Creative Education
2012. Vol.3, Supplement, 59-62
Published Online December 2012 in SciRes (http://www.SciRP.org/journal/ce) DOI:10.4236/ce.2012.37B014
Copyright © 2012 SciRes.
59
A Legacy System Encapsulation Strategy Based on Web Service
D. J. Chen1, X. Y. Li2
1Dali Comprehensive Inspection Centre of Quality and Technical Supervision, Dali, P.R. China
2School of Mathematics and Computer Science, JiangHan University, Wuhan, P.R. China
Email: 652187048@QQ.com, Lixy_76@163.com
Received 2012
In order to migrate the enterprise legacy system to the web, a multi-agent based legacy system encapsula-
tion model is proposed. Firstly, the characteristics of legacy system are analyzed, and then the data and
functions that need to be published are confirmed. Secondly the legacy system is wrapped into web com-
ponents with common interface, and these components are managed by the application server. Thirdly,
the clients can send requests to the application server, and receive the return result from the application
server. Due to adoption of wrapping technology for legacy system, original security and stability of
legacy system are guaranteed in the web components. Finally, the validity and practicability of the migra-
tion technology are verified through the application in the encapsulation of Matlab as web components.
Keywords: Legacy System; Multi-agent; System Migration; System Encapsulation
Introduction
At present, there isn't a uniform definition on le gacy syste m,
in 1995, there was a definition proposed in [1]: if a system is
important to the enterprise, and at the same time, the enterprise
doesn’t know how to update it to meet the new need, the system
can been called as a legacy system. To legacy system, nowadays,
there are three basic models to rebuild it to meet the new need.
Giving up legacy system: the model will give up a legacy
system and a new system is developed to replace it at the same
time [2]. The model will produce waste to the enterprise's re-
source, so it isn't used by most enterprise.
Wrapping legacy system: the model will hide the real im-
plementation of legacy system, and its functions are published
as public interfaces, then we can access the legacy system
through these public interfaces [3]. In the model, the security
and stability of legacy system aren’t affected, and the rebuild-
ing cost is low.
Rebuilding legacy system: in the model, before rebuilding
the legacy system, we must analyze its functions and demands,
and then we can implement its functions in the new system [4].
In fact, for the half-baked documents of legacy system, it is
very difficult to analyze its functions.
In the three models, the second model is adaptive to enter-
prise. In the model, the inherent security and stability of legacy
system are not affected in the new system, and this is very im-
portant to an enterprise.
The reminder of this paper is structured as in the following:
section 2 gives some background information in legacy system
integration technology, section 3 gives an introduction on the
multi-agent based system encapsulation model. In section 4
gives a case study to describe how to encapsulate the Matlab as
web components based on multi-agent technology, and finally,
section 5 contains the conclusion.
Legacy System Integration Tech nology
XML Integration
The Extensible Markup La nguage (XML ) is a bro adly adopted
format for structured documents and data on the Web. XML is
a simple and flexible text format derived from standard genera-
lized markup language and developed by the World Wide Web
Consortium. XML is expanding from its origin in document
processing and becoming a solution for data integration. XML
excels in inter-application data exchange because of its flexible
and extensible method for describing data and its capability to
communicate over the Internet using the standard HTTP proto-
col.
This flexibility makes XML a powerful mechanism for busi-
ness-to-business (B2B) application integration. B2B integration
is the automated exchange of information between systems
from different organizations. B2B, for example, improves ex-
ternal processes such as supply chain integration or ship-
ping/logistics tracking. XML-based B2B is gaining momentum
as XML vocabularies emerge in specific business domains such
as finance. In addition, a growing number of commercial enter-
prise application solutions are embracing XML [5].
Database Replication
Database replication is the process of copying and maintain-
ing database objects in multiple databases that make up a dis-
tributed database system. Changes applied at one site are cap-
tured and stored locally before being forwarded and applied to
the centralized repository. Replication provides users with fast,
local access to shared data and greater availability to applica-
tions because alternati ve data access opt ions exi sts. Even if one
site becomes unavailable, users can continue to query, or even
update, data at other locations. Database replication is often
used to enable decentralized access to legacy data stored in
mainframes. In this approach, local instances of modern data-
bases are replicated parts of a centralized database. We can get
a similar effect using a remote database gateway, but it is
sometimes undesirable to use a software layer to wrap legacy
data. Instead, we want a local copy of the data that wraps and
buffers the original data source. New applications using the
data receive the benefits of local access to a modern database
D. J. CHEN, X. Y. LI
Copyright © 2012 SciRes.
60
instead of the problems of remote access to an obsolete data
repository.
Object-Oriented Wrapping
Objects have been used to implement complex software sys-
tems successfully. Object oriented systems can be designed and
implemented in a way that closely resembles the business
processes they model. Additionally, the use of abstraction, en-
capsulation, inheritance, and other object orientation (OO)
techniques make object-oriented systems easier to understand.
To support object distribution, we need a more powerful
means of communication than that provided by normal inter-
object communication mechanisms. Distributed object tech-
nology (DOT) is the combination of distributed technology
with OO. In effect, DOT extends object technology to the
net-centric information systems of modern enterprises by using
object middleware. The most prevalent object middleware is
the Common Object Request Broker Architecture (CORBA)
from OMG, with its platform-neutral object specification lan-
guage, robust remote method calls, interoperable protocols, and
rich set of services. The conceptual model of object-oriented
wrapping is deceptively simple: individual applications are
represented as objects; common services are represented as
objects; and business data is represented as objects. In reality,
object-oriented wrapping is far from simple and involves sever-
al tasks including code analysis, decomposition, and abstraction
of the OO model [6].
Legacy System Encapsulation Model Based on
Multiagent Technology
Based on multi-agent techniques, every function of legacy
system can be encapsulated as a web component, which may
have various interfaces for different functions. The multi-agent
based encapsulation architecture is shown Figure 1.
External wrapper provides various interfaces to accept SOAP
request, which can be managed by the client’s power, and inner
wrapper are developed for the legacy system in the institutions.
The SOAP requests from online user are transmitted to external
wrapper, and then retransmitted to inner wrapper. Then the
inner wrapper decomposes the requests into a sequence of sta-
tus values, according to which web component will change the
state, and related web components are called to respond to the
request. Finally the result should be sent back to external
wrapper which in the same way sends it to the online user.
Figure 1.
System encapsulation based on multi-agent technology.
External Wrapper
External wrapper accepts requests and cooperates with inner
wrapper. Its main tasks are to analyze and decompose the re-
quests, including creating service detailed lists, decomposing
request to status switch, defining communications protocols
and the format of input and output, registering and issuing web
components in the UDDI.
External wrapper is closed to the actual operation platform,
and access request through inner wrapper, external wrapper is
used to establish and maintain the safe communication among
different web components. When communication happens
among different legacy system systems, the conversion between
bytes and certain data type must be run to ensure platform-
independent. The communication information is described and
defined uniformly in XML.
Inner Wrapper
Inner wrapper needs to coordinate with external wrapper, and
interact with legacy system systems directly. Therefore, they
should be standardized and adaptable to the operation environ-
ments of legacy systems. The architecture of inner wrapper is
shown in Figure 2.
The net communication agent receives service request from
external wrapper, and sends it to service decomposing agent,
which decomposes the service request to the status switch set
based on transaction logic library. The net communication
agent is composed of external communication agent and inner
communication agent. Then according to the status data, the
monitoring operation agent manipulates related subsystems and
sends the needed output to external communication agent.
Based on multi-agent technology, the legacy system is en-
capsulated with common interfaces, and other systems can
communicate information with the legacy system based on the
common interfaces, but to the clients who don’t be familiar
with the common interfaces, we can’t send the request with
right format to the legacy system and get the right result.
For providing the access interface to clients, as shown in
Figure 3, an encapsulation architecture of legacy system is
proposed, as following describes how to clients to send the
request to the legacy system and get the return results.
Clients send the request message to the application mes-
sage with SOAP forma t .
Figure 2.
The architecture of inner wrapper.
D. J. CHEN, X. Y. LI
Copyright © 2012 SciRes.
61
Figure 3.
The encapsulation architecture of
legacy system.
Application server receives the request and sends it to the
legacy system access middleware based on the JDBC,
ODBC or OPC interface.
The legacy system access middleware sends the request
the request to the legacy system wrapper adapter, which
includes external wrapper and inner wrapper.
The legacy system wrapper adapter sends the request to
the legacy system based on their common access interfac-
es.
The legacy system receives the request, and returns the
result to the legacy system adapter, and finally, the result
is returned to the clients.
Based on the proposed encapsulation architecture, the legacy
system is encapsulated as common web components, and then
we can integrate the legacy system with other legacy systems,
and these legacy systems can share message with SOAP format.
Due to adoption of wrapping technology for legacy system,
original security and stability of legacy system are guaranteed
in the web components.
Case Study
Matlab is professional software developed by MathWorks
Corporation. It has powerful function on data analysis, data
emendation, and image process etc. Based on the proposed
system encapsulation model, we can encapsulate it as Web
component to provide the computation functions, linear regres-
sion analysis, nonlinear regression analysis, matrix normaliza-
tion, covariance matrix computation [7,8]. Regression analysis
is mainly use to analyze real-time data for the product devel-
opment process.
As shown in Figure 4, the Matlab is encapsulated as com-
mon web components, as shown in Figure 5, the SOAP mes-
sage can be sent to the Matlab common component to request
linear regression function, and then the result will return with
SOAP message as shown in Figure 6.
Figure 4.
The encapsulation model of Matlab.
Figure 5.
SOAP request message of linear regression service.
D. J. CHEN, X. Y. LI
Copyright © 2012 SciRes.
62
Figure 6.
SOAP response message of linear regression service.
Conclusion
For encapsulating the enterprise legacy system as common
web components, a multi-agent based legacy system encapsula-
tion model is proposed. Due to adoption of wrapping tech-
nology for legacy system, original security and stability of leg-
acy system are guaranteed in the web components. Finally, the
validity and practicability of the migration technology are veri-
fied through the application in the encapsulation of Matlab as
web components.
REFERENCES
W. Proie, “Legacy computer systems migration,” AISTech - Iron and
Steel Technology Conference P roceedings, 2, pp. 873-881, 2009.
Y. Qu, “Research and application of reusing enterprise legacy systems
based on Web Service,” 2009 WRI World Congress on Software En-
gineering, 3, pp.325-329, 200 9.
X. Li, “A service-oriented partitioning and encapsulation method for
java legacy systems,” Jisuan ji Xuebao/Chin ese Journal of Co mputers,
32(9), pp.1804-1815, 2009.
W. Gee, “Systems engineering requirements for legacy hardware up-
grade and sustainment requirements definition, analysis, and valida-
tion,” 2009 IEEE Intern ational Systems Conference Pro ceedings, pp.
431-436, 2009.
B. Medjahed and A. Bouguettaya, “Composing web components on the
semantic web,” Journal on Very Large Databases, 4 , p.
333-351(2003).
J. Y. Sheng and J. Hang, “CORBA based tele-collaboration workflow
mode,” Com puter Applications, 9, p. 19-25(2002).
Y. Zhan, J . A. Clark, "A search-based framework f or automatic testing
of MATLAB/Simulink models," The Journal of Systems and Soft-
ware, 81(2), pp.262-285, 2008.
C. Y. Chen, Y. C. Shun, "MATLAB-based rapid controller develop-
ment platform for control applications," Journal of Mechanical En-
gineering Science, pp.1461-1473, 2007.