Abstract.

In article, we are considering the problems of processing SQL queries with a high specific weight of join operations to conservative type databases (with occasional updating of data in specially allocated time) with increased data volumes on the GPU-cluster platform. Modified architecture of the previously developed parallel DBMS Clusterix-N (N - from New). We are proposed the methods for organizing dynamic segmentation of intermediate / temp relationships on a dedicated node with GPU accelerators and full load of processor cores of nodes on JOIN and IO levels. The performance of this DBMS is compared with the original DBMS PerformSys on a limited TPC-H test (without write operations) with VDB=60 GB and VDB=120 GB.

Keywords:

conservative type databases, increased volumes of data, modification of parallel DBMS architecture, dynamic segmentation of intermediate / temp relations, application of graphic accelerators, full load of processor cores, comparative performance estimates.

PP. 108-118.

DOI 10.14357/207186321804011

References

1. Matt Turck Firing on All Cylinders: The 2017 Big Data Landscape [Web page] // URL: http://mattturck.com/bigdata2017/
2. Paradigm4: Creators of SciDB a computational DBMS URL: https://www.paradigm4.com/
3. In-Memory Database | VoltDB URL: https://www.voltdb.com/
4. Postgres-XL | Open Source Scalable SQL Database Cluster URL: https://www.postgres-xl.org/
5. Abramov E.V. Parallel'naya SUBD Clusterix. Razrabotka prototipa i ego naturnoe issledovanie [Parallel DBMS Clusterix. Development of prototype and its full-scale studits] // Herald of KSTU named after A.N. Tupolev. Kazan. 2006. № 2. P.50-55.
6. Raikhlin, V.A. Informacionnye klastery kak dissipativnye sistemy // V.A. Raikhlin, D.O. Shageev // Nonlinear world. Vol.7. 2009. №5. P.323-334.
7. V.A. Rajhlin, R.K. Klassen Sravnitel'no nedorogie gibridnye tekhnologii konservativnyh SUBD bol'shih ob"emov [Relatively inexpensive hybrid technology of large volumes conservative DBMS] // Journal of Information Technologies and Computing Systems. Moscow. 2018. P. 46-59.
8. TPC BenchmarkTM H Standard Specification Revision 2.17.1, URL: http://www.tpc.org/tpc_documents_current_versions/pdf/tpc-h_v2.17.1.pdf
9. Implementation of HASH on GPU. URL: https://github.com/rozh1/gpuhash
10. Martin, J. Computer database organization. Second Edition — Prentice-Hall, Inc., Eng-lewood Cliffs, New Jersey 07632, 1977.
11. Klassen R.K. Programma regional'noj balansirovki nagruzki k baze dannyx konser-vativnogo tipa na klasternoj platforme «PerformSys». [The program for regional load balancing to a conservative type database on the cluster platform «PerformSys».] Svidetel'stvo o gosudarstvennoj registracii programmy dlya E'VM №2017611785 ot 09.02.2017. [Certificate of state registration of the computer program No. 2017611785 of 09.02.2017].
12. Klassen R.K. Povyshenie e'ffektivnosti parallel'noj SUBD konservativnogo tipa na klaster-noj platforme s mnogoyadernymi uzlami [Improving efficiency of parallel conservative type DBMS on a cluster platform with multi-core nodes] //Herald of KSTU named after A.N. Tupolev. 2015. No.1. P.112-118.
13. Vadim A. Raikhlin, Roman K. Klassen Can GPU-accelerator significantly increase the effectiveness of conservative DBMS considerable volumes on cluster platforms? //2017 International Siberian Conference on Control and Communications (SIBCON). 2017. P. 1-5. DOI: 10.1109/SIBCON.2017.7998474.
14. Raikhlin V.A., Minyazev R.Sh. Analiz processov v klasterax konservativnyx baz dannyx s pozicij samoorganizacii [Analysis of processes in clusters of conservative databases from self-organization perspective] //Herald of KSTU named after A.N. Tupolev. 2015. No.2. P.120-126.
15. PerformSys. URL: https://github.com/rozh1/PerformSys
16. Clusterix-N. URL: https://bitbucket.org/rozh/clusterixn