{"id":168656,"date":"2015-08-17T00:00:00","date_gmt":"2015-08-17T00:00:00","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/msr-research-item\/congestion-control-for-large-scale-rdma-deployments\/"},"modified":"2018-10-16T20:39:53","modified_gmt":"2018-10-17T03:39:53","slug":"congestion-control-for-large-scale-rdma-deployments","status":"publish","type":"msr-research-item","link":"https:\/\/www.microsoft.com\/en-us\/research\/publication\/congestion-control-for-large-scale-rdma-deployments\/","title":{"rendered":"Congestion Control for Large-Scale RDMA Deployments"},"content":{"rendered":"<div class=\"asset-content\">\n<p>Modern datacenter applications demand high throughput (40Gbps) and ultra-low latency (<em><<\/em> 10 microsecond per hop) from the network, with low CPU overhead. Standard TCP\/IP stacks cannot meet these requirements, but Remote Direct Memory Access (RDMA) can. On IP-routed datacenter networks, RDMA is deployed using RoCEv2 protocol, which relies on Priority-based Flow Control (PFC) to enable a drop-free network. However, PFC can lead to poor application performance due to problems like head-of-line blocking and unfairness. To alleviates these problems, we introduce DCQCN, an end-to-end congestion control scheme for RoCEv2. To optimize DCQCN performance, we build a fluid model, and provide guidelines for tuning switch buffer thresholds, and other protocol parameters. Using a 3-tier Clos network testbed, we show that DCQCN dramatically improves throughput and fairness of RoCEv2 RDMA traffic. DCQCN is implemented in Mellanox NICs, and is being deployed in Microsoft\u2019s datacenters.<\/p>\n<\/div>\n<p><!-- .asset-content --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Modern datacenter applications demand high throughput (40Gbps) and ultra-low latency (<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","footnotes":""},"msr-content-type":[3],"msr-research-highlight":[],"research-area":[13547],"msr-publication-type":[193716],"msr-product-type":[],"msr-focus-area":[],"msr-platform":[],"msr-download-source":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-168656","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-systems-and-networking","msr-locale-en_us"],"msr_publishername":"ACM - Association for Computing Machinery","msr_edition":"SIGCOMM","msr_affiliation":"","msr_published_date":"2015-08-17","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"204204","msr_publicationurl":"http:\/\/conferences.sigcomm.org\/sigcomm\/2015\/pdf\/papers\/p523.pdf","msr_doi":"10.1145\/2785956.2787484","msr_publication_uploader":[{"type":"file","title":"lossless.pdf","viewUrl":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2016\/02\/lossless.pdf","id":204204,"label_id":0},{"type":"url","title":"http:\/\/conferences.sigcomm.org\/sigcomm\/2015\/pdf\/papers\/p523.pdf","viewUrl":false,"id":false,"label_id":0},{"type":"doi","title":"10.1145\/2785956.2787484","viewUrl":false,"id":false,"label_id":0}],"msr_related_uploader":"","msr_attachments":[{"id":0,"url":"http:\/\/conferences.sigcomm.org\/sigcomm\/2015\/pdf\/papers\/p523.pdf"}],"msr-author-ordering":[{"type":"user_nicename","value":"yibzh","user_id":36065,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=yibzh"},{"type":"text","value":"Yibo Zhu","user_id":0,"rest_url":false},{"type":"text","value":"Haggai Eran","user_id":0,"rest_url":false},{"type":"user_nicename","value":"fstone","user_id":35969,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=fstone"},{"type":"text","value":"Daniel Firestone","user_id":0,"rest_url":false},{"type":"user_nicename","value":"chguo","user_id":31398,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=chguo"},{"type":"text","value":"Marina Lipshteyn","user_id":0,"rest_url":false},{"type":"text","value":"Yehonatan Liron","user_id":0,"rest_url":false},{"type":"text","value":"Jitendra Padhye","user_id":0,"rest_url":false},{"type":"text","value":"Shachar Raindel","user_id":0,"rest_url":false},{"type":"text","value":"Mohamad Haj Yahia","user_id":0,"rest_url":false},{"type":"user_nicename","value":"mzh","user_id":33042,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=mzh"},{"type":"user_nicename","value":"padhye","user_id":33179,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=padhye"}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[],"msr_group":[],"msr_project":[171491],"publication":[],"video":[],"download":[],"msr_publication_type":"inproceedings","related_content":{"projects":[{"ID":171491,"post_title":"DCQCN: Data Center QCN","post_name":"dcqcn-data-center-qcn","post_type":"msr-project","post_date":"2015-08-11 14:24:41","post_modified":"2016-05-25 11:20:38","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/dcqcn-data-center-qcn\/","post_excerpt":"DCQCN is a congestion control protocol for large scale RDMA networks, developed jointly by Microsoft and Mellanox.","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/171491"}]}}]},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/168656"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":1,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/168656\/revisions"}],"predecessor-version":[{"id":439200,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/168656\/revisions\/439200"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=168656"}],"wp:term":[{"taxonomy":"msr-content-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-content-type?post=168656"},{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=168656"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=168656"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=168656"},{"taxonomy":"msr-product-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-product-type?post=168656"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=168656"},{"taxonomy":"msr-platform","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-platform?post=168656"},{"taxonomy":"msr-download-source","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-download-source?post=168656"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=168656"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=168656"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=168656"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=168656"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=168656"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=168656"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=168656"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}