Evaluation time of multiquery processing

Аuthors

Brekhov O. M.^1*, Wunna K. K.²

1. Moscow Aviation Institute (National Research University), 4, Volokolamskoe shosse, Moscow, А-80, GSP-3, 125993, Russia
2. Defence services academy, Pyin OO Lwin, Republic of the Union of Myanmar

*e-mail: obrekhov@mail.ru

Abstract

Reducing the total time of multi-query queries processing can be achieved through co-processing of a subset of elementary queries that form an intersection of queries with multi-query. It is well known [1, 2] that the order of elementary queries processing is a significant parameter of query processing time. Here, in elaboration of our work [3] the task of forming a multi query execution plan based on the order of elementary queries processing in multiprocessor database aerospace systems is discussed.
We prove the two statements.
Statement 1. Let multi query consists of n queries which form ordered sets with successive numbers of elementary queries.
Let the conditions met:
Embedding of sets:.
Processing elementary queries from subsets   in order provides a reduction in multi-query queries multijob time.
Statement 2. Let the intersection of elementary queries sets be a subset . If a subset s is formed of elementary queries with successive numbers, starting with the first elementary query EQ₁, the execution of elementary queries of subset s first off provides a multijob time reduction compared to inconsistent query execution by the time ,    where — processing time of elementary queries of subset s.
We also obtained analytical evaluation of multi-query processing time for the two patterns (geometric or arithmetic progressions) of elementary queries parameter setting in multiprocessor database when queries are defined by the following parameters:
k — the number of elementary queries that form queries of multi queries Elementary queries from d groups of u elementary queries each. Each query , consists of two groups of elementary queries with numbers: 1st group:       2nd group 
Analytical experiments have shown that

• success probability parameter during query running isasignificant parameter affecting the selection ofeither multijob and inconsistent methods ofmulti-query processing ordetermination the number ofprocessors;
• multijob query processing provides time inuse which isnot always minimal inrelation toinconsistent processing;
• increasing the number ofprocessors may lead not toreduction but toincrease ofmulti-query processing time.

Keywords:

database, multiprocessor computer, multi-query, distribution of elementary queries, optimization

References

1. Paura S. M. Tsai, Arbee L. P. Chen. Optimizing Queries with Foreign Function in a Distributed Environment", IEEE Trans. On Knowledge and data engineering, vol.14,No.4,July/August 2002.
2. Brekhov O.M. Uspekhi sovremennoi radioelektroniki, 2012, vol. 12, no. 7, pp. 37-45.
3. Brekhov O.M., Vunna Dzho Dzho, Tan Khlaing M’int Naukoemkie tekhnologii, 2014, no. 1, pp. 101-106.

Download