Presentation 2015-08-05
Improving Dynamic Node Joining Performance in KVS by Utilizing Page Cache
Shohei Miyokawa, Taiki Tokuda, Saneyasu Yamaguchi,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The number of accesses to a service in the Internet remarkably changes every hour. Thus, dynamic scaling of a system, such as expanding a system in daytime for processing large number of accesses and reducing the system in midnight for saving power consumption, is expected. KVS (Key-Value Store) is widely used DBMS in large scale Internet services. In this paper, we focus on Cassandra, a famous open source KVS, and discuss methods for improving dynamic scaling performance. Firstly, we evaluate node joining time, which is time to add a node to a running KVS system, and show that its bottleneck process is disk I/O in the existing nodes. Secondly, we analyze disk accesses in the bottleneck nodes and point it out that some heavily accessed files are cause of performance decline. Thirdly, we propose two methods for decreasing node joining time. Lastly, we evaluate our methods and demonstrate that our method can improve scaling-up performance of Cassandra.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) KVS / Cassandra / Database / Page cache
Paper # DE2015-13
Date of Issue 2015-07-29 (DE)

Conference Information
Committee DE / IPSJ-DBS / IPSJ-IFAT
Conference Date 2015/8/5(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Todaiji Culture Center
Topics (in Japanese) (See Japanese page)
Topics (in English) Management, Information Retrieval, Knowledge Discovery, etc.
Chair Masato Oguchi(Ochanomizu Univ.)
Vice Chair Makoto Onizuka(Osaka Univ.) / Masashi Toyoda(Univ. of Tokyo)
Secretary Makoto Onizuka(Univ. of Electro-Comm.) / Masashi Toyoda(Kyushu Univ.)
Assistant Mayuki Ueda(Univ. of Marketing and Distrbution Science) / Daisuke Kitayama(Kogakuin Univ.)

Paper Information
Registration To Technical Committee on Data Engineering / Special Interest Group on Database System / Special Interest Group on Information Fundamentals and Access Technologies
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Improving Dynamic Node Joining Performance in KVS by Utilizing Page Cache
Sub Title (in English)
Keyword(1) KVS
Keyword(2) Cassandra
Keyword(3) Database
Keyword(4) Page cache
1st Author's Name Shohei Miyokawa
1st Author's Affiliation Kogakuin University(Kogakuin Univ.)
2nd Author's Name Taiki Tokuda
2nd Author's Affiliation Kogakuin University(Kogakuin Univ.)
3rd Author's Name Saneyasu Yamaguchi
3rd Author's Affiliation Kogakuin University(Kogakuin Univ.)
Date 2015-08-05
Paper # DE2015-13
Volume (vol) vol.115
Number (no) DE-177
Page pp.pp.7-12(DE),
#Pages 6
Date of Issue 2015-07-29 (DE)