Cited time in webofscience Cited time in scopus

Full metadata record

DC Field Value Language Kim, Minsub - Lee, Sungjin - 2021-01-29T07:23:23Z - 2021-01-29T07:23:23Z - 2020-10-15 - 2020-08-25 -
dc.identifier.isbn 9781450380690 -
dc.identifier.uri -
dc.description.abstract Most recommender systems are designed to comply with service level agreement (SLA) because prompt response to users' requests is the most important factor that decides the quality of service. Existing recommender systems, however, seriously suffer from long tail latency when the embedding tables cannot be entirely loaded in the main memory. In this paper, we propose a new SSD architecture called EMB-SSD, which mitigates the tail latency problem of recommender systems by leveraging in-storage processing. By offloading the data-intensive parts of the recommendation algorithm into an SSD, EMB-SSD not only reduces the data traffic between the host and the SSD, but also lowers software overheads caused by deep I/O stacks. Results show that EMB-SSD exhibits 47% and 25% shorter 99th percentile latency and average latency, respectively, over existing systems. © 2020 ACM. -
dc.language English -
dc.publisher Association for Computing Machinery -
dc.title Reducing Tail Latency of DNN-based Recommender Systems using In-storage Processing -
dc.type Conference Paper -
dc.identifier.doi 10.1145/3409963.3410501 -
dc.identifier.scopusid 2-s2.0-85092208085 -
dc.identifier.bibliographicCitation 11th ACM SIGOPS Asia-Pacific Workshop on Systems, APSys 2020, pp.90 - 97 -
dc.citation.conferencePlace JA -
dc.citation.conferencePlace Tsukuba -
dc.citation.endPage 97 -
dc.citation.startPage 90 -
dc.citation.title 11th ACM SIGOPS Asia-Pacific Workshop on Systems, APSys 2020 -
Files in This Item:

There are no files associated with this item.

Appears in Collections:
Department of Electrical Engineering and Computer Science Data-Intensive Computing Systems Laboratory 2. Conference Papers


  • twitter
  • facebook
  • mendeley

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.