Journal of Software Engineering and Applications, 2013, 6, 33-36
doi:10.4236/jsea.2013.63b008 Published Online March 2013 (http://www.scirp.org/journal/jsea)
33
Selection and Research for Online Registration System’s
Database System
Yabing Chen1, Haiyong You2
1Department of mechanical and electrical engineering, Yantai Engineering & Technology College, Yantai, China; 2Electrified educa-
tion centre, Yantai Engineering & Technology College, Yantai, China.
Received 2013
ABSTRACT
Based on the Online Registration System (ORS) characteristics and key technology analysis, this paper points out that
that a good performance and high stability of the ORS lies in the choice of the system database. Database clustering
technology which has merits such as concurrent processing, easy expansion, and high security is proposed to achieve
database subsystem of ORS, and the design of the database cluster system framework is available in this paper. Finally,
we also explore the database load balancing of the cluster system, heterogeneous database replication technology.
Keywords: Online Registration System; Database System; Database Cluster System; Load Balancing
1. Introduction
With the rapid development of modern information
technology, the internet characterized with convenient,
fast, as well as rich information becomes ubiquitous in
most cities and even countryside. Furthermore it is also
run widely in various educational resources including the
common entrance examination, examination of th e adult,
independent recruitment examination institutions at all
levels, even distance education as well. As the most fo-
cused interface of the network education, the online reg-
istration technology has been developed rapidly.
Project of Online Registration System (ORS) research
and development was audited by the Shandong Provin-
cial Planning Leading Group of the Education & Science
in May 2011. The choice of database system, one of top-
ics in the project is fully researched in this paper.
2. The Characteristics and Key Technologies
of ORS
In my studies the main framework adopts three B/S lay-
ers which are browsers and negligible size of main-
tainance base installed by clients, the Web server kept in
the middle layer to probe the client's requests, as well the
underlyi ng database which is visit e d by the web server.
Unique features of online registration technology
mainly focus on below points.
1) The data access period converged relatively. In
usual, simple work just to do the routine maintenance
because of seldom accesses by user. Nevertheless during
the examination stage, amount of database access will
surge abruptly, not only by administrators to frequently
maintenance but by users to access.
2) The lethal factor of data security. Once the info rma-
tion leakage occurs, it is not only the personal informa-
tion leakage problems, but even affect the entire test
process. The risk of large-scale examination security is
increased proportionally.
3) Data types which are mainly text data and a few of
date or figures type that supports candidates photos are
relatively simple. However, the surged candidates take
vast volume data for their private information storage.
In the ORS technology, data processing is consisted
with several complicated working processes. One of
them, the data response speed is the key technical point.
Therefore, capability of soon data response depends on
the high reliability and fast response ability of database
in the whole project.
In light of the above analysis, the ideal state for the
ORS technology should meet the flexible requirement
that the usage of the database system should occupy less
hardware resources as possible in usual while be able to
facilitate expansion, integrated the necessary hardware
and software resources to meet the spurt of the great ac-
cess demand. Ideally, the structure of ORS is shown in
Figure 1.
3. Technical Requirements Towards the
Database by ORS
To meet the requirements of surge accesses and tight
security towards users signed in information, the data-
base of the ORS should adopt the following strategies:

Selection and Research for Online Registration System’s Database System
34
Figure 1. The favorable structure of ORS.
1) The huge data storage, processing, processing ca-
pacity
2) The ideal response time for user requests
3) The stability and robustness towards physical data-
base runni ng
4) The facility of the database itself
5) The security of the data confidentiality.
For most stylish database system, desirable services
are available towards above stability an d robu stness. And
most of databases have not addressed the security audit
data analysis tools, in order to improve and perfect the
current databases security audit mechanism, we can use
data mining technology, data mining technology to the
database security audit, the characteristics of database
records for analysis.
But on the view of strategies huge data processing and
response time for user requests, efforts to improve possi-
bility of efficient database system usage by users are
never suspended as well. In fact, the strategies huge data
processing and response time for user requests require
not only a database management system (DBMS) single
effort but perplexing factors including a database system
hardware configuration, Web connected and so on. Con-
sistent with the needs of huge data processing and re-
sponse time for user requests, it is conventiona l argued to
improve the capability of database server and storage
volume, adjust the DBMS parameters and other means
while database system selection and preparation. How-
ever the factual effect is controversial because of the
wasted cost of mainframe and limited improvement of
overall performance of the database system. Paralleled or
distributed database systems are evolutionary methods.
Paralleled database systems with high availability and
scalability focus on parallel data processing capability [1]
and emphasize one single database system that each of its
node does not have autonomy so much so flexibility is
constrained [2]. Distributed database systems emphasize
the centralized control for the system. It is characterized
with the distribution of data and database coordination,
as well it aims to realize the autonomous site and data
sharing globally and transparently without requiring the
usage of every node in the network to improve system
performance [3]. Therefore, we prefer to use a database
cluster system to meet the online application system for
database system requirements.
4. The Choice of Database Cluster System
As far as the single database system selection, it is key to
choose the proper DBMS which is selective from rich
and varied candidates on current IT market. Mainstream
DBMS products include DB2, Oracle, Microsoft SQL
Server, Sybase SQLServer, Informix and MySQL which
have their respective features and advantages. However,
to the database designers, it seems obsolete DBMS
products on project schedule until the operation funds
and other preparations are available sometimes no matter
how latest he chooses. Theoretically speaking, it is valid
to use database cluster system to effectively alleviate the
suffering mentioned and deal with more and more com-
plex demanding for database system nowadays. Database
cluster system refers a group of database server as one
whole to manage through the network so that they can
work coordinately to supply services for users, wh ich can
be heterogeneous. The cluster system has few limitation
for the server operating system and all database servers
are mutually independent to perform a given user request
by itself, at the same time it is available to replicate data
for backup, r eco ver data and o ther fu n ction s in th e group.
When there are concurrent requests arrive, it can be in-
visibly for user to choose one certain database server
node to process requests [4-5]. On the basis of the char-
acteristics of ORS, one framework is proposed to use a
database cluster system as shown in Figure 2 which can
Figure 2. database cluster system framework.

Selection and Research for Online Registration System’s Database System
35
conceive the volume of access in ORS to adjust on data-
base cluster. It is negligible effect on other databases
when a single database is adjusted, as well the storage
area can be used as public data center.
For the database cluster system, features of reliability,
high availability and high extensibility are quite suitable
for ORS. Meanwhile there are some issues such as the
load balancing, heterogeneous database replication,
transparent access and database cluster technology which
should be fixed in the practical application. The weighted
least connection method is adopted in the processing of
load balance which administrator assigns corresponding
weight value for each node based on the database server
nodes (hereinafter referred to as the node) performance
because the value is representative of its processing per-
formance and default is 1. So when a connection request
arrives, the higher priority will assign to the nodes ac-
cording to their weight value. That is to say, higher valu e
node will get more links. After a while, it is possible that
the link number of nodes assigned to becomes propor-
tional to nodes weigh t valu e [6-7]. Based on this th ought,
the database administrator can consider total quantity
value node according to the results by formula (1).
SN=W×B (1)
In the formula, SN means the maximum number of
links that a node will receive. W means weight values,
and B means average minimum number of links by each
node. The W is valued by the hardware and software
performance of database server node, while B is com-
posed of number of database cluster head nodes and in-
stant total visit volume. When the SN is beyond or close
to the node DBMS tolerance v alue, it means cluster node
number is too small and need to increase the database
server nod es, whereas th e SN is too low, means database
node surplus and it is better to reduce the node properly.
On the side of heterogeneous database replication,
most major database vendors provide exuberant schemes,
such as Oracle converter, Sybase LTM solution, IBM
CCD solutions, as well as the Microsoft Publisher / sub-
scriber scheme [8]. One unified data interface is pro-
posed in paper [9] for heterogeneous databases. It is also
one kind of anticipated scheme to solve the data replica-
tion problem through the JDBC and XML.
It works on middleware form to fix the issue of access
the database cluster transparently [10-11]. Middleware is
the best option for the communication protocol of data-
base cluster standardization and modularization. Mid-
dleware also can provide mechanism to solve heteroge-
neous problems among database cluster system.
5. Summary
In this paper, the analysis contents focus on the ORS
basic structure and work characteristics, and then some
research on the database system proceeded. According to
the study, it is lethal factor of the whole application sys-
tem under the circumstance of surge access during the
register period. Because it is hard to meet needs of the
performance of the database with conventional methods
including purely raising the performance of database
hardware, modifying the DBMS parameters and using
connection pool technology, this paper puts forward one
database cluster system with favorable concurrency,
flexible expansion and high-security based on the
framework of database system. What’s more, researches
also proceed in designing of the database cluster system
framework, and issues of load balancing, data replica-
tion.
REFERENCES
[1] R.Buyya, T.Cortes, H.Jin. Single System Image (SSI).
Journal of High Performance Computing Applications,
2001,15(2):124~135
[2] Gong Weihua. Study on the key issues of database cluster
system[D]. Huazhong University of Science and Tech-
nology，2006.
[3] Li Chongxin. Design and implementation snapshot for
distributed database HBase[D]. Zhejiang Un1versity,
2011.
[4] Murthy, V.K. High performance cluster computing using
component-oriented distributed systems[C].1st Interna-
tional Conference on e-Science and Grid Computing，
Melbourne， Australia: First International Conference
on e-Science and Grid Computing，e-Science 2005，
2005:522~529.
[5] Z Shao, H.Jin.HARTS. High availability cluster architec-
ture with redundant TCP stacks[C].IEEE Initernational
Performanee, Computing and Communications Confer-
ence，Phoenix，AZ, United states: Institute of Electrical
and Electronics Engineers Inc., 2003:28~29.
[6] Keong Loh Peter Kok, Hsu Wen Jing, Wentong Cai. How
network topology affects dynamic load balancing[J].
IEEE Parallel and Distributed Technology, 1996,
4(3):25~35.
[7] Zhu Xiongfeng. Research and analysis on database cluster
middleware MySQL proxy[D]. Wuhan University, 2011.
[8] Gai J Y，Zhang Z N，Xiao H． Data replication technique
analysis and application in distributed database system [J].
Computer Applications and Software，2005，22( 7) :
36-38，41( in Chinese) ．
[9] Zhang D P，Chen C，Xu Z． Research and implemen-
tation of heterogeneous database replication technology
[J]. Journal of Graduate University of Chinese Academy
of Sicences，2012，29( 1) : 101-108．
[10] C.L.Pape, S.Gancarski, P.Valduriez. Refresco: Improving
query performance through freshness control in a data-
base cluster. in: Proceeding of On the Move to Meaning-
ful Internet Systems 2004:CoopIS, DOA and ODBASE,
Agia Napa Cyprus.Berlin: Springer-Verlag, 2004, (1):

Selection and Research for Online Registration System’s Database System
36
174~193
[11] S.Narayanan, U.Catalyurek, T.Kurc et al. Applying data-
base support for large scale data driven science in distrib-
uted environments. in: Proceedings of the Fourth Interna-
tional Workshop on Grid Computing.Los Alamitos:IEEE
Compurter Society,2003.141~148