Presentation 2017-09-19
A load balancing mechanism using shared state for tree processing in actor model
Kouhei Sakurai,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) For machine learning and data mining, there are methods that deal with tree structure, and also their parallelization and distribution are required against big-data. The actor model is a practical choice for the purpose; we can define actors as nodes in the tree, and then it can achieve scaling out by those actors concurrently processing input data. However, since the root of the tree always takes the all data as the entrance of it, the root becomes a serious performance bottleneck. In this work, we extend a library of actor model and introduce replicas for the root node-actor which share the state of the node in order to realize load balancing.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) data stream mining / decision trees / incremental learning / actor model / shared state / data replication
Paper # KBSE2017-21
Date of Issue 2017-09-12 (KBSE)

Conference Information
Committee KBSE
Conference Date 2017/9/19(1days)
Place (in Japanese) (See Japanese page)
Place (in English) Doshisha Univ. Tokyo Branch Office
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Shigeo Kaneda(Doshisha Univ.)
Vice Chair Fumihiro Kumeno(Nippon Inst. of Tech.)
Secretary Fumihiro Kumeno(Kanagawa Inst. of Tech.)
Assistant Takuya Saruwatari(NTT DATA) / Kosaku Kimura(Fujitsu labs.)

Paper Information
Registration To Technical Committee on Knowledge-Based Software Engineering
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A load balancing mechanism using shared state for tree processing in actor model
Sub Title (in English)
Keyword(1) data stream mining
Keyword(2) decision trees
Keyword(3) incremental learning
Keyword(4) actor model
Keyword(5) shared state
Keyword(6) data replication
1st Author's Name Kouhei Sakurai
1st Author's Affiliation Kanazawa University(Kanazawa Univ.)
Date 2017-09-19
Paper # KBSE2017-21
Volume (vol) vol.117
Number (no) KBSE-214
Page pp.pp.1-6(KBSE),
#Pages 6
Date of Issue 2017-09-12 (KBSE)