Sei sulla pagina 1di 6

JOURNAL OF COMPUTING, VOLUME 2, ISSUE 12, DECEMBER 2010, ISSN 2151-9617

HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 88

Mobile Database Transaction Using Mobile


Agents
A.Priya, Dr. R.Dhanapal, T.P.Vijayalakshmi

Abstract—Mobile agents are programs that can move through a network under their own control, mi- grating from host to host and inte-
racting with other agents and resources on each. In mobile database systems, new features such as mobility, disconnection, heterogeneity,
self-command and distribution make traditional transaction processing schemes such as query processing technology, transaction manage-
ment technology and data distribution technology no longer suited. To solve this problem, a new mobile database model based on agent was
proposed. And transaction processing process based on this model was set forth. The model can shield the heterogeneous among all mem-
bers, and made them transparent so as to bring great conveniences to the entering of heterogeneous members. The introduction of mobile
agent contributes to decomposition of long-lived transaction, as to complete query optimization of mobile database.

Index Terms— mobile agent, mobile database, multi agent system, transaction processing.

——————————  ——————————

1 INTRODUCTION

M
obile computers have become increasingly pre- anytime and anywhere [1]. In the mobile computing envi-
valent as professionals discover the benefits of ronment, each mobile computing device has its own da-
having their electronic work available at all tabase system. All the database systems in the mobile
times. Developing distributed applications that make effec- computing environment created a mobile database cluster
tive use of networked resources from a mobile platform, (called “MDBC” for short). The traditional distributed
however, is difficult for several reasons. First, mobile com- database technology has been difficult to support mobile
puters do not have a permanent connection into the net- computing, because of the mobility, frequent disconnec-
work and are often disconnected for long periods of time. tion, network diversity, non-symmetry of network com-
Second, when the computer is connected, the connection munications in mobile computing environment, and the
often has low bandwidth and high latency and is prone to mobility, self-command, distribution, heterogeneity and
sudden failure, such as when a physical obstruction blocks other features of the mobile database in the MDBC [2].
the signal from a cellular modem. In order to effectively access the database in mobile
Third, since the computer may be forced to use differ- computing environment, mobile database technology
ent transmission channels depending on its physical loca- came into being. Mobile database technology as an exten-
tion, the performance of its network connection can vary sion of the distributed database technology has many
dramatically from one session to another. Finally, de- strengths of distributed database and unique features [3].
pending on the nature of the transmission channel, the Mobile database system not only can manage datum and
computer might be assigned a different network address information of the local terminals, but also can connect
each time that it connects. In short, any distributed appli- the central database to process all kinds of transactions, so
cation that works on a mobile platform must deal with as to meet the needs of information processing of differ-
unforgiving network conditions. ent industries and users in the mobile environment and
With the rapid development of internet and wireless greatly improve the timeliness of information.
network technology the computer-centered mobile com- In mobile database, the transactions launched by the
puting technology has been widely used and developed, mobile computer go by the name of mobile transaction.
so that people can access to any required information at Mobile transaction usually belongs to long-lived transac-
tion because of limited communication bandwidth and
————————————————
frequent disconnection. In the processing of mobile
 Asstistant Professor, Department of Computer Applications, Ve- transactions, the location change of mobile computers
lammal Engineering College, Chennai – 600 066, Tamil Nadu, India.
Research & Development Centre, Bharathiar University, Coimbatore. will bring about complex switchover problem of region.
In addition, the processing of mobile transaction is more
 Professor, Department of Computer Applications, Easwari Engineer- prone to error [4, 5]. Mobile transaction processing has
ing College, Bharathi Salai, Ramapuram, Chennai – 600 089, Tamil Nadu,
India.
become a challenging study field because of the above
 SAP Data Archiving Consultant, Wipro Technologies, Client:Apple characteristics of mobile transaction. In order to improve
Inc, Cupertino, California, U.S. the processing efficiency of mobile database system, a
new mobile database model based on agent was pro-
posed. And transaction processing process based on this

© 2010 Journal of Computing Press, NY, USA, ISSN 2151-9617


http://sites.google.com/site/journalofcomputing/
JOURNAL OF COMPUTING, VOLUME 2, ISSUE 12, DECEMBER 2010, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 89

model was set forth. The model can shield the heteroge- attention.
neous among all members, and made them transparent so
as to bring great conveniences to the entering of hetero-
3 ARCHITECTURE OF MDB MODEL BASED
geneous members. The introduction of mobile agent con-
tributes to decomposition of long-lived transaction, as to ON AGENT
complete query optimization of mobile database. Based on the above analysis and researches, a MDB mod-
el was designed, and the architecture of the model is
shown in Fig. 1[7].

Figure 1. Architecture of MDB model based on agent


3.1 Mobile Terminal Layer
This layer includes all mobile terminals having the ability
of transaction processing. The model provides a two-way
2 RELATED TECHNOLOGIES OF AGENTS strategy that not only is applied to requesting transaction,
Agent is a kind of computer system that is in a specific submitting transaction, but also is applied to accepting
environment and can perceive surroundings. Agent is a transaction and processing transaction in this layer. The
computing entity or program that can achieve a series of layer includes management agent and interface agent.
goals of designers and user, and can flexibly and auto- The main responsibility of management agent is to preli-
nomously operate in that kind of environment. Intelligent minarily analysis and process transaction submitted by
agent provides a new way of computing and problem users or user applications, and to management local
solving, and a way of intelligent remote programming. transactions of database. The main responsibility of inter-
face agent is to process global transactions and remote
2.1 Multi Agent System transactions, or to act as an interface between local data-
Multi agent system is consisted of many interactive and base and wireless net database server.
interrelated agents. Multi agent system’s ability of solu- The functions of the two kinds of agents and the other
tion and collaboration is over a single Agent. In multi units are described as follow.
agent system, people are concerned about how to organ-  Management agent: Its main function is to manage the
ize the existent agents to complete the work a single agent local computing resources of mobile computing devic-
can not do. es. These resources can be used to execute mobile
agents. The agent preliminarily analyses and processes
2.2 Mobile Agent transaction submitted by users or user applications,
Mobile Agent can freely move among the hosts of net- and manages local transactions of database.
work. After a mobile agent is constructed in an environ-  Interface agent: As an interface, the agent needs to
ment, it can purposefully and independently move to create a context of processing transactions for mobile
another environment carrying its own codes and states agents and to provide a public data form of expression
and continue to run. Mobile agent is a kind of software for local datum stored in mobile devices. Thinking of
entity completing distributed tasks by local interaction the characteristics of XML, XML can be used as the
with the distributed resources [6]. Mobile agent is an ex- public data form of expression. An interface agent run-
pand way of traditional Client/Server. It can bring more ning in a given mobile unit needs to temporarily trans-
flexibility, efficiency, reliability and intelligence to distri- fer one or more services to the other mobile devices in
buted calculation, so mobile Agent has more and more the mobile database group for improving the efficiency.
JOURNAL OF COMPUTING, VOLUME 2, ISSUE 12, DECEMBER 2010, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 90

At the same time, once some crises happen, the inter-  Operating agent: The main task of it is to implement
face agent will also bear the responsibility of the above. the tasks that the mobile units need to complete in the
So, when any mobile agent visits the interface agents of MDBC. These tasks include data querying, data updat-
the mobile unit, it must declare the service that comes ing or mobile database model updating.
from which mobile device.  Transmitting agent: The main task of it is to transmit
 Transaction pretreatment unit: The unit initially database querying results to the mobile units that need-
processes the transactions submitted by users or appli- ing the querying results.
cations. Then the unit analyses whether or not the  Pick-up agent: The main task of it is to collect a variety
transactions can be processed in the local database, and of heterogeneous database model in the MDBC when a
divides these transactions into local transactions and new mobile database unit is added into one MDBC.
global transactions. In the end, the local transactions are  Temporary agent: It can define the mobile unit as a
sent to the local database server, the global transactions temporary storage device, and temporarily store the
are sent to the global/remote transaction manager. querying results in the storage device, when a sub-
 Global/Remote transaction manager: After the manag- query transaction gets a part of querying results from a
er receives the global transactions coming from the mobile unit.
transaction pretreatment unit, the interface agent re- When mobile agents arrive at one database node, its
sided in the manager will select the right node in the subquery can be extracted by a mobile agent platform,
MDBC to process these global transactions by the re- and be transmitted to the database agent in the node. The
sources searched from database. The manager can de- database agent that having related connecting drivers of
compose the global transactions into many sub- local database and related implementation interfaces can
transactions, and attach these sub-transactions to mo- further reduce the amount of data carried by the mobile
bile agents. These sub-transactions move to correspond- agent. Database agent can transfer sub-query to local
ing units with the mobile agents to be operated in the query, and trigger local database transactions also. Then
MDBC. database agent transfers the query results to the uniform
 Mobile agent platform: The module is responsible for XML form, and returns them to the mobile agent plat-
the creation of all kinds of mobile agents (the four form.
agents described above), and the summary and analysis
of all kinds of information fed back by mobile agents. 4 AGENT NAVIGATION AND ADAPTATION
The world of an agent is dynamic and uncertain. Ma-
3.2 Mobile Agent Layer chines go up and down, the information stored in reposi-
Mobile agents are responsible for the implementation tories changes, and the exact sequence of destinations and
of the logic mobile functions in this mobile database ar- steps needed to complete an information-gathering task
chitecture. It should have the ability to transfer itself to often is not completely known at the time that the agent is
another member from one member of MDBC. In this mo- launched into the world. An autonomous agent is crip-
bile database architecture, the basic abilities of the mobile pled without external state (what the agent can perceive
agents are to carry and transmit visiting codes and visit- about the state of its world) since it has no way of per-
ing results. ceiving and adapting to the dynamic changes in its envi-
ronment. In this section it is described the sensors that
allow an agent to determine its external state and a me-
chanism that uses these sensors for adaptive navigation.

Network sensing, at least the ability for a laptop to


detect the state of its network connection, is an integral
part of our laptop docking system described in the pre-
vious sub-section [8]. It performs an even more important
task, however, when providing agents with information
about the expected transit time across the network and
about whether a network site is reachable at all. This in-
formation enables agents to adapt to changing network
conditions. Consider an agent that needs to visit informa-
tion resources at several sites. A smart agent should be
able to adapt to the fact that some sites may currently be
unreachable, and to visit other sites first. An even smarter
Figure 2. The structure of mobile agents
agent may be able to plan a sequence of visits given an
estimate of the current network delay to each site. Other
agents may wish to tailor their behavior to the current
Mobile Database among every member in the MDBC.
bandwidth available, such as the amount or format of the
The mobile agent layer includes operating agent, trans-
data that it carries with them.
mitting agent, pick-up agent and temporary agent, as
A set of network sensing tools that the agents can use
shown in the Fig. 2. The main tasks of these mobile agents
to gather information about the status of the network as
are as follows.
JOURNAL OF COMPUTING, VOLUME 2, ISSUE 12, DECEMBER 2010, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 91

shown below: yellow pages are not static entities. The adaptive learning
 A tool to determine whether the local host is physically methods to keep the virtual yellow pages up to date.
connected. This tool \pings" the broadcast address on New services register with one or more navigation
the local subnet; if there is any response in a short in- agents to advertise their location. For example, in Figure
terval, the network is connected. 3, service 1 registers with navigation agent 2 by the fol-
 A tool to determine whether a specific host is reachable; lowing protocol: service 1 first contacts the specialist
this is just the standard \ping." agent on its machine which knows the location of naviga-
 A tool to determine the expected bandwidth to a re- tion agent 2. Service 1 then sends a registration message
mote host, so that agents can choose their destination or to navigation agent 2 which adds service 1 to the data-
amount of data based on the bandwidth. Rather than base.
measuring the bandwidth by sending lots of data to the An application agent locates a list of navigation agents
remote host, which would often take as much time as by querying the specialist agent on the local host (Figure
sending the agent itself, the attempt to predict band- 3). The application agent then consults the navigation
width from experience. A track monitor agent at each agents by providing a list of keywords. The navigation
site tracks information about all recent communications agent returns a list of matching services from its database.
(bytes moved and time required), which is provided by After visiting some of the services, the application agent
the local agent server. Application agents contact the revisits the navigation agents to provide feedback on
network monitor to obtain estimates of bandwidth or which of the sites were useful and which were useless.
latency, which are computed from the recorded infor- These \consumer reports" enable the navigation agents to
mation. If there are no recent communications with the prioritize their lists.
requested site, the track monitor may use data from re- Agents that discover services accidentally report the
cent communications with \similar" sites, that is, other corresponding sites to the navigation agents. For exam-
sites in the same subnet or domain as the requested site. ple, services relevant to one task may be discovered while
handling a different but related task. Such a situation
might arise if an agent handles textual queries about dif-
ferent topics; while finding documents relevant to one
topic, it may discover document collections that relate to
another. Alternatively, an agent might receive different
site information from two navigation agents; it feeds the
differences back to the navigation agents.
Each machine has a number of agents running on it
(denoted by rectangular blocks.) The specialist agents
know about the location of one or more navigation
agents. There are two navigation agents shown here: one
on machine 1 and one on machine 2. The navigation agent
on machine 2 knows about service 1, but the navigation
agent on machine 1 does not. The specialist agent on ma-
chine 3 knows about both navigation agents.
The customer agent on machine 3 uses the following
protocol to locate service 1. It first contacts its local spe-
cialist agent and finds the location of navigation agents 1
Figure 3: An example of navigation. and 2. Then it migrates to machine 1 and queries naviga-
tion agent 1 about service 1. This navigation agent does
Navigation agents. To locate other agents that can not know about service 1 since service 1 is only registered
serve their needs, agents need access to a dynamic index with navigation agent 2. The customer agent then mi-
of service agents and their locations. By consulting these grates to machine 2 where it queries navigation agent 2
navigation services and using their network sensing tools, and finds the location of service 1. Finally, the customer
agents can formulate adaptive navigation plans to visit agent migrates to the location of service 1. Application
some of the services. The virtual yellow pages are a dis- agents construct an initial plan for accomplishing their
tributed database of service locations maintained by a task by using the prioritized list of services that it receive
hierarchical set of navigation agents. Services register from the navigation agents. Most applications will want
with the navigation agents that are scattered throughout to visit either one or all of the sites on the list.
the system (Figure 3). Each machine has a specialist agent
that knows the location of some of the navigation agents
5 TRANSACTION PROCESSING PROCESS
(which in turn know the locations of services and other
navigation agents). In general, by consulting the local OF THE MDB MODEL
specialist agent and then visiting one or more navigation The model in Fig. 1 is an abstract model. In order to ex-
agents, an application agent can obtain the necessary list plain its operation process, assume the followings:
of services and their locations.  Assuming a mobile unit MU plan to construct a
Since the information landscape changes, the virtual MDBC[9] based on its own.
JOURNAL OF COMPUTING, VOLUME 2, ISSUE 12, DECEMBER 2010, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 92

 In order to the MDBC can be constructed, first of all, send a sign of completing the transaction to the interface
MU must declare to its own interface agent that it will agent of the MU. The interface agent will process the re-
be declared as initialization organizer in this MDBC to lated information of the transaction in the local MU, and
be built. show the result to the user or application of the mobile
 Then the MU sends the information of the new MDBC database. At last, the interface agent will come to process
to all mobile terminals covered by the MU’s wireless the next transaction.
network.
 When new members plan to join this new MDBC, the
6 CONCLUSIONS
organizer will no longer be required.
 Organization function will be completed by all kinds of Mobile transaction processing has become a challenging
agents in the MDBC. study field. In order to improve the processing efficiency
of mobile database system, a new mobile database model
5.1 Instance of Mobile Transaction Processing based on agent was proposed. And transaction
Based on the MDB Model processing process based on this model was set forth. The
First of all, in order to illustrate the details of all database model can shield the heterogeneous among all members,
models, a clear statement of the MU must be given by the and made them transparent so as to bring great conve-
local interface agent. At the same time, a pick-up agent niences to the entering of heterogeneous members. It’s
will be constructed and be initialized by the local inter- clear that the introduction of mobile agent can contribute
face agent. Then the pick-up agent moves in all mobile to decomposition of long-lived transaction, as to complete
units to pick up their local database models. The local query optimization of mobile database.
interface agent will provide local database model, and
transfer it to XML form, the formal form of model. REFERENCES
When the pick-up agent of the MU returns to the MU,
[1] A.Sripriya, Dr. R.Dhanapal, “Mobile Software Solutions using Mobile
it will bring back database models of all database mem- Database Design Methodology”, IEEE Explorer Proceedings & Interna-
bers, and will divert them to the interface agent of MU. tional Conference on Computing Communication and Networking Technol-
Thus, MU owns the database models of all mobile data- ogies ICCCNT 2010.
bases in MDBC [9]. Users or mobile applications of MU [2] A.Sripriya, Dr. R.Dhanapal, “A Cache Operation in Mobile Database
will have the ability to visit the datum stored in other Using Genetic Algorithm”, International Journal of Recent Trends in En-
gineering, Vol 2, No. 4, November 2009 (ISSN 1797 – 9617), 2009 Acade-
mobile devices in the MDBC through the interface pro- my Publisher, Finland (Page no. 204-206).
vided by the local interface agent [10]. When the interface [3] A.Sripriya, Dr. R.Dhanapal, “ Building Data Warehouses Using Mo-
agent of MU receives a global query request, it will con- bile Database Agents”, National Conference on Emerging Trends in In-
struct an operation agent in the MU, and send the infor- formation and Communication Technology 2009.
mation of global query request to the operation agent. [4] A.Sripriya, Dr. R.Dhanapal, “Authentication for Wireless Network
Then the operation agent begins to process the query Security using Mobile Agent”, National Conference on Recent Trends in
Computer Applications 2009.
transaction.
[5] Ren Wei, “Research on the System Framework based on the Agent
Mobile Database”, Micro Computer Information, 2007.5, pp. 157- 159.
5.2 Process of a Query Transaction
[6] Yajuan Zhang, Muning Kang, Zhenghong Deng, and Peng Zhang,
First of all, the interface agent of the MU constructs an “Research on Embedded Mobile Database System Based on Mobile
operation agent, then the operation agent move to the Agent”, Computer Engineering and Design, 2007.7, pp. 3429-3432
MU to process the transaction, because the MU includes [7] Zhao Yang, “Research on Model of Transaction Processing Based on
the target database of processing the query transaction Mobile Database”, Computer engineering, 2006. 5, pp.68-79.
[8] Madria S K, Mohania M, “Mobile Data and Transaction Manage-
[11,12,13].
ment”, Information Science, 2002.141, pp.279-309.
When the operation agent arrives at the MU, it will [9] S Buchholz, T Ziegert, A Schill et al, “ Transaction processing in a
submit the information of transaction processing to the mobile computing environment with alternating client hosts”, The
local interface agent of the MU. The local interface agent 10th Int’1 Work shop on Research Issues in Data Engineering
needs not to think of the type of the transaction. The type (RIDE’00) , San Diego, 2000.
of the transaction is transparent for the local interface [10] Seydim, A., “An Overview of Transaction Models in Mobile Envi-
ronments.” Retrieved December 29, 2003, from
agent because of the agency role of the operation agent, so http://engr.smu.edu/~yasemin/mobile_trans.pdf
as to improve the mobile transaction processing efficien- [11] Oracle Corp. “Oracle9i Lite.” April 22, 2004, from
cy. http://otn.oracle.com/products/lite/htdocs/o9ilite_datasheet.
When the result of transaction processing comes out, htm
the local interface agent will send the result to operation [12] Sybase, Inc. “UltraLite Database User’s Guide” Retrieved May 28,
2004, from http:// download.sybase.com /pdfdocs/ awg0900e/
agent. If the result is too large, and the operation agent
ulfoen9 .pdf
must move to other mobile units to process transaction at [13] Microsoft Corporation. “SQL Server CE 2.0 Product Overview.”
the same time, the operation agent can construct a trans- Retrieved December 31, 2003, from http:// www. microsoft.
mitting agent and send the result to the agent. The trans- com / sql/ce/productinfo/overview.asp
mitting agent will return to the starting node of the opera-
tion agent that constructing the agent along the moving
path of the operation agent.
When the operation agent returns to the MU too, it will
A.Priya is presently pursuing her Ph.D Computer Science in Bhara-
JOURNAL OF COMPUTING, VOLUME 2, ISSUE 12, DECEMBER 2010, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 93
thiar University, Coimbatore. She got her Masters degree in Com-
puter Science from Avinashilingam University, Coimbatore and Mas-
ter of Philosophy in Computer Science from Manonmaniam Sundar-
narar University, Tirunelveli. She is currently working as an Assistant
Professor in the Department of Computer Applications, Velammal
Engineering College, Chennai and Affiliated to Anna University of
Technology Chennai, Tamil Nadu, India. She has 11 years of Teach-
ing and Administrative experience and 2 years of research expe-
rience. She worked as a Controller of Examinations for the Auto-
nomous College affiliated to Thiruvalluvar University, Vellore. She
guided about 8 M.Phil Scholars in several universities such as Ala-
gappa University, Karaikudi, Prist University, Thanjavur, Mother Tere-
sa Women’s University, Kodaikanal. She is a Life time Member in
ISTE Chapter. She organized DRDO sponsored National Confe-
rence in the Department of Computer Applications, Velammal Engi-
neering College, Chennai. Her publications are one International
Journal, one International Conference (in IEEE Proceedings) and six
National Conferences.

Dr.R.Dhanapal obtained his Ph.D in Computer Science from Bhara-


thidasan University, India. He is currently Professor of the Depart-
ment of Computer Applications, Easwari Engineering College, Affi-
liated to Anna University of Technology Chennai, Tamil Nadu India.
He has 25 years of teaching, research and administrative expe-
rience. Besides being Professor, he is also a prolific writer, having
authored twenty one books on various topics in Computer Science.
His books have been prescribed as text books in Bharathidasan
University and autonomous colleges affiliated to Bharathidasan Uni-
versity, Tiruchirapalli. He has served as Chairman of Board of Stu-
dies in Computer Science of Bharathidasan University, member of
Board of studies in Computer Science of several universities and
autonomous colleges. Member of standing committee of Artificial
Intelligence and Expert Systems of IASTED, Canada and Senior
Member of International Association of Computer Science and In-
formation Technology (IACSIT), Singapore. He has Visited USA,
Japan, Malaysia, and Singapore for presenting papers in the Interna-
tional conferences and to demonstrate the software developed by
him. He is the recipient of the prestigious ‘Life-time Achievement’ and
‘Excellence’ Awards of Govt. of India. He is serving as Principal In-
vestigator of UGC sponsored innovative, major and minor research
projects about 1.6 crore. He is the recognized supervisor for re-
search programmes in Computer Science leading to Ph.D and MS
by research in several universities including Anna University of
Technology Chennai, Bharathiar University, and Manonmaniam
Sundaranar University. He has got 47 papers on his credit in interna-
tional and national journals.

T.P.Vijayalakshmi obtained her Masters in Computer science from


Avinashilingam University, Coimbatore and M.Phil in Computer
Science from Mother Teresa University, Kodaikanal, India. She is
currently working as a SAP Data Archiving Consultant, Wipro Tech-
nologies, Client - Apple Inc, Cupertino, California, U.S. She has 14
years of IT, Teaching and Administrative experience.

Potrebbero piacerti anche