[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-290607-105":59,"doc-detail-290607-en":131},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","scheduling-jobs-across-geo-distributed-datacenters","Scheduling Jobs Across Geo-distributed Datacenters","","Growing data volumes spread across geo-distributed datacenters make centralized aggregation inefficient, especially when bandwidth costs and performance degrade as jobs require data from multiple sites. A distributed execution trend moves computation to data locations, but raises coordination challenges for job scheduling across datacenters. This paper develops low-overhead job scheduling algorithms that coordinate execution across datacenters while achieving near-optimal performance, validated through simulation with realistic job traces, delivering up to 50% lower average completion time versus SRPT.",{"@graph":69,"@context":123},[70,84,106],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/scheduling-jobs-across-geo-distributed-datacenters/290607/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/scheduling-jobs-across-geo-distributed-datacenters/290607.png","ImageObject",300,407,{"name":92,"@type":93},"Pentious","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-19","2026-09-17",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",13,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"Why are centralized job execution approaches becoming inefficient for geo-distributed datacenters?","Question",{"text":113,"@type":114},"Because data volumes increase and jobs often require data from multiple datacenters, centralized approaches cause substantial network traffic and increased job completion time.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"What is the main idea behind the distributed job execution trend?",{"text":118,"@type":114},"Run tasks at the datacenters where the needed data are stored, and only aggregate results when the job completes.",{"name":120,"@type":111,"acceptedAnswer":121},"How much improvement do the proposed algorithms achieve over SRPT-based approaches?",{"text":122,"@type":114},"Simulation with realistic job traces shows up to 50% improvement in average job completion time compared with SRPT-based approaches.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},290607,1789642188,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":140,"language":141,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":67,"update_tm":130,"read_time":145,"preview_page_count":4},1374404730887,"https://ap-avatar.wpscdn.com/davatar_6f874abed73319feea01a86fa6f0fab8","Scheduling Jobs Across Geo-distributed Datacenters  \nChien-Chun Hung, Leana Golubchik, Minlan Yu  \nUniversity of Southern California  \n{chienchun. hung, leana, [minlanyu](minlanyu}@usc.edu)[}](minlanyu}@usc.edu)[@usc.edu](minlanyu}@usc.edu)  \nAbstract  \nWith growing data volumes generated and stored across geodistributed datacenters, it is becoming increasingly inef􀀂 -cient to aggregate all data required for computation at a single datacenter. Instead, a recent trend is to distribute computation to take advantage of data locality, thus reducing the resource (e.g., bandwidth) costs while improving performance. In this trend, new challenges are emerging in job scheduling, which requires coordination among the datacenters as each job runs across geo-distributed sites. In this paper, we propose novel job scheduling algorithms that coordinate job scheduling across datacenters with low overhead, while achieving near-optimal performance. Our extensive simulation study with realistic job traces shows that the proposed scheduling algorithms result in up to 50% improvement in average job completion time over the Shortest Remaining Processing Time (SRPT) based approaches.  \n1. Introduction  \nData intensive jobs run by cluster computing systems (e.g., Hadoop[3], Spark[39], Dryad[17]) have recently generated signi􀀂cant workloads for datacenters, providing services such as web search, consumer advertisements and product recommendations, user behavior analysis and business intelligence. These jobs are composed of numerous tasks. Each task reads a partition of input data and runs on available computing slots in parallel; the job is 􀀂nished upon the completion of all of its tasks [5, 7, 8] . To serve the increasing demands of various data analytics applications, major cloud providers like Amazon[1], Microsoft[4] and Google[19] each deploy from tens to hundreds of geo-distributed datacenters; AT&T has thousands of datacenters at their PoP locations.  \nPermission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for pro􀀂t or commercial advantage and that copies bear this notice and the full citation on the 􀀂rst page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior speci􀀂c permission and/or a fee. Request permissions [from permissions@acm.org](from permissions@acm.org).  \nSoCC’15, August 27-29, 2015, Kohala Coast, HI, USA.  \nCopyright 􀀍c 2015 ACM ISBN 978-1-4503-3651-2/15/08...$15.00 . [http://dx.doi.org/10.1145/2806777.2806780](http://dx.doi.org/10.1145/2806777.2806780)  \nConventional approaches perform centralized job execution, with each job running within a single datacenter. In such case, when a job needs data from multiple datacenters, atypical approach is to 􀀂rst collect all the required data from multiple datacenters at a single location, and then run the computation at that datacenter [11, 13, 16, 20, 25] . However, as data volumes continue to grow in an unprecedented manner, such an approach results in substantial network traf-􀀂c [27, 35, 36] and increased job completion time [14] . Moreover, it is becoming increasingly impractical to replicate a large data set across multiple datacenters [24]. Finally, some data are restricted to certain datacenters due to security and privacy constraints (e.g., must be kept within a particular nation [35, 36]), and therefore cannot be moved.  \nConsequently, instead of data aggregation at a single datacenter, a recent trend is to conduct distributed job execution, i.e., running ajob’s tasks at the datacenters where the needed data are stored, and only aggregating the results at job completion time. Recent research efforts show that distributed job execution achieves 250× bandwidth savings [35, 36] and reduces the 90th − percentil","cbCaik8DP1RqRzVL","https://ap.wps.com/l/cbCaik8DP1RqRzVL","pdf",584998,14,"English","# Abstract\n# Introduction","[{\"question\":\"Why are centralized job execution approaches becoming inefficient for geo-distributed datacenters?\",\"answer\":\"Because data volumes increase and jobs often require data from multiple datacenters, centralized approaches cause substantial network traffic and increased job completion time.\"},{\"question\":\"What is the main idea behind the distributed job execution trend?\",\"answer\":\"Run tasks at the datacenters where the needed data are stored, and only aggregate results when the job completes.\"},{\"question\":\"How much improvement do the proposed algorithms achieve over SRPT-based approaches?\",\"answer\":\"Simulation with realistic job traces shows up to 50% improvement in average job completion time compared with SRPT-based approaches.\"}]","Scheduling Jobs Across Geo-distributed Datacenters | PDF",35]