Classification and A Brief Analysis of Existing Developments For The Search Optimization in Databases

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 2

CLASSIFICATION AND A BRIEF ANALYSIS OF EXISTING

DEVELOPMENTS FOR THE SEARCH OPTIMIZATION IN


DATABASES

Nowadays there is a development and dissemination of "cloud


computing" technology. The growing demand for services offering a broad
range of cloud computing services for large numbers of users all over the
world, therefore increase in the number of applications, the purpose is to
process large data sets. The operation of the database in the cloud leads to
the need to find new instruments.
Most often this concept is treated as a "complex of information
collection and procedures: management, updates, information retrieval and
post-processing - which allows to accumulate, store, update and provide
information", processing and organizing information, should take into
account. User no longer divides his activities on information search object
can not be clearly defined in advance [1].
The search tools and technologies used for the implementation of
information requirements, depend on the type and condition of the problem
to be solved by the user operations. The process user interaction with the
system is determined by the level of knowledge of user resource content
(completeness, reliability of the source, etc.) and the functionality of the
system as a tool. In general, these factors are usually limited to the notion
of professionalism - information (trained / untrained user) and objective
(professional / amateur). For solving the problem of query optimization in
the cloud storage system should also be taken into account. For network
topology used two methods:
1) Query Optimization in the Cloud SQL type database architecture are as
follows: all files are stored in the local file system; cloud database is
designed to store and manage huge amounts of index files and metadata;
enter the query and get the results performed by the web user interface;
user request is executed the current search query plan (as a subset).
2) Query optimization in cloud databases NoSQL type since there is one
sever: used programming model map-reduce (MR) platform for cloud
computing, which allows analysis of large amounts of data in the cloud.
MR facilitates parallel execution for long-term problems of the analysis of
large data. In MR each task is represented as a map for reduce tasks. MR
Kernel for storage and data uses a distributed file system with MRQL
language. It allows user to write own scripts for a large tasks in declarative
form, and at the same time itself to optimization. This procedure consists of
the following steps: simplification of the request, building a query graph
and query graph representation in algebraic form [2].
In this work another method is proposed for similar purposes, based
on virtual machines and processing centers. The main interest of this
approach lies in the fact that its purpose is to balance the load of virtual
machines in the cloud environment which indirectly leads to an increase in
the quality of the search (it is obvious that the speed with which the cloud
responds to user requests, is one of the search quality criteria). For resource
management in large-scale data centers are developed and implemented a
centralized solution, but in this case the occurrence of a failure at the
control node, resulting in malfunction of the whole system. The average
length of the vector unit load is equal to the number of iterations of sending
the index. A load information will be stored in a decentralized, in order to
avoid trouble in the event of a failure of the node, another positive aspect is
that the network traffic is distributed across all active nodes (as opposed to
the scheme with centralized management, where all packets should go
through common node). The decision on the virtual machine migration can
be taken in two cases: when the CPU usage exceeds a certain level (the
upper limit), freeze mode or when the CPU usage is below a certain level
(the lower limit) transferred to the "sleep" mode.
Each methods have both advantages and disadvantages. Common to
all is a lack of synthetic nature of the results, that the introduction of
statistics obtained in artificial systems created just for testing approach.
Optimization such developments in the near future will be in demand.

REFERENCES
1. Vakkari P., Hakala N. An Appropriate Boolean Query
Reformulation Interface for Information Retrieval Based on Adaptive
Generalization. In WIRI, 2015. p. 145-150.
2. Ullman J.D., Hector G.M., Widom J. Database Systems. USA:
Pearson, 2013. 256 p.

http://www.techrepublic.com/blog/10-things/10-things-you-should-
know-about-nosql-databases/
http://www.hadoop360.com/blog/advantages-and-disadvantages-of-
nosql-databases-what-you-should-k
https://www.thoughtworks.com/insights/blog/nosql-databases-
overview

You might also like