Optically Connected Memory for Disaggregated Data Centers

Recent advances in integrated photonics enable the implementation of reconfigurable, high-bandwidth, and low energy-per-bit interconnects in next-generation data centers. We propose and evaluate an Optically Connected Memory (OCM) architecture that disaggregates the main memory from the computation...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2020-08
Main Authors: Gonzalez, Jorge, Gazman, Alexander, Hattink, Maarten, Palma, Mauricio G, Bahadori, Meisam, Rubio-Noriega, Ruth, Orosa, Lois, Glick, Madeleine, Mutlu, Onur, Bergman, Keren, Azevedo, Rodolfo
Format: Article
Language:English
Subjects:
Computation
Computer centers
Computer simulation
Connectors
Data centers
Dynamic random access memory
Energy consumption
Nodes
Performance evaluation
Photonics
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Gonzalez, Jorge
Gazman, Alexander
Hattink, Maarten
Palma, Mauricio G
Bahadori, Meisam
Rubio-Noriega, Ruth
Orosa, Lois
Glick, Madeleine
Mutlu, Onur
Bergman, Keren
Azevedo, Rodolfo
description Recent advances in integrated photonics enable the implementation of reconfigurable, high-bandwidth, and low energy-per-bit interconnects in next-generation data centers. We propose and evaluate an Optically Connected Memory (OCM) architecture that disaggregates the main memory from the computation nodes in data centers. OCM is based on micro-ring resonators (MRRs), and it does not require any modification to the DRAM memory modules. We calculate energy consumption from real photonic devices and integrate them into a system simulator to evaluate performance. Our results show that (1) OCM is capable of interconnecting four DDR4 memory channels to a computing node using two fibers with 1.07 pJ energy-per-bit consumption and (2) OCM performs up to 5.5x faster than a disaggregated memory with 40G PCIe NIC connectors to computing nodes.
format article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2437282180</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2437282180</sourcerecordid><originalsourceid>FETCH-proquest_journals_24372821803</originalsourceid><addsrcrecordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mSw9C8oyUxOzMmpVHDOz8tLTS5JTVHwTc3NL6pUSMsvUnDJLE5MTy9KTU8ESbgkliQqOKfmlaQWFfMwsKYl5hSn8kJpbgZlN9cQZw_dgqL8wtLU4pL4rPzSojygVLyRibG5kYWRoYWBMXGqAHXdNkU</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2437282180</pqid></control><display><type>article</type><title>Optically Connected Memory for Disaggregated Data Centers</title><source>ProQuest - Publicly Available Content Database</source><creator>Gonzalez, Jorge ; Gazman, Alexander ; Hattink, Maarten ; Palma, Mauricio G ; Bahadori, Meisam ; Rubio-Noriega, Ruth ; Orosa, Lois ; Glick, Madeleine ; Mutlu, Onur ; Bergman, Keren ; Azevedo, Rodolfo</creator><creatorcontrib>Gonzalez, Jorge ; Gazman, Alexander ; Hattink, Maarten ; Palma, Mauricio G ; Bahadori, Meisam ; Rubio-Noriega, Ruth ; Orosa, Lois ; Glick, Madeleine ; Mutlu, Onur ; Bergman, Keren ; Azevedo, Rodolfo</creatorcontrib><description>Recent advances in integrated photonics enable the implementation of reconfigurable, high-bandwidth, and low energy-per-bit interconnects in next-generation data centers. We propose and evaluate an Optically Connected Memory (OCM) architecture that disaggregates the main memory from the computation nodes in data centers. OCM is based on micro-ring resonators (MRRs), and it does not require any modification to the DRAM memory modules. We calculate energy consumption from real photonic devices and integrate them into a system simulator to evaluate performance. Our results show that (1) OCM is capable of interconnecting four DDR4 memory channels to a computing node using two fibers with 1.07 pJ energy-per-bit consumption and (2) OCM performs up to 5.5x faster than a disaggregated memory with 40G PCIe NIC connectors to computing nodes.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Computation ; Computer centers ; Computer simulation ; Connectors ; Data centers ; Dynamic random access memory ; Energy consumption ; Nodes ; Performance evaluation ; Photonics</subject><ispartof>arXiv.org, 2020-08</ispartof><rights>2020. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://www.proquest.com/docview/2437282180?pq-origsite=primo$$EHTML$$P50$$Gproquest$$Hfree_for_read</linktohtml><link.rule.ids>780,784,25752,37011,44589</link.rule.ids></links><search><creatorcontrib>Gonzalez, Jorge</creatorcontrib><creatorcontrib>Gazman, Alexander</creatorcontrib><creatorcontrib>Hattink, Maarten</creatorcontrib><creatorcontrib>Palma, Mauricio G</creatorcontrib><creatorcontrib>Bahadori, Meisam</creatorcontrib><creatorcontrib>Rubio-Noriega, Ruth</creatorcontrib><creatorcontrib>Orosa, Lois</creatorcontrib><creatorcontrib>Glick, Madeleine</creatorcontrib><creatorcontrib>Mutlu, Onur</creatorcontrib><creatorcontrib>Bergman, Keren</creatorcontrib><creatorcontrib>Azevedo, Rodolfo</creatorcontrib><title>Optically Connected Memory for Disaggregated Data Centers</title><title>arXiv.org</title><description>Recent advances in integrated photonics enable the implementation of reconfigurable, high-bandwidth, and low energy-per-bit interconnects in next-generation data centers. We propose and evaluate an Optically Connected Memory (OCM) architecture that disaggregates the main memory from the computation nodes in data centers. OCM is based on micro-ring resonators (MRRs), and it does not require any modification to the DRAM memory modules. We calculate energy consumption from real photonic devices and integrate them into a system simulator to evaluate performance. Our results show that (1) OCM is capable of interconnecting four DDR4 memory channels to a computing node using two fibers with 1.07 pJ energy-per-bit consumption and (2) OCM performs up to 5.5x faster than a disaggregated memory with 40G PCIe NIC connectors to computing nodes.</description><subject>Computation</subject><subject>Computer centers</subject><subject>Computer simulation</subject><subject>Connectors</subject><subject>Data centers</subject><subject>Dynamic random access memory</subject><subject>Energy consumption</subject><subject>Nodes</subject><subject>Performance evaluation</subject><subject>Photonics</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>PIMPY</sourceid><recordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mSw9C8oyUxOzMmpVHDOz8tLTS5JTVHwTc3NL6pUSMsvUnDJLE5MTy9KTU8ESbgkliQqOKfmlaQWFfMwsKYl5hSn8kJpbgZlN9cQZw_dgqL8wtLU4pL4rPzSojygVLyRibG5kYWRoYWBMXGqAHXdNkU</recordid><startdate>20200825</startdate><enddate>20200825</enddate><creator>Gonzalez, Jorge</creator><creator>Gazman, Alexander</creator><creator>Hattink, Maarten</creator><creator>Palma, Mauricio G</creator><creator>Bahadori, Meisam</creator><creator>Rubio-Noriega, Ruth</creator><creator>Orosa, Lois</creator><creator>Glick, Madeleine</creator><creator>Mutlu, Onur</creator><creator>Bergman, Keren</creator><creator>Azevedo, Rodolfo</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20200825</creationdate><title>Optically Connected Memory for Disaggregated Data Centers</title><author>Gonzalez, Jorge ; Gazman, Alexander ; Hattink, Maarten ; Palma, Mauricio G ; Bahadori, Meisam ; Rubio-Noriega, Ruth ; Orosa, Lois ; Glick, Madeleine ; Mutlu, Onur ; Bergman, Keren ; Azevedo, Rodolfo</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_24372821803</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Computation</topic><topic>Computer centers</topic><topic>Computer simulation</topic><topic>Connectors</topic><topic>Data centers</topic><topic>Dynamic random access memory</topic><topic>Energy consumption</topic><topic>Nodes</topic><topic>Performance evaluation</topic><topic>Photonics</topic><toplevel>online_resources</toplevel><creatorcontrib>Gonzalez, Jorge</creatorcontrib><creatorcontrib>Gazman, Alexander</creatorcontrib><creatorcontrib>Hattink, Maarten</creatorcontrib><creatorcontrib>Palma, Mauricio G</creatorcontrib><creatorcontrib>Bahadori, Meisam</creatorcontrib><creatorcontrib>Rubio-Noriega, Ruth</creatorcontrib><creatorcontrib>Orosa, Lois</creatorcontrib><creatorcontrib>Glick, Madeleine</creatorcontrib><creatorcontrib>Mutlu, Onur</creatorcontrib><creatorcontrib>Bergman, Keren</creatorcontrib><creatorcontrib>Azevedo, Rodolfo</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni)</collection><collection>ProQuest Central</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>ProQuest - Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Gonzalez, Jorge</au><au>Gazman, Alexander</au><au>Hattink, Maarten</au><au>Palma, Mauricio G</au><au>Bahadori, Meisam</au><au>Rubio-Noriega, Ruth</au><au>Orosa, Lois</au><au>Glick, Madeleine</au><au>Mutlu, Onur</au><au>Bergman, Keren</au><au>Azevedo, Rodolfo</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Optically Connected Memory for Disaggregated Data Centers</atitle><jtitle>arXiv.org</jtitle><date>2020-08-25</date><risdate>2020</risdate><eissn>2331-8422</eissn><abstract>Recent advances in integrated photonics enable the implementation of reconfigurable, high-bandwidth, and low energy-per-bit interconnects in next-generation data centers. We propose and evaluate an Optically Connected Memory (OCM) architecture that disaggregates the main memory from the computation nodes in data centers. OCM is based on micro-ring resonators (MRRs), and it does not require any modification to the DRAM memory modules. We calculate energy consumption from real photonic devices and integrate them into a system simulator to evaluate performance. Our results show that (1) OCM is capable of interconnecting four DDR4 memory channels to a computing node using two fibers with 1.07 pJ energy-per-bit consumption and (2) OCM performs up to 5.5x faster than a disaggregated memory with 40G PCIe NIC connectors to computing nodes.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2020-08
issn 2331-8422
language eng
recordid cdi_proquest_journals_2437282180
source ProQuest - Publicly Available Content Database
subjects Computation
Computer centers
Computer simulation
Connectors
Data centers
Dynamic random access memory
Energy consumption
Nodes
Performance evaluation
Photonics
title Optically Connected Memory for Disaggregated Data Centers
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-10T12%3A32%3A14IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Optically%20Connected%20Memory%20for%20Disaggregated%20Data%20Centers&rft.jtitle=arXiv.org&rft.au=Gonzalez,%20Jorge&rft.date=2020-08-25&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2437282180%3C/proquest%3E%3Cgrp_id%3Ecdi_FETCH-proquest_journals_24372821803%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2437282180&rft_id=info:pmid/&rfr_iscdi=true