 doi: 10.1109/TNNLS.2019.2944481 – ident: ref6 doi: 10.1109/ICC.2019.8761315 – year: 2018 ident: ref23 article-title: Cooperative SGD: A unified framework for the design and analysis of communication-efficient SGD algorithms publication-title: arXiv 1808 07576 – ident: ref28 doi: 10.1109/MLSP.2014.6958872 – ident: ref35 doi: 10.1017/CBO9781107298019 – year: 2020 ident: ref27 article-title: FedAT: A communication-efficient federated learning method with asynchronous tiers under non-IID data publication-title: arXiv 2010 05958 – start-page: 1273 year: 2017 ident: ref9 article-title: Communication-efficient learning of deep networks from decentralized data publication-title: Proc AISTATS – ident: ref26 doi: 10.1109/TII.2019.2942179 – ident: ref33 doi: 10.1109/INFOCOM41043.2020.9155268 – ident: ref32 doi: 10.1109/INFOCOM.2018.8486403 – year: 2016 ident: ref8 article-title: Federated optimization: Distributed machine learning for on-device intelligence publication-title: arXiv 1610 02527 – start-page: 19 year: 2014 ident: ref15 article-title: Communication efficient distributed machine learning with the parameter server publication-title: Proc Adv Neural Inf Process Syst – ident: ref30 doi: 10.1109/ICDM.2016.0012 – ident: ref1 doi: 10.1109/TSMCC.2009.2032660 – year: 2018 ident: ref10 article-title: Federated learning with non-IID data publication-title: arXiv 1806 00582 – ident: ref18 doi: 10.1109/INFOCOM41043.2020.9155494 – start-page: 4615 year: 2019 ident: ref12 article-title: Agnostic federated learning publication-title: Proc Int Conf Mach Learn – year: 2019 ident: ref20 article-title: Asynchronous online federated learning for edge devices with non-IID data publication-title: arXiv 1911 02134 – ident: ref22 doi: 10.1109/TC.2020.2994391 – ident: ref24 doi: 10.1109/TWC.2020.3008091 – volume: 398 year: 2013 ident: ref34 publication-title: Applied Logistic Regression doi: 10.1002/9781118548387
SSID	ssj0014482
Score	2.6857984
Snippet	Federated learning (FL) involves training machine learning models over distributed edge nodes ( i.e. , workers) while facing three critical challenges, edge...
SourceID	proquest crossref ieee
SourceType	Aggregation Database Enrichment Source Index Database Publisher
StartPage	3654
SubjectTerms	Accuracy Adaptive learning Algorithms Analytical models Cognitive tasks Collaborative work Completion time Computational modeling Convergence Data models Edge computing Federated learning Heterogeneity Machine learning Mathematical models non-IID Parameters semi-asynchronous mechanism Servers Staff participation Training data
Title	FedSA: A Semi-Asynchronous Federated Learning Mechanism in Heterogeneous Edge Computing
URI	https://ieeexplore.ieee.org/document/9562538 https://www.proquest.com/docview/2599211024
Volume	39
WOSCitedRecordID	wos000720517900009&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVIEE databaseName: IEEE/IET Electronic Library (IEL) (UW System Shared) customDbUrl: eissn: 1558-0008 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0014482 issn: 0733-8716 databaseCode: RIE dateStart: 19830101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1La9wwEB6SpYfk0EcedNtt0SGnUCe2JVtSbyZkCYWGwqY0NyNLo7CQ9ZbsppB_nxnbWVoaCr0YH2aM0cia-Twz3wAcNRGLJpUqcZiHRCGdg84amWjlSYMCXOVUN2xCX16a62v7bQs-bXphELErPsMTvu1y-WHp7_lX2anlaF2abdjWuux7tTYZA4IZXcZAS5kwCBgymFlqT7_MqjNCgnlGAJXHmxR_-KBuqMpfJ3HnXqav_u_FXsPLIYwUVW_3N7CF7R7s_kYuuA8_phhm1WdRiRku5km1emg9M-ES1BdTppCgKDOIgV_1RnxF7gGerxZi3ooLLpJZ0t5Clj4PNyj68Q8keQDfp-dXZxfJMEYh8eTL14l2NthI10b6xpQyjR6DjoqATsCy0YhRuUw7DBQNSGu8MljEzFNkp0kR5SGM2mWLb0GULmgdi6Z00jNRu3E5FgpzY6LOYpONIX1a2NoPHOM86uK27rBGamu2Rc22qAdbjOF4o_KzJ9j4l_A-L_5GcFj3MUyerFcPn-CqJlxnGd3m6t3zWu9hh5_d16ZMYLS-u8cP8ML_Ws9Xdx-73fUIqvDM1g
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Ra9RAEB5qFawPam3Fs1X3oU9ibJLd3O72LZQep7aHcC3tW9jszpYDmyu9q-C_dyZJD0tF8CXkYYaEnc3OfJmZbwD26ohFnUqVOMxDopDOQWeNTLTypEEBrnKqHTahJxNzcWG_r8GnVS8MIrbFZ_iZb9tcfpj7W_5Vtm85WpfmETwulMrTrltrlTMgoNHmDLSUCcOAPoeZpXb_67Q8JCyYZwRRecBJcc8LtWNVHpzFrYMZvfi_V3sJz_tAUpSd5TdhDZtX8OwPesEtOB9hmJYHohRTvJol5eJX45kLl8C-GDGJBMWZQfQMq5fiBLkLeLa4ErNGjLlMZk67C1n6KFyi6AZAkOQ2nI2OTg_HST9IIfHkzZeJdjbYSNda-toMZRo9Bh0VQZ2Aw1ojRuUy7TBQPCCt8cpgETNPsZ0mRZSvYb2ZN_gGxNAFrWNRD530TNVuXI6FwtyYqLNYZwNI7xa28j3LOA-7-FG1aCO1FduiYltUvS0G8HGlct1RbPxLeIsXfyXYr_sAdu-sV_Uf4aIiZGcZ3-bq7d-1PsDT8enJcXX8ZfJtBzb4OV2lyi6sL29u8R088T-Xs8XN-3an_QZ8_dAd
openUrl	ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=FedSA%3A+A+Semi-Asynchronous+Federated+Learning+Mechanism+in+Heterogeneous+Edge+Computing&rft.jtitle=IEEE+journal+on+selected+areas+in+communications&rft.au=Ma%2C+Qianpiao&rft.au=Xu%2C+Yang&rft.au=Xu%2C+Hongli&rft.au=Jiang%2C+Zhida&rft.date=2021-12-01&rft.pub=The+Institute+of+Electrical+and+Electronics+Engineers%2C+Inc.+%28IEEE%29&rft.issn=0733-8716&rft.eissn=1558-0008&rft.volume=39&rft.issue=12&rft.spage=3654&rft_id=info:doi/10.1109%2FJSAC.2021.3118435&rft.externalDBID=NO_FULL_TEXT
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0733-8716&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0733-8716&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0733-8716&client=summon