Frontend computing paradigm toward accelerating large-scale distributed computing platform

2015 Fiscal Year Final Research Report

• Project/Area Number: 26540040
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Multimedia database
• Research Institution: Hokkaido University
• Principal Investigator: Motomura Masato 北海道大学, 情報科学研究科, 教授 (90574286)
• Co-Investigator(Kenkyū-buntansha): Asai Tetsuya 北海道大学, 情報科学研究科, 準教授 (00312380)
• Project Period (FY): 2014-04-01 – 2016-03-31
• Keywords: データマイニング / ビッグデータ / ストリーム処理

Outline of Final Research Achievements

As an example for accelerating large-scale distributed processing platform based on front-end computing concept, we have worked on acceleration of memcached on network interface, and its hardware implementation. We have found 3.5 times speed improvement compared to existing software solution. We have also worked on approximate nearest-neighbor search algorithm LSH, especially Exact Euclidean Locality Hashing. By improving the algorithm and by implementing it on hardware, we have found 1/9 latency and around 25 times throughput improvement over existing software solution.

Free Research Field

離婚フィギュラブルシステム