6th International Conference on Machinery, Materials, Environment, Biotechnology and Computer (MMEBC 2016)

Research on Database Application Performance Optimization Method

Jiangang Zhang1, a
1
Nanchang Institute of Science & Technology, Nanchang, Jiangxi, China 330108,
a
email,

Keywords: Database, Application, Performance Optimization, Method

Abstract. Optimization of database systems for the normal operation of the whole system plays a
vital role, but it is a very complex task. The final decides the database performance database
availability. A database can be up and running, but if you want to take a long time to return results,
then their use will be limited. From the user's point of view, if the actual throughput for a given
period of time in the database is high, the actual performance of the database is higher. On the other
hand, the poor performance of the database to run a large complex query seems to stop at some
point, like a database does not exist, so was anxious to let the user. Therefore it emphasizes
optimization is very important.

Introduction
Database application system currently used by the main database platform Oracle, Sybase, SQL,
Server, DB2 other large relational database, Oracle for its openness, efficiency, reliability and
confidentiality has become the first choice of many developers. However, due to the huge Oracle
architecture, and the management is very complex, so a lot of database application system after
Oracle as the database backend development for some time, there have been some problems in the
use of performance, such as running slower reaction time becomes longer and the like. Usually
developers often done at the application layer optimization process, such as the optimization of SQL
statements, but sometimes the effect is not obvious, this is because many performance optimization
must be adjusted to the DB server to complete.
Database is a database management system is more widely used, but the response time with the
information age and the rapid expansion of the continuous generation of data, database applications
have increased. How to network traffic, disk I / O and CPU time is minimized, so that the shortest
response time of each query and maximize the throughput of the entire database server, to optimize
the performance of database applications, it is particularly important. When the database
optimization requires a thorough analysis of the requirements of the application, understand the
logical and physical structure of the data, co-ordination from the viewpoint of the operating system,
application design, indexes, stored procedures, transactions, etc., form the overall scheme of the
system this paper made some research on this.
Consideration of performance issues should be run through the whole process of the
development phase. The life cycle of a database system can be divided into: the design,
development, and finished three stages. In the design and development phase of database
performance optimization into wood a minimum, the largest gains.

Database Design
Logic Design. Logical database design, including relationships between tables, which is the core
relational database optimization. A good logical database design can lay a good foundation for the
optimization of databases and applications. Comply with the rules of database design usually
comprises a large number, there is the relationship between the narrow tables instead of wide data
table with many columns, columns and more tables to produce less, thus reducing data redundancy,
thus reducing the page for storing data, to improve the efficiency and reduce the possibility of the
application due to data inconsistencies caused the error. However, table relationships may need to
merge through a complex process that will reduce system performance. Some degree of
© 2016. The authors - Published by Atlantis Press 2236
standardization of the non-performance of the system can be improved, non-standardized procedure
can be carried out using a variety of different ways depending on performance considerations, the
preferred strategy is standardized design as a starting point, and for specific reasons selective some
tables.
Physical Design. To achieve physical design optimization, we must understand and make good
use of database access format and operating characteristics of hardware resources, especially
memory and disk subsystem I / O. Associated with each table column's data type should reflect the
minimum required data storage space, especially for indexed columns even more so. For example,
be able to use type do not use integer type, so the index fields can be read faster, and more data can
be placed in rows on a data page, therefore reducing the I / O operations.

Steps of Performance Optimization

Many people believe that optimize database application system is in operation of the system when
there is a problem, which is when this view is wrong. Optimization work during the system
planning and design phase has already begun, and has been throughout the operating cycle of the
system. In the process of tuning the database requires application designers, application developers,
database administrators and system administrators together. The following order of priority given
step optimization:
Optimize the Business Logic. For best performance, the database design process to use business
logic. This is related to the overall system level analysis and design. Like whether the system-wide
adoption of such threaded server configuration issue is the need to consider the issue at this stage.
Such designers with the performance requirements of specific business requirements directly
linked;
Adjust the Data Structure Design. In the data design phase, the designer must decide what data
is required for the application, but also consider the relationship between what is important and
what are their properties, but also to determine which attributes can be used as the primary key
(Primary Key) or foreign key ( Foreign Key). Finally, to construct a good information to better meet
performance goals. This phase requires a careful analysis of the data, to avoid data redundancy,
taking into account a reasonable distribution of data in order to prevent congestion data access, such
as whether to use the database partitioning feature, and for frequently accessed database tables need
to establish whether local or global indexes;
Adjust the Application Structure. This step needs to consider the application uses what
architecture is the use of the traditional Client / Serve: two-tier architecture, or use the three-tier
architecture Browser / Web / Database of. Database resources of different application architecture
requirements are different;
Optimize the Logical Structure of the Database This stage is through increase, decrease or
adjust the logical structure, such as consideration other than indexing on the primary and foreign
keys to add other indexes to improve system performance;
Operation
Database Optimization. This phase before optimize the database server, first make sure that the
application has been fully utilized in order to improve the performance and design of SQL and some
special features. As Oracle database array processing, Oracle optimizer, row-level locking
management and PL / SQL and so on. Execution of the application will ultimately boil down to the
database SQL statement is executed, and therefore the efficiency of SQL statements the final
decision of the properties of the database performance analysis and optimization of database
application system. Whether it is writing a new adjustment or had some performance issues with
SQL statements should use less CPU or disk UO and other resources as the main objective;
The Access Path Optimization. In order to ensure efficient data access path, consider using
cluster (clusters), B * tree, bitmap index, and optimizer hints. Taking into account the table and
histogram analysis to analyze each column, in order to help the optimizer determine the best query
plan;
Optimize Memory Allocation. Memory allocation is optimized configuration information in the
2237
system is running. In Oracle, for example, a database administrator can be adjusted according to the
database Health System Global Area (SGA area) data buffer, the log buffer size and shared pool;
you can also adjust the program global area (PGA area) in size. Note that, SGA zone is not the
bigger the better, SGA region over the General Assembly takes up memory the operating system
caused the virtual memory page swapping, so it will reduce the system. A reasonable allocation of
memory resources can improve cache performance, reduce SQL statement parsing time, and reduce
paging and paging;
Optimize the Physical Structure of the Database and FO. Database data is ultimately stored
in the physical disk, access the data is read and write to these physical disk, so the physical storage
of these optimization is an important part of system optimization. Although the physical storage
optimization is not possible to reduce the number of physical storage read and write, but it can make
them read and write as much as possible in parallel, to reduce disk reading and writing competition,
thereby enhancing the efficiency of physical storage can also be precisely calculated to reduce
unnecessary expand the physical storage structure, to improve system performance;
Adjust the Underlying Platform. Based on the underlying operating system, it adjusts to make
the database more efficient operation. For example: UNIX operating system running on an Oracle
database, you can adjust the size of the UNIX data buffer pool, each process can use the memory
size and other parameters;
Adjust Network Performance. Large databases running inseparable is from the network, so the
network configuration is good or bad has a direct impact on the efficiency of the database
application systems. For example, the same application is executed on a client faster, and in another
client but slow running, this indicates a problem on the network, the network needs to be adjusted.
In fact, the linkages between the database optimization step, influence each other. Sometimes with
several steps were sometimes done in the back of the adjustments need to return to the previous step
and then adjust, and so forth constantly, until achieve the desired result.

Optimization of Database Application System

Optimize the Database Server. Fully understand the size of the computer's memory and main
memory database usage and understanding of the working principle of the main memory will help
tune the database. Here's to the Oracle database to illustrate the optimization of the performance of
the memory. In Oracle, tuning memory usually refers to the adjustment of the Shared Global Area
(SGA). This includes monitoring and adjusting the shared pool (library cache and data dictionary)
and database cache. Adjust and adjust the application memory and input / output closely linked,
since a major objective of a reduction or elimination of adjustment contention in all aspects of the
adjustment process involves.
Optimize SQL Statements. And application design, as adjusted for the SQL statement seems to
have nothing to do with the responsibility of database management. However, database
administrators should check the SQL involved in a part of the application is written. A
well-designed application, SQL structure is not ideal if you are using will still experience
performance issues. In a properly designed database application design and SQL performance issues
cause most of the problems. In a relational database, application design, not as good as the physical
location of data is important to its logical position. However, the database must find the data, so that
it can return the user to perform the query. Therefore, the key is to make adjustments SQL database
to find the most simplified data path.
The optimization process is an internal mechanism used by Oracle, it can calculate the most
efficient access paths to obtain the data required to satisfy SQL statements. Oracle provides two
internal optimizers: rule-based optimizer and cost-based optimizer.
Optimize Operating System Platforms. The database server is heavily dependent on the
operating system to run the server, if the operating system cannot provide the best performance,
then the adjustment in any case, the database server cannot play its proper performance. Here
mainly to the Windows NT platform, for example, instructions to run on the NT platform like
Oracle to deal with large database operating system to do what changes.
2238
 Optimize Network Performance. Running large databases inseparable from the network, the
network performance will directly affect the performance of the database. The main adjustment is to
minimize network Internet data traffic on the network in order to reduce resource contention. Here's
to oracle database as an example, explain how to adjust network performance. The Oracle SQL *
Net is a network protocol or Nets, it is located in the session layer Open Systems Interconnection
seven layer structure, which provides transparent connectivity to Oracle servers and clients. The
so-called transparent connectivity means that SQL * Net / Net8 received from the Oracle
application SQL statements via the interface layer and these statements packaged and sent to the
database in a standard format. Oracle client application using SQL * Net / Net8 to communicate
with the Oracle server, and SQL * Net / Net8 is responsible for establishing and maintaining (final
safety shut-off) to open the client (or server) session between the server and the network
connection.

Conclusion
Database optimization is a complex, involving very broad question, there is no fixed pattern to
solve the need to analyze specific issues. Given above is only a general principle of optimization, in
front of the main article in the Oracle database, for example, from the database server, SQL
statements, running the operating system and network performance and other aspects of the Oracle
database application system may encounter various performance issues provides a general solution
under different circumstances, we hope to provide a database application system the general idea of
optimization methods.

References
[1] Guo Weihua. Computer Engineering and Applications, Vol. 6 (2004) No 53, p.25-26
[2] Lu Chenyin, Wang Qunyong. Computer Engineering, Vol. 12 (2005) No 27, p.74-76
[3] Jing Jianfen, Hou XuSiem. Computer Engineering, Vol. 30 (2004) No 19, p.144-145
[4] Wang Kuailiang. Mining machinery, Vol. 29 (2008) No 27, p.21-23
[5] Zhang Gongxu, Sun Jing. New Quality Management, Vol. 8 (2003) No 27, p.57-60

2239