1987 Fiscal Year Final Research Report Summary
A Testbed System for High Performance Database System Development
Project/Area Number |
60850066
|
Research Category |
Grant-in-Aid for Developmental Scientific Research
|
Allocation Type | Single-year Grants |
Research Field |
計算機工学
|
Research Institution | Kyushu University |
Principal Investigator |
YAHIKO KAMBAYASHI Kyushu University, Proderror, 工学部, 教授 (00026311)
|
Co-Investigator(Kenkyū-buntansha) |
最所 圭三 九州大学, 工学部, 助手 (50170486)
田中 克己 神戸大学, 工学部, 助教授 (00127375)
HIROSHI IMAI Kyushu University, Assouate Professor, 工学部, 助教授 (80183010)
KEIZO SAISHO Kyushu University, Research Associate
KATSUMI TANAKA Kobe Unirivesity, Assouate Professor
|
Project Period (FY) |
1985 – 1987
|
Keywords | Concurrency Control / Query processing / Rollback / Simulation / Testbed / Large Main Memory / 動的負荷 |
Research Abstract |
It is extremely improtant to develop high performance database systems. Usually, the literaturesd investigate concurrenty control, Query processing or some other techniques separately. However, wr need to evaluate the combined effect to develop a high performance system. To this end, the testbed system, Hich is implemented on a 32-bit workstation by C programmins language, is devised. We also use it to evaluate several performance improvement techniques we already proposed; they are listed below: (1)Observable of controllable concurrentyc control strategies, which are important to connect database system using different concurrency control algorithms; (2)Concurrency control utilizing relationships among data; (3) A teo phase locking algorithm which aboids deadlocks caused by read-write conflicts; (4)A timestamp-besed algorithm which uses generalized timestamp and can select rollback bojects; (5)A scheme which reduces the number of data transfer between main memoty and secondary satorages. This scheme applies to large main memory systems; (6)Dynamic concurrency control; These techniques habe several advantages over the conventional ones.
|