Edge/Cloud Infinite-Time Horizon Resource Allocation for Distributed Machine Learning and General Tasks

Edge computing has emerged as a computing paradigm where the application and data processing takes place close to the end devices. It decreases the distances over which data transfers are made, offering reduced delay and fast speed of action for general data processing and store/retrieve jobs. The b...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE eTransactions on network and service management 2024-02, Vol.21 (1), p.697-713
Hauptverfasser: Sartzetakis, Ippokratis, Soumplis, Polyzois, Pantazopoulos, Panagiotis, Katsaros, Konstantinos V., Sourlas, Vasilis, Varvarigos, Emmanouel
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 713
container_issue 1
container_start_page 697
container_title IEEE eTransactions on network and service management
container_volume 21
creator Sartzetakis, Ippokratis
Soumplis, Polyzois
Pantazopoulos, Panagiotis
Katsaros, Konstantinos V.
Sourlas, Vasilis
Varvarigos, Emmanouel
description Edge computing has emerged as a computing paradigm where the application and data processing takes place close to the end devices. It decreases the distances over which data transfers are made, offering reduced delay and fast speed of action for general data processing and store/retrieve jobs. The benefits of edge computing can also be reaped for distributed computation algorithms, where the cloud also plays an assistive role. In this context, an important challenge is to allocate the required resources at both edge and cloud to carry out the processing of data that are generated over a continuous ("infinite") time horizon. This is a complex problem due to the variety of requirements (resource needs, accuracy, delay, etc.) that may be posed by each computation algorithm, as well as the heterogeneous resources' features (e.g., processing, bandwidth). In this work, we develop a solution for serving weakly coupled general distributed algorithms, with emphasis on machine learning algorithms, at the edge and/or the cloud. We present a dual-objective Integer Linear Programming formulation that optimizes monetary cost and computation accuracy. We also introduce efficient heuristics to perform the resource allocation. We examine various distributed ML allocation scenarios using realistic parameters from actual vendors. We quantify trade-offs related to accuracy, performance and cost of edge/cloud bandwidth and processing resources. Our results indicate that among the many parameters of interest, the processing costs seem to play the most important role for the allocation decisions. Finally, we explore interesting interactions between target accuracy, monetary cost and delay.
doi_str_mv 10.1109/TNSM.2023.3312593
format Article
fullrecord <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_crossref_primary_10_1109_TNSM_2023_3312593</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10242079</ieee_id><sourcerecordid>2923116971</sourcerecordid><originalsourceid>FETCH-LOGICAL-c246t-4587998a766cf4f53fffba6faa07c65a98f6e6b634bcd68bb35a3aefe1f80b343</originalsourceid><addsrcrecordid>eNpNkM1OwzAQhC0EEqXwAEgcLHFO65_EiY9VKW2lFiQoZ8tx1sUldYqdHODpSVUOPe1oNbM7-hC6p2REKZHjzcv7esQI4yPOKcskv0ADKjlL0oznl2f6Gt3EuCMkK6hkA7SdVVsYT-umq_DSW-ddC8nG7QEvmuB-G4_fIDZdMIAndd0Y3bp-Z5uAn1xsgyu7Fiq81ubTecAr0ME7v8XaV3gOHoKu8UbHr3iLrqyuI9z9zyH6eJ5tpotk9TpfTierxLBUtH3DIpey0LkQxqY249baUgurNcmNyLQsrABRCp6WphJFWfJMcw0WqC1IyVM-RI-nu4fQfHcQW7Xr2_v-pWKScUqFzGnvoieXCU2MAaw6BLfX4UdRoo481ZGnOvJU_zz7zMMp4wDgzM9SRnLJ_wBh73LH</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2923116971</pqid></control><display><type>article</type><title>Edge/Cloud Infinite-Time Horizon Resource Allocation for Distributed Machine Learning and General Tasks</title><source>IEEE Electronic Library (IEL)</source><creator>Sartzetakis, Ippokratis ; Soumplis, Polyzois ; Pantazopoulos, Panagiotis ; Katsaros, Konstantinos V. ; Sourlas, Vasilis ; Varvarigos, Emmanouel</creator><creatorcontrib>Sartzetakis, Ippokratis ; Soumplis, Polyzois ; Pantazopoulos, Panagiotis ; Katsaros, Konstantinos V. ; Sourlas, Vasilis ; Varvarigos, Emmanouel</creatorcontrib><description>Edge computing has emerged as a computing paradigm where the application and data processing takes place close to the end devices. It decreases the distances over which data transfers are made, offering reduced delay and fast speed of action for general data processing and store/retrieve jobs. The benefits of edge computing can also be reaped for distributed computation algorithms, where the cloud also plays an assistive role. In this context, an important challenge is to allocate the required resources at both edge and cloud to carry out the processing of data that are generated over a continuous ("infinite") time horizon. This is a complex problem due to the variety of requirements (resource needs, accuracy, delay, etc.) that may be posed by each computation algorithm, as well as the heterogeneous resources' features (e.g., processing, bandwidth). In this work, we develop a solution for serving weakly coupled general distributed algorithms, with emphasis on machine learning algorithms, at the edge and/or the cloud. We present a dual-objective Integer Linear Programming formulation that optimizes monetary cost and computation accuracy. We also introduce efficient heuristics to perform the resource allocation. We examine various distributed ML allocation scenarios using realistic parameters from actual vendors. We quantify trade-offs related to accuracy, performance and cost of edge/cloud bandwidth and processing resources. Our results indicate that among the many parameters of interest, the processing costs seem to play the most important role for the allocation decisions. Finally, we explore interesting interactions between target accuracy, monetary cost and delay.</description><identifier>ISSN: 1932-4537</identifier><identifier>EISSN: 1932-4537</identifier><identifier>DOI: 10.1109/TNSM.2023.3312593</identifier><identifier>CODEN: ITNSC4</identifier><language>eng</language><publisher>New York: IEEE</publisher><subject>Accuracy ; Algorithms ; Cloud and edge computing ; Cloud computing ; Computational modeling ; Costs ; Data processing ; Delay ; distributed computing ; distributed machine learning ; Edge computing ; inference ; Integer programming ; Internet of Things ; Linear programming ; Machine learning ; Parameters ; Resource allocation ; Resource management ; Task analysis ; Training</subject><ispartof>IEEE eTransactions on network and service management, 2024-02, Vol.21 (1), p.697-713</ispartof><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c246t-4587998a766cf4f53fffba6faa07c65a98f6e6b634bcd68bb35a3aefe1f80b343</cites><orcidid>0000-0001-8908-7867 ; 0000-0003-0633-5859 ; 0000-0003-0725-5463 ; 0000-0003-1041-7750</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/10242079$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>315,782,786,798,27931,27932,54765</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/10242079$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Sartzetakis, Ippokratis</creatorcontrib><creatorcontrib>Soumplis, Polyzois</creatorcontrib><creatorcontrib>Pantazopoulos, Panagiotis</creatorcontrib><creatorcontrib>Katsaros, Konstantinos V.</creatorcontrib><creatorcontrib>Sourlas, Vasilis</creatorcontrib><creatorcontrib>Varvarigos, Emmanouel</creatorcontrib><title>Edge/Cloud Infinite-Time Horizon Resource Allocation for Distributed Machine Learning and General Tasks</title><title>IEEE eTransactions on network and service management</title><addtitle>T-NSM</addtitle><description>Edge computing has emerged as a computing paradigm where the application and data processing takes place close to the end devices. It decreases the distances over which data transfers are made, offering reduced delay and fast speed of action for general data processing and store/retrieve jobs. The benefits of edge computing can also be reaped for distributed computation algorithms, where the cloud also plays an assistive role. In this context, an important challenge is to allocate the required resources at both edge and cloud to carry out the processing of data that are generated over a continuous ("infinite") time horizon. This is a complex problem due to the variety of requirements (resource needs, accuracy, delay, etc.) that may be posed by each computation algorithm, as well as the heterogeneous resources' features (e.g., processing, bandwidth). In this work, we develop a solution for serving weakly coupled general distributed algorithms, with emphasis on machine learning algorithms, at the edge and/or the cloud. We present a dual-objective Integer Linear Programming formulation that optimizes monetary cost and computation accuracy. We also introduce efficient heuristics to perform the resource allocation. We examine various distributed ML allocation scenarios using realistic parameters from actual vendors. We quantify trade-offs related to accuracy, performance and cost of edge/cloud bandwidth and processing resources. Our results indicate that among the many parameters of interest, the processing costs seem to play the most important role for the allocation decisions. Finally, we explore interesting interactions between target accuracy, monetary cost and delay.</description><subject>Accuracy</subject><subject>Algorithms</subject><subject>Cloud and edge computing</subject><subject>Cloud computing</subject><subject>Computational modeling</subject><subject>Costs</subject><subject>Data processing</subject><subject>Delay</subject><subject>distributed computing</subject><subject>distributed machine learning</subject><subject>Edge computing</subject><subject>inference</subject><subject>Integer programming</subject><subject>Internet of Things</subject><subject>Linear programming</subject><subject>Machine learning</subject><subject>Parameters</subject><subject>Resource allocation</subject><subject>Resource management</subject><subject>Task analysis</subject><subject>Training</subject><issn>1932-4537</issn><issn>1932-4537</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNpNkM1OwzAQhC0EEqXwAEgcLHFO65_EiY9VKW2lFiQoZ8tx1sUldYqdHODpSVUOPe1oNbM7-hC6p2REKZHjzcv7esQI4yPOKcskv0ADKjlL0oznl2f6Gt3EuCMkK6hkA7SdVVsYT-umq_DSW-ddC8nG7QEvmuB-G4_fIDZdMIAndd0Y3bp-Z5uAn1xsgyu7Fiq81ubTecAr0ME7v8XaV3gOHoKu8UbHr3iLrqyuI9z9zyH6eJ5tpotk9TpfTierxLBUtH3DIpey0LkQxqY249baUgurNcmNyLQsrABRCp6WphJFWfJMcw0WqC1IyVM-RI-nu4fQfHcQW7Xr2_v-pWKScUqFzGnvoieXCU2MAaw6BLfX4UdRoo481ZGnOvJU_zz7zMMp4wDgzM9SRnLJ_wBh73LH</recordid><startdate>20240201</startdate><enddate>20240201</enddate><creator>Sartzetakis, Ippokratis</creator><creator>Soumplis, Polyzois</creator><creator>Pantazopoulos, Panagiotis</creator><creator>Katsaros, Konstantinos V.</creator><creator>Sourlas, Vasilis</creator><creator>Varvarigos, Emmanouel</creator><general>IEEE</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope><orcidid>https://orcid.org/0000-0001-8908-7867</orcidid><orcidid>https://orcid.org/0000-0003-0633-5859</orcidid><orcidid>https://orcid.org/0000-0003-0725-5463</orcidid><orcidid>https://orcid.org/0000-0003-1041-7750</orcidid></search><sort><creationdate>20240201</creationdate><title>Edge/Cloud Infinite-Time Horizon Resource Allocation for Distributed Machine Learning and General Tasks</title><author>Sartzetakis, Ippokratis ; Soumplis, Polyzois ; Pantazopoulos, Panagiotis ; Katsaros, Konstantinos V. ; Sourlas, Vasilis ; Varvarigos, Emmanouel</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c246t-4587998a766cf4f53fffba6faa07c65a98f6e6b634bcd68bb35a3aefe1f80b343</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Accuracy</topic><topic>Algorithms</topic><topic>Cloud and edge computing</topic><topic>Cloud computing</topic><topic>Computational modeling</topic><topic>Costs</topic><topic>Data processing</topic><topic>Delay</topic><topic>distributed computing</topic><topic>distributed machine learning</topic><topic>Edge computing</topic><topic>inference</topic><topic>Integer programming</topic><topic>Internet of Things</topic><topic>Linear programming</topic><topic>Machine learning</topic><topic>Parameters</topic><topic>Resource allocation</topic><topic>Resource management</topic><topic>Task analysis</topic><topic>Training</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Sartzetakis, Ippokratis</creatorcontrib><creatorcontrib>Soumplis, Polyzois</creatorcontrib><creatorcontrib>Pantazopoulos, Panagiotis</creatorcontrib><creatorcontrib>Katsaros, Konstantinos V.</creatorcontrib><creatorcontrib>Sourlas, Vasilis</creatorcontrib><creatorcontrib>Varvarigos, Emmanouel</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><collection>CrossRef</collection><jtitle>IEEE eTransactions on network and service management</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Sartzetakis, Ippokratis</au><au>Soumplis, Polyzois</au><au>Pantazopoulos, Panagiotis</au><au>Katsaros, Konstantinos V.</au><au>Sourlas, Vasilis</au><au>Varvarigos, Emmanouel</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Edge/Cloud Infinite-Time Horizon Resource Allocation for Distributed Machine Learning and General Tasks</atitle><jtitle>IEEE eTransactions on network and service management</jtitle><stitle>T-NSM</stitle><date>2024-02-01</date><risdate>2024</risdate><volume>21</volume><issue>1</issue><spage>697</spage><epage>713</epage><pages>697-713</pages><issn>1932-4537</issn><eissn>1932-4537</eissn><coden>ITNSC4</coden><abstract>Edge computing has emerged as a computing paradigm where the application and data processing takes place close to the end devices. It decreases the distances over which data transfers are made, offering reduced delay and fast speed of action for general data processing and store/retrieve jobs. The benefits of edge computing can also be reaped for distributed computation algorithms, where the cloud also plays an assistive role. In this context, an important challenge is to allocate the required resources at both edge and cloud to carry out the processing of data that are generated over a continuous ("infinite") time horizon. This is a complex problem due to the variety of requirements (resource needs, accuracy, delay, etc.) that may be posed by each computation algorithm, as well as the heterogeneous resources' features (e.g., processing, bandwidth). In this work, we develop a solution for serving weakly coupled general distributed algorithms, with emphasis on machine learning algorithms, at the edge and/or the cloud. We present a dual-objective Integer Linear Programming formulation that optimizes monetary cost and computation accuracy. We also introduce efficient heuristics to perform the resource allocation. We examine various distributed ML allocation scenarios using realistic parameters from actual vendors. We quantify trade-offs related to accuracy, performance and cost of edge/cloud bandwidth and processing resources. Our results indicate that among the many parameters of interest, the processing costs seem to play the most important role for the allocation decisions. Finally, we explore interesting interactions between target accuracy, monetary cost and delay.</abstract><cop>New York</cop><pub>IEEE</pub><doi>10.1109/TNSM.2023.3312593</doi><tpages>17</tpages><orcidid>https://orcid.org/0000-0001-8908-7867</orcidid><orcidid>https://orcid.org/0000-0003-0633-5859</orcidid><orcidid>https://orcid.org/0000-0003-0725-5463</orcidid><orcidid>https://orcid.org/0000-0003-1041-7750</orcidid></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 1932-4537
ispartof IEEE eTransactions on network and service management, 2024-02, Vol.21 (1), p.697-713
issn 1932-4537
1932-4537
language eng
recordid cdi_crossref_primary_10_1109_TNSM_2023_3312593
source IEEE Electronic Library (IEL)
subjects Accuracy
Algorithms
Cloud and edge computing
Cloud computing
Computational modeling
Costs
Data processing
Delay
distributed computing
distributed machine learning
Edge computing
inference
Integer programming
Internet of Things
Linear programming
Machine learning
Parameters
Resource allocation
Resource management
Task analysis
Training
title Edge/Cloud Infinite-Time Horizon Resource Allocation for Distributed Machine Learning and General Tasks
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-05T08%3A36%3A51IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Edge/Cloud%20Infinite-Time%20Horizon%20Resource%20Allocation%20for%20Distributed%20Machine%20Learning%20and%20General%20Tasks&rft.jtitle=IEEE%20eTransactions%20on%20network%20and%20service%20management&rft.au=Sartzetakis,%20Ippokratis&rft.date=2024-02-01&rft.volume=21&rft.issue=1&rft.spage=697&rft.epage=713&rft.pages=697-713&rft.issn=1932-4537&rft.eissn=1932-4537&rft.coden=ITNSC4&rft_id=info:doi/10.1109/TNSM.2023.3312593&rft_dat=%3Cproquest_RIE%3E2923116971%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2923116971&rft_id=info:pmid/&rft_ieee_id=10242079&rfr_iscdi=true