XML-BASED IMPACT ANALYSIS USING CHANGE-DETECTION

APPROACH FOR SYSTEM INTERFACE CONTROL

Namho Yoo

Department of Computer Science, The George Washington University,

801 22

Street, N.W., Room 730, Washington, DC 20052, USA

Keywords: XML, Impact Analysis, System Engineering, Interface Control, Change Detection, XPath

Abstract: In this paper, an XML-based approach is presented for the impact analysis of interface control in sustained

systems. Once a system is completed developed, it goes into a sustained phase supported by many

interfaces. As new technologies develop, updating and maintaining such systems require non-trivial efforts.

A clear pre-requisite before the deployment of a new system is to clarify the influence of changes on other

systems connected through interfaces. However, as each sustained system manages its own information

separately, integrating relevant information among the interfaced systems is a major hurdle. In our approach,

the XML technology is applied to support impact analysis for interface control architecture using change-

detection approach for the reference. In particular, I focus on messaging interface issues in Health Level

Seven typically used in medical information system and propose a scheme to represent message information

that can be used for the decision support of interface impact between sustained systems.

1 INTRODUCTION

System engineering efforts are vital to make

decisions before making system changes associated

with new requirements. In large-scale system, many

supporting engineers would be involved in the

engineering process through the technical

committee. As system environments change rapidly

and engineering change efforts for impact analysis

increase significantly, many documentation is

considered to carefully examine in terms of non-

functional view as well as functional perspective.

Engineering Change Proposal (ECP), which is one

of the key documentation for configuration

management, contains the description about change

requests/requirements and how they should be

reviewed and implemented(Arnold, 1993). The

proposed approach, therefore, is the extended

framework with impact analysis for interface

control using change-detection approach with XML

like Figure 1.

This paper suggests an effective process and

method to consider interface control requirement

during the impact analysis. For dealing with the

interface control, change detection approach (CDA)

for impact analysis is discussed as a lightweight tool.

In order to handle, interface control requirement,

engineering change proposal (ECP) for impact

analysis are covered.(Yoo, 2004)

Impact Analysis for

Interface Control

IA-IC

Application

IA-IC

Parts

IA-IC

Effectiveness

HL7 Message Interface

XML Technology and DOM Tree

Change Detection Approach

Testing & Implementation

Impact Analysis for

Interface Control

IA-IC

Application

IA-IC

Parts

IA-IC

Effectiveness

HL7 Message Interface

XML Technology and DOM Tree

Change Detection Approach

Testing & Implementation

Figure 1: Extended Impact Analysis for Interface Control

(IA-IC) Framework

Also, since there is system resource information

without specification, common vehicle to exchange

information effectively is required. In this paper,

Extensible Markup Language (XML) technology is

used for supporting engineering change efforts for

impact analysis. An XML is widely used as a

standard for information exchange on the World

Wide Web.(W3C, 2000)

506

Yoo N. (2005).

XML-BASED IMPACT ANALYSIS USING CHANGE-DETECTION APPROACH FOR SYSTEM INTERFACE CONTROL.

In Proceedings of the Seventh International Conference on Enterprise Information Systems, pages 506-509

DOI: 10.5220/0002534305060509

 SciTePress

2 BACKGROUND AND PROBLEM

STATEMENTS

Under considering an world-wide deployed US

health system involving 8 sub-systems A through H

with more than 50 to 200 sites, a impact analysis for

interface control is essential for decision-making.

Consider the interface between systems A and B.

The Health Level Seven (HL7) protocol messaging

supported by the interface engine between A and B

is given in Figure 2 as an example from.

Reposit

ory

Database

(future)

* A has a large number of external system interfaces that have not been shown.

Reposit

ory

Database

(future)

* A has a large number of external system interfaces that have not been shown.

Figure 2: Sustained System Interface Architecture

In table I, the several cases for change are

presented. In change cases (1-3), even though it

may takes for a while to find out the result whether

or not to analyze the impact, system engineer can

say Yes or No. But the analysis result is dependent

to their knowledge and experience. And there are

many cases existed not for sure.

Table I: Change Case Table for the system

Change

Case

System

1 Insert Yes/No

2 Delete Yes/No

3 Update Yes/No

4 Volume N/A

5 Environment N/A

6 Combined N/A

If change cases (4-6) are considered, the situation

for system impact analysis for interface control is

getting harder than previous case. For example, if

the messaging volume of the system A is 25%

increasing, how much of system B is affected by

this change in terms of performance. What about

the system interface impact when the system A

change the security policy and install the new

security patch at the same environment? Is this

change providing impact for the system B? In order

to solve the problem caused by change, the system

testers need to be involved more system and

sustained system project schedule is delayed

accordingly. Creating ECP for Configuration

Management (CM) should be required.

3 XML-BASED IMPACT

ANALAYSIS FOR INTERFACE

CONTROL

In this paper, the way to do system impact analysis

is discussed for system engineer more effectively

and accurately in time manner. The situation of

change request based on health system using HL7

protocol is considered. And the effective

representation is addressed using XML.

<?xml version="1.0" encoding="EUC-KR"?>

<Segment1>MSH

</Element1>

<Element4>LABADT</Element4>

<Element6>Security</Element6>

</Element7>

</Segment1>

<Segment2>EVN

</Segment2>

<Segment3>PID

<Field1>PATID12345</Field1>

</Element2>

<Field6>USSSA</Field6>

</Element4>

<Field1>JOHNES</Field1>

<Field2>WILLIAM</Field2>

</Element5>

<Field4>STREET</Field4>

<Field6>GREENSBORO</Field6>

</Element8>

<Field1>PATID12345001</Field1>

</Segment3>

<Segment4>NK1

<Field1>JOHNES</Field1>

<Field2>BARBARA</Field2>

</Element2>

</Element3>

</Element4>

</Segment4>

<Segment5>PV1

</Element3>

<Field2>LEBAUER</Field2>

<Field3>SIDNEY</Field3>

</Element4>

</Segment5>

</HL7import>

<?xml version="1.0" encoding="EUC-KR"?>

<Segment1>MSH

</Element1>

<Element4>LABADT</Element4>

<Element6>Security</Element6>

</Element7>

</Segment1>

<Segment2>EVN

</Segment2>

<Segment3>PID

<Field1>PATID12345</Field1>

</Element2>

<Field6>USSSA</Field6>

</Element4>

<Field1>JOHNES</Field1>

<Field2>WILLIAM</Field2>

</Element5>

<Field4>STREET</Field4>

<Field6>GREENSBORO</Field6>

</Element8>

<Field1>PATID12345001</Field1>

</Segment3>

<Segment4>NK1

<Field1>JOHNES</Field1>

<Field2>BARBARA</Field2>

</Element2>

</Element3>

</Element4>

</Segment4>

<Segment5>PV1

</Element3>

<Field2>LEBAUER</Field2>

<Field3>SIDNEY</Field3>

</Element4>

</Segment5>

</HL7import>

Figure 3: Converted XML from HL7 Sample Message

Figure 3 is the corresponding representation of

HL7 example. XML representation is more

understandable for system engineer and scalable to

different format for future needs. Whenever

XML-BASED IMPACT ANALYSIS USING CHANGE-DETECTION APPROACH FOR SYSTEM INTERFACE

CONTROL

507

appropriate, it can be possibly triggered and

converted by the development.

Figure 4(upper part) depicts the ideal data flow

among the different modules in this interface. The

HL7 message received by the communication

modules from the external system goes to the HL7

engine module. After the HL7 engine module edits

the HL7 message, it passes the decoded data to the

queue and the encoded acknowledgement back to

the communication module. The communication

module then sends back the acknowledgement to

the external system. The parsed data stored in the

queue will then be mapped and sent to the receiving

system/HL7 internal file. Likewise, if the edit is not

successful, no data will be sent to the

parsed/destination file.

Figure 4 (lower part) shows the XML based

information architecture mapping to each offline or

online assets. Our focus is regarding offline assets

such as documents and data dictionary. Using these

artifacts, the DTD and Document XML are

generated. However as there exist some

inconsistency or some gaps between the document

and real system, it is also required to verify the

information using online log file and parsed data.

HL7

Engi ne

HL7 Msg

Ack

Queue

Decoded

Da ta

Mapped Data

Par sed

Data

Appli cation

Database

Log File

Dat a Dict ionar y

Doc ume nt s

Impact Analysis

For Interface Control

Res ul t

Log XML

Me s s age

XML

DTD XML

Doc u X ML

SysA DOM Checker

SysB DOM Checker

Compare DOM

Filtering

Saving Difference

Maintain IIA

Verifying Result

CSV_ 2_ XML

XML_2_ DOM

Up load

Web

TXT_2_XML

Tract

HTML_2_XML

DAT _ 2_ XML

Save

Loc al

Dis k

Rul e File

HL7

Engi ne

HL7 Msg

Ack

Queue

Decoded

Da ta

Mapped Data

Par sed

Data

Appli cation

Database

Log File

Dat a Dict ionar y

Doc ume nt s

Impact Analysis

For Interface Control

Res ul t

Log XML

Me s s age

XML

DTD XML

Doc u X ML

SysA DOM Checker

SysB DOM Checker

Compare DOM

Filtering

Saving Difference

Maintain IIA

Verifying Result

CSV_ 2_ XML

XML_2_ DOM

Up load

Web

TXT_2_XML

Tract

HTML_2_XML

DAT _ 2_ XML

Save

Loc al

Dis k

Rul e File

Figure 4: Proposed XML based IA-IC Architecture

As defined by HL7 standard, each message is

composed of logical groupings of data entities,

named message, segment, and elements. Some

segments may be optional or repeatable in a

message. Figure 5 depicts the logical structure of

the entities and some examples.

Message

Segment

Element

ADT/A01

HL DG ADMIT PATIENT

PID

HL DG PATIENT ID

PV1

HL DG PATIENT VISIT

PID

HL DG PATIENT ID INTERNAL

HL COUNTRY CODE

PV1

HL DG ADMISSION TYPE

ADMITTING DOCTOR

Not

Applicable

System B

Figure 5: HL7 Example for IA-IC

Based on the testing for the research with XML-

based representation, an HL7 message transferred

from system A to system B used is found only about

30% of the time in terms of message-level and only

50% of the messages are being used if I consider

element-level. In other words, by using XML-based

message modeling the meaningful portion of

messages is identified for decision-making and is

managed the part required to enhance interface

impact analysis as well as updating impact analysis

for each system.

4 CHANGE-DETECTION

APPROACH

The time for efforts and potential error is mitigated.

IA-IC XML DOM Tree

IC Rule Tree

(System A)

IC Rule Tree

(System B)

IC Message Tree

(System A)

IC Message Tree

(System B)

IA-IC XML DOM Tree

IC Rule Tree

(System A)

IC Rule Tree

(System B)

IC Message Tree

(System A)

IC Message Tree

(System B)

Figure 6: IA-IC XML DOM Tree

This paper suggests extended IA-IC tree by

considering combined scheme with HL7 testing and

change detection approach by XML technology.

ICEIS 2005 - INFORMATION SYSTEMS ANALYSIS AND SPECIFICATION

508

Figure 6 is an example of DOM integration with

message trees and rule trees.

For traversing the information under the XML

DOM tree structure, given algorithm using XPath

and XSLT is used. XPath takes a navigational

approach for specifying the nodes to be selected,

hence a large number of navigational axes have

been defined in XPath.

5 IMPLEMENTING XML-BASED

IA-IC

A common environment as shown(Figure 7) can be

built for the simulation of the performance

requirement and the impact analysis on the interface.

(a) Extracting the change of artifacts

(b) Comparing the difference of control documents

Figure 7: Executable IA-IC Architecture Implementation

6 CONCLUSIONS AND FUTURE

WORK

The engineering issues are discussed to meet

changing functional requirement in systems

maintenance phase. In particular, the impact

analysis for interface control is focused on in the

presence of interface requirement and performance

requirement in large scaled sustained system. A

baseline using XML-based representation to handle

changing functional requirements is considered with

the process for impact analysis. The XML-based

change-detection approach enables supporting

engineers’ collaboration effectively to meet the

limited time requirement. Through a health system

example, engineering steps are addressed for

integrating information and sharing analysis result

during ECP process.

REFERENCES

Arnold, R., Bohner, S., 1993, Impact Analysis – Toward

A Framework for Comparison, In Proceeding of

Conference on Software Maintenance, pp 27-30,

September.

W3C, 2000, Extensible Markup Language (XML) 1.0 ,

W3C Recommendation, October

Yoo, N., 2004, Impact Analysis using Performance

Requirement with Application Response

Measurement in Sustained System, In Proceedings of

the ISOneWorld Conference.

Yoo, N., 2004, An XML-based Engineering Change

Impact Analysis with Non-Functional Requirements,

In Proceeding of SERP.

XML-BASED IMPACT ANALYSIS USING CHANGE-DETECTION APPROACH FOR SYSTEM INTERFACE

CONTROL

509