Proceedings of the 29th Annual Hawaii International
Conference on System Sciences - I996
Building an O rganizational Decision Support System for Inves
Performance Management : Korea Telecom Case
Young-GUIKim, Hee-WoongKim
KAIST, Departmentof M IS
P.O. Box 201, Cheong-Ryang,Seoul, 130-650,Korea
Internet: ygkim, hwkim@msd.kaist.ac.kr
Jae-WookYoon, Ho-Sung Ryu
Korea Telecom Research Institute
17 Woomyeon-dong,Seocho-gu,Seoul, Korea
Internet:jwyoon, ryoohosg@ktrc.kotel.co.kr
Abstract
are taking place in the Korean telecommunications
industry. Reqmding to the ever increasingcustomer
demandfor higher qualily, Koreantelecommunications
Organizational decision support systems (ODSL$ are a
new type of decision support systems(DSS)focusing on
companies are stepping up their qua&y-based
competition. Korea Telecom(KT), which had
the organization-wide issues rather than individual,
monopolizedthe domestictelecommunicationsservice
group, or departmental issues. Because of its
organization-wide scope, a typical ODSS cuts across
marketuntil recently,needsto shakeup its old operating
mode drasticallyto improve its customerservice and
organizational functions or hierarchical lqers. Thus,
sustain its competitiveadvantage.Thanks to Korea’s
seamless integration with organization’s diverse IS
telecommunicationmodernizationdrive started in the
applications running on heterogeneous platforms
198Os,
becomes a critical issuefor building a successful ODSS.
hardware
part of the KT’s
--lY
In this paper, we analyzed the Korea Telecom (22)‘s
telecommunications
ti at&uctme (electronicswitching
Operations and Maintenance (O&11) division focusing
devices, fiber-optic communicationlines, etc.) has
on its investment strategies.
We developed a
improvedsign&antly over the years. It is now the
conceptual framework which links O&M investment
softwarepart (operationsand maintenance)of KT that
decisions to performance of its operational branches
needs to be the fixus of the KTs new drive for
across the nation. In this framework, performance of
improvingits customerservicequality.
each local branch is evaluated on two dimensions:
In this study, we introduce a prototype
quality of service (QOS) and d&f&x.&y of the local
Organizational Decision Support System(ODSS)to
operating environment (CDI: context di$iculty index).
support the investment performance management
This approach will enable O&A4 management to
processwithin ICl?sOperations& Maintenance(O&M)
explicate the major causes behind differences in service
division. This ODSS is built upon the integration of
quality among local branches and to determine proper
KT’s nwneroustransactionprocessingsystemsacross
scale and area of O&M investment for competing
the country and will be used by not only the O & M
branches. To support the above framework, we also
headquartersoffice but also the ten district business
developed a prototype for the KTOM-ODSS with an
offices. Insteadof just supportingthe current decision
EIS-like user-friendly interface. When a complete
process,we proposedto “reengineer”
the processfirst to
ODSS is implemented on top of various KT transaction
bettermeetKT’s internal and externalcustomerneeds.
processing systems, it will become a critical component
With KT management’
agreement,
s
we then developed
of the O&M Integrated Decision Supp6rt EHvironmerrt
the ODSSprototype to support the reengineered
O&M
(IDS@.
process.
The rest of the paperproceedsas follows. In section
1. Introduction
2, we will introduceKT, KT O & M division, and KT
O & M division%-motivation to develop au ODSS.
With the opening trends of the international Section 3 provides overview of the ODSS literature
telecommunications
servicemarket,significantchanges relevant to our study. Section 4 describesthe KT
* This paper was supported by KT(Korea Telecom) in 1994.
1060-3425/96
$5.00 0 1996 IEEE
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
62
Proceedings of the 29th Annual Hawaii International
O&M division’s investment performancemanagement
process and how we reengineeredit. Section 5
illustrates the architecture of the O&M ODSS and
explains how we implementedits prototype.In section
6, we draw the conclusion of the paper and suggest
Iirture research directions related to extending the
current ODSS.
2. Background
Korea Telecom (KT) is Korea’s largest
telecommunicationsservicecompany,which is a semiprivate companywhose stocksare owned partly by the
governmentand partly by the public. KT provides a
broad set of telecommunication services, including
phone and data communication,with its nation-wide
communicationnetwork. In 1994,it had more than six
billion dollars in revenue. It has about 60,000
employeesand has ten district businessoffices,which in
turn are in chargeof 360 local phoneoffices.
Operations& Maintenance(O&M) division is one of
KT’s 12 business divisions. It is the most laborintensive division, employing 30000 of the KT’s 60000
employees. These are the peoplewho work closestto
KT customers. The major O&M tasks are installation
of new phone lines, changeof customerphoneservices,
troubleshootingand restorationof failed phoneservices,
and day to day operationand maintenanceof various
hardwareand softwarecommunicationfacilities.
To maintain high customer service quality, the
O&M division must make continued investments,
spending 15% of the KT’s total annual budget. Here,
the term “investment”meansthe budgetfor the various
O&M activities which is spent to raise the
communication service quality. To make these
investment decisions effectively, performanceof each
district businessoffice and local phone offices needsto
be evalwted. In the past, these performancerelated
data had to be collected and processedmanually from
the local phone offices and submitted to the
headquartersthrough the district businessoffices. At
the headquarters,the data were scatteredover different
division offices. The whole processwas very timeconsumingand the data collectedwere often incomplete,
incorrect, and hard to integrate. As a result, O&M
managementfelt the need to establish a systematic
investmentperformancemanagementprocess,utilizing
an integrateddecisionsupportsystemcapabilitiesbased
on O&M database,network, and various transactions
processingsystems.
We first looked into the traditional decisionsupport
Conference on System Sciences - 1996
system (DSS) and the more recent group decision
supportsystems(GDSS) researchand practices. Both
were found not very well-suited to our purpose. DSS
implementationsto date were mostly for the individual
decision makers and GDSS’s have also been used
typically to support a group of people, gatheredin a
specially equipped decision room or on a local area
network, solving a common problem during a
designatedgroup meeting. Thus the organizationwide
scope and integration requirements of the O&M
decisionsupportenvironmentled us to the development
of an organizationaldecisionsupportsystem(ODSS)for
the O&M division’s investment performance
management process. For the development of the
system,we adoptedthe developmentprocedureof the
Issue-BasedDSS[S] and Business Reengineering[l4]
perspectivefor the more e&ctive implementation.Next
sectionwill provide a brief overviewof what an ODSS
is.
3. Organizational Decision Support System
(ODSS) Overview
3.1 Definition of ODSS
Philippakis[l7] introducedthe term, Organizationwide DecisionSupportSystem,as an integrateddecision
supporttool at the organizationallevel, differentiating it
from the moretraditional decisionsupportsystem(DSS)
to support individuals within a certain organizational
unit. Watson[21] defined ODSS to be a system for
supporting the cross-functional decision process to
achieve organizational goals, utilizing computer and
communicationtechnologies. Swanson[191interpreted
ODSS to be a systemto supportthe distributeddecision
making across the organization and used the term,
DDSS@istributed Decision Support System).
Considering all these definitions, we can say that
organizationaldecision support systems(ODSS)are a
new class of decision support systemsfocusing on the
organization-wideissuesrather than individual, group,
or departmentalissues.
According to George[7] and Aggarwal and
Mirani[l], ODSS deals with a decision processwhich
involvesmultiple organizationalunits and influencesthe
entire organization, necessitating the use of an
integratedcommunicationnetwork. Table 1 compares
DSS, GDSS, and ODDS in terms of their goal, target
decision maker, decision scope,and major technology
components.
63
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
Proceedings of the 29th Annual Hawaii International Conferenceon SystemSciences- 1996
Table1. Comparisonof DSS,GDSS,andODSS
DSS
GDSS
Enhancing group
Improving
individual decision decision making
performance
making
Individual
Group
Decision Maker
Team/department
Personal
Decision Scope
Coordinationtool
Major Technology Modelbase
Decisionroom
Database
Component
I Local network
I Userinterface
System Goal
I
ODSS
Facilitating
organizationaldecision
process
Individual& group
Organization-wide
Integrationtool
Enterprise-wide
comm’nnetwork
I
3.2 ODSS Architecture
There have beenmany kinds of ODSS architecture.
Each architecture has its own characteristicsand
elements.Philippakis and Green [17] categorizedthe
componentsof the ODSS into organization-levelDSS
functions and DSS resources. Here DSS functions
include CPS(CorporatePlanning System)to support
strategicplanning process,EIS(ExecutiveInformation
System)tosupport top management,FDSS(Functional
DSS) to support function-leveldecision making, and
LDSS(Local DSS) to support the FDDSs. DSS
resourceswere human resource,software/hardware,
model/tool,and database.
Watson[21]suggestedan ODSS architecturein the
form of an extendedGDSS. His .
architecture consists of decision
database, expertise database,.
corporatedatabase,public database,
and knowledgebase, all integrated
through a Meeting Support System
operating on a corporate-wide
network.
Miller and Nilakanta[l5]‘s ODSS
architecture differentiate the three
DSS
components (knowledge’
subsystem,databasesubsystem,model subsystem)into
internal/external and private/public dimensions.
Miller[l3] later proposedan ODSS architecturewhere
model subsystem,knowledgesubsystem,and database
subsystemexist in the frameworkof the organizational
information environment,interconnectedthrough the
various communication interfaces. Because of the
diversitiesof ODSSarchitecture,we shouldconsiderthe
applicationdomainas well as the characteristics
of each
ODSSarchitecturewhenwe build an ODSS.
4. Investment Performance Management
@PM) at KT O&M Division
4.1 Reengineered IPM Process
64
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
As Illiaifar and Nilakama[l8] suggested,
thereis an
important relationshipbetween ODSS and business
processes
redesigning. Accordingto the concept,lPM
processwas aualyzd anddiagnosed. Oneof the major
decision making within the O&M division’s IPM
processhas been how to allocate its annual facility
budgetamongthe nation-widedistrict businessoffices
and local phone offices most fairly and effectively.
Here “ftily” means the use of open and objective
criteria in the budgetallocationprocessupon which all
lower level offices can agree. “Effectively”suggests
how to maximize performancegains out of its
investment.
P&m
Evaluation
Figure 1. CurrentO&M BudgetInvestmentAllocation
The currentbudgetallocationprocessis perceivedto
be neither fair nor effective.As seenin figure 1, the
most criticsI problemwas the lack of link betweenthe
investment decision(budget allocation) and the
performanceevaluation.Budget allocation was done
mostly basedon the hardwarefacility status[lO],while
performanceevaluationcenteredon quality of service
@OS) index. Sincefair evaluationof hardwarefacility
statuswas very difficult due to the lack of any agreedupon measurement standards, budget allocation
decisions were frequently subject to internal and
external influences and complaints. On the
petformanceevahtationside, data collectedfor QOS
Proceedings of the 29th Annual Hawaii International Conference on System Sciences - 19%
intensive researchsince 1988 to establish its QOS
framework, which consists of network performance
qualityandoperationsqualityas in Figure3. [2,7,11,12]
Networkperformancequality measuresthe level of
successful transmissionof data(voiceand non-voice)
and connectionthroughoutthe network. Currently,
non-voiceserviceson the phonenetwork include data
transmission through modem, video&, and fax
,ion. Quality of voice servicebreaks down
into transm&ion quality which is afWed
by transmissionloss and line noise and
connectionquality which occurswhen one
subscriir tries to be co~ected with the
index was often incomplete or incorrect for the
following reasons. First, people at the local phone
officesfound it too time consumingto enterQOS data
which will not be reflectedin the budget allocation
processat the headquarter. Second,theywerereluctant
to have their performancejudged solely on QOS data
without any mechanismto accountfor the diBiculty of
their operatingenvironments.
I---+
CD1 1
other subscn’ber.
Figure2. ProposedO&M lPM ProcessModel
When the target processis unstablelike this,
developingan ODSSwithoutIirst stabilizingthe process
may not help much. So, we proposeda new,
“reengineered”
lPM processmodelas in Figure2. The
model consists of O&M budget, O&M operations,
CDI(Context Difficulty Index) and QOS(Quality of
Service)of the local office. The modelconsistsof O&M
budget,O&M operations,CDI(ContextDifficulty Index)
andQOS(Qualityof Service)of the localoffice.
Each of theseconstructswill be explainedin more
detail in the following sections.The majorchangesfrom
Figure 1 are:first, contextdifficulty of eachlocal office’s
operatingenvironmentis formally evahratedas CDI,
second,both CD1 and QOS are incorporatedinto the
performanceevaluation,third, performanceevahration
results become the basis for the budget allocation
decisions.The proposedredesignideaswere congrmed
by the KT managersin chargeof the targetprocesses
for
feasibihty purposes.This reengineeredIPM process
model will be the conceptualframeworkupon which
O&M ODSSwill be built.
Operation quality is measuredthrough
manyserviceitemswhosedataare gathered
from the operationof the phonenetwork.
These items are: rate of billing errors,
averagefaihrrerate,averagerate of normal
repairing, average rate of normal
installation, and normal 114 responserate. Here
normal servicemeansa servicewhich was provided
within the prespe&iedtime limit,
4.3 Context Difficulty Index: CD1
QOS is an important means of evahrating
performance
of the O&M operationsat local and district
offices. Over the years, however,while QOS was
treated as the major criteria in evaluating O&M
performance,analysis of what factors caused or
contributedto such QOS valueswas virtuahy ignored.
This in turn led to wide-spreadconcernand complaints
among the local offices over the fairness of their
performanceevaltion results. To address this
problemand evaluatethe O&M performanceon a more
reasonable
basis,we introducedthe conceptof Context
Difliculty Index (CDI) which will be usedto evahrate
each local or district office’s O&M operating
environment.
CDI, which representsthe difllculty of the O&M
operatingenvironmentconsistsof four components,as
in Figure 4. Level of backboneinstallation has two
items: backbonedeficiency rate and level of no4.2 Quality of Service: QOS
backboneareawherebackbone(mostly madeof copper
cableor fiber optics)refersto portion of the telephone
According to the InternationalTelecommunication line which startsfrom the local phoneoffice and ends
Union whitebook [6], QOS is definedas “the overall where it branchesinto the generalsubscriber’phone
s
intluenceof the serviceperformancewhich determines lines.Hackbonedeficiencyrateindicatesthe relativerate
the setice user’s satisfaction”.For a phone setice
of backbonedeficiencycomparedto its demandat the
company,QOS is an importantindex which indicates area.Level of no-backbone
arearefersto the rateof nothe extentta which its currentphonenetworkmatches backboneareaamongthe entireareaservicedby a local
an ideal network performance. KT has initiated
65
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
Proceedings of the 29th Annual Hawaii International Conference on System Sciences - 1996
Quality of Service
I
I
Network Performance
Quality
Q
Voice service
Quality
Bit error rate
Non-error rate
Exceeding rate
of excessive error
Transmission loss
Line noise
Phase error level
Noise versus signal
Loss pf PAD
I
Delay of sent signal
Delay after dialing
Connection rate
Figure 3. KT QOS framework
inWling and replacing new and old facilities and is
all&
according to KT’s long-term facility planning.
These budgets are first allocated among the district
business offices, where regular budget is allocated
among the local phone offices while facility budget is
sent to and managedby the construction department. At
each local office, the regular budget is divided into
allocations for machine, lines, transmission, and
electricity.
The relationship among QOS, CDI, and O&M
budget is shown in Figure 5. In Figure 5, CD1 and
QOS are evahated once at each local phone office and
secondly at each district business office, comprising the
two-dimensional performance evaluation of O&M
operations. Based on these evaluation results, O&M
management at the headquarters will be able to make
O&M investment decisionsin a fair and effective way.
Performance evaluation of O&M operations can be
done, utilizing the Paretto curve between QOS and CD1
as in Figure 6. Figure 6 shows that QOS and CDX
move in opposite directions from each other. That is,
in general, when CD1 is high (difticult operating
environment), QOS is likely to be low, and vice versa.
Most offices will be fOMd around the Paretto curve. But,
there can be exceptions. For instans, if a phone office
is located in the “A” quadrant of Figure 6, it means they
are delivering the high quality service to their
customers despite the dif&ult operating environments,
implying exceptional management capability. On the
other hand, a phone office found inside the “C” quadrant
away from the Paretto curve signals a potential
management problem since its service quality is low
while it operatesin a favorable environment.
Level of undergrounding cabling
Man power supply status
Facility status
Figure 4. Context Ditsculty Index Components
phone office. Level of underground cabling specifies
the ratio of the underground cabling among the entire
cabling.
Manpower supply status reveals the
over/under stafling of the O&M personnel at a given
local office. Facility status representsthe obsoleteness
of various facilities (switches, lines, etc.) operatedby the
local office. CD1 will be higher if a phone office has
less backbone, less underground cabling, understat&g,
and more obsoletefacilities. More detailed explanation
on CD1 items and their measurementcan be faund in
@ I.
4.4.0&M
Rate of billing error
Average failure rate
Average rate of normal repairing
Avera e rate of normal installation
NormaP 114 reaponse rate
Budget Allocation
O&M budgets are broken into regular and facility
budgets. Regular budget is a budget for maintaining
facilities and is used for electric facility management,
communication network management, line facility
maintenance, etc. Facility budget is a budget for
66
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
Proceedings of the 29th Annual Hawaii International Conferenceon System Sciences- 1996
5. KTOM-ODSS: Organizational Framework
5.1 KTOM-ODSS: Integrated
Network Perspective
O&M Budget
Fi;gure5. RelationshipsbetweenCDI, QOS,andO&M
BudgetAllocation
C
D
IDW
>
HigqPoor)
Low(Qood)
CD1
I
Figure6. RelationshipbetweenQOSand CD1
Basedon this two-dimensionalO&M performance
evahration,fair and effective O&M budget allocation
now seemspossible. First, for offices locatedin “A”
area,allocateextra budgetson the first priority so that
their operatingconditionscan improve,thus they can
move to area “B.” This may createdoubleeffect of
rewarding“A” officesas well as stimulating“D” off&s
to improve their QOS so that they can also operate
under better conditions. Second,offices in the “D”
quadrant will receive extra budgets on the second
priority so that they can movetoward“B” quadrantover
time. Third, officesin the “B” quadrantmay continue
to receivethe currentlevel of budgets. Finally, offices
in the “C” quadrantfirst need to receive a special
diagnosisof its managementprocessto identify the
cause of low service quality despite the favorable
operatingenvironment.
67
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
To
implement the
reengineered
O&M lPM process
in the most systematic and
effective way, we decided to
develop an ODSS, named
KTOM-ODSS. In Figure 7,
KTOM-ODSS is shown to
interactwith manyKT business
processesand their supporting
systems horizontally while
vertically integrating with the
lower level O&M transaction
processing systems such as
OMAS(Operationsand Maintenance Administration
NPMS (Network Performance
System) and
Management System). On the horizontal side,
interactiontakesplacein the form of dataexchangeover
For instance, human resource
the network.
provide
management
andfacility management
processes
CD1datato KTOM-ODSS,while KTOM-ODSSprovide
cr~ial evaluation data to task managementand
investmentdatato accountingmanagement.
On the vertical integration side, all the O&M
managementdata at district businessoffices and local
phone offices are entered into the O&M business
databasethroughthe OMAS serverlocatedat each of
the ten district business offices. OMAS server
performsdata file receptionfrom local phone offices,
databasemanagementat the district level, creationanalysis-evaluation
of QOS data for each local office.
NPMS is a system to measurethe connectionand
transmissionquality of a phone network. NPMS
providesQOS’snetwork performancequality data for
input into the O&M database.
5.2 KTOM-ODSS: Architectural Perspective
Before preparingKTOM-ODSS architecture,we
investigatedthe characteristicsof O&M process.The
main characteristicsof KT O&M processare processhierarchy(headquarter,
district businessoffice,telephone
office) and horizontal interrelationshipswith other
internal processes(task management, employee
management,
accountingmanagement,
etc.).
Proceedings of the 29th Annual Hawaii International Conference on System Sciences - 1996
offices through the network. One
of the core components of the
KTOM-ODSS is the investment
performance e&ration
model
(QOS/CDI), and OMAS and NPMS
correspond to the software/
hardware DSS resources. Lastly,
all the O&M personnel involved in
the data entry, hardware/software
operation
and
maintenance,
database
and
network
administration, and final use of the
ODSS models will comprise the
human resource category of DSS
resource.
6. KTOM-ODSS: Prototype
Implementation
. . . . .
I
1
Prototype of KTOM-ODSS was
developed on a 486 PC under the Windows 3.1
environment, using the FOCUS[4] and FOCUS/EIS[S]
Execution of the prototype requires the
tools.
minimum memory of 8 MB. FOCUS language was
selected because of its non-procedural nature and to
facilitate the integration with the KT O&M division’s
main trausaction processingsystem (OMAS) which was
developedin FOCUS. KTOM-ODSS main system has
four modules at the top level: target selection,
investment support, performance evahration, and data
management(Figure 9).
Figure 7. KTOM-ODSS : Integrated Network
Perspective
Considering these aspects, we selected and
developed KTOM-ODSS architecture in Figure 8
following the ODSS architecture of Philippakis and
Green[171. For their organization-wide planning or
strategy development at the highest DSS function level,
KTOM-ODSS has O&M strategy Development in the
area of investment performance management At the
second DSS function level, Philippakis and Green
suggest an Executive Information System (EIS) to
analyze and evaluate the r
target business process. In
KTOM-ODSS, we have
O&M
investment
performance evahtation and
management
functions.
At the third level, they
utilize the functional and
local DSSs to support the
within-department business
DatabaseManagement System
tasks. KTOM-ODSS has a
data management system for
this level to manage data
organizational DSS Function
transfer, data tra&ormation,
and data retrieval through
the
organization-wide I
Figure 8. KTOM-ODSS : Archi~
Perspective
integrated network.
Organizational DSS resources involve dambase,
Target selection module supports the selection of a
model/tool, software/hardware, and human resource
district
or local office by name or list search.
[17]. In KTOM-ODSS, we have O&M databasewhich
Investment
support module containshistorical and
collects and stores O&M data from district and local
68
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
Proceedings of the 29th Annual Hawaii International
Conference on System Sciences - 1996
integrated with its
sm-roullding
transactionprocessing
systemsin the future,
it will becomea very
Maill
powerful model case
System
of an ODSS which
will afkct the entire
phone
nation’s
network operations
and maintenance.
The very natureof
ITS
a phone service
Coonection
company providing
serviw
to
My
Figure9. SystemConfiguration
customerco~ected to its phonenetwork seemsto fit
nicely with the ODSSphilosophyof providing decision
currentinformationon the regularandfacility budgetof
supportto the decisionmakersat headquarterand each
each local and district office. It will be extendedto
district business office co~ected to the integrated
include the performance forecasting function that
corporate network. Applying the “reengineering”
enablesusersto forecastperformance(QOS
and CDI) for
principleto our targetdecisionprocesswasvery crucial.
a given investment(regular
and facility budget)amount Without it, we may have createdan ODSS which will
and total portfolio. Performanceevaluation module supportthe “unfair”and “ine&ctive” decisionprocessto
calculatesQOSand CD1dataof everyoffice and display run more“efficiently.”
their statistical distribution in both text and graphic
After reengineering
the decisionprocess,we selected
forms along with a budgetallocationstrategy.Figure 10
an ODSS architecturesuggestedby Philippakis[l7]
SllOWS performance
of O & M works at the
consideringboth the characteristics
of KT O & M process
‘GwanghwamoonOflice in Seoul Business Area’.
and the architecturesuggested.But the architecture
neededsome modificationsfor real implementation.
Gwanghwamoonoffice got 93.02% QOS point and 4
CD1 level. The scoredistributionrevealsthat the office
First, the strategy-planningcomponent was not
systematizedbecause the strategy planning is a
obtained high QOS point under poor O & M
environment.And, if the ‘Action’ menu is selected, management
policy issuewhich was beyondthe O & M
anotherscreen(f;igure11) will display QOS and CD1
scope.Second,we considerednot only a centralized
distribution with Gwanghwamoonoffice’s status and
database
but also a distributeddatabase(O&Mdatabase,
suggestspecific investmentaction. Data management each managementrelated database)that will be
module interacts with the various transactions interfaced through the network. Finally, although
Philippakisseparated
DSS levelsinto FDSSand LDSS,
processingsystemsto retrieve,convert,and trausmitthe
DSS levels in KTOM-ODSS were integrated and
O & M investmentand performancerelateddatafrom the
otherbusinessdepartments
andlocal anddistrict offices.
databasemanagementfunctionality was emphasized.
These modificationswere made to reflect KT O & M
7. Conclusion and Future Directions
processmoreadequately.
For the future extensionof our KTOM-ODSS,three
Researchand practical implementationsof ODSS directions seem promising. First, develop diverse
conceptsare not as matureor activeas thoseof DSS or
decision models (e.g., performanceforecasting)and
GDSS. But, with more and more organizations knowledge-basedcomponentsto make the system
integratingtheir departmentalized
informationsystems behave more “intelligently”. Second, refine the
into an organization-wide
onebasedon the sophisticated definition and operationalizationof QOS and CD1 by
communicationnetwork,the importanceof ODSS will
empiricallyvalidatingthem on a larger scalethan in our
continue to grow. This paper reports on the
study. Third, establishthe actual network connection
implementationof an ODSS prototypefor the Korea between the KTOM-ODSS and its surrounding
Telecomcompany. As a stand-aloneprototypesystem transactionprocessingsystemsandotherKT databases.
itself, it is fairly small and simple. But when it is fully
69
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
Proceedings of the 29th Annual Hawaii International
Conference on System Sciences -
Figure10.Performance
Evaluation
Figure 11. StatusandAction aboutQOSand CD1
70
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE
1996
Proceedings of the 29th Annual Hawaii International
Conference on System Sciences - 1996
systems,”Decision SupportSystems,vol. 9, 1993, pp.
201- 215.
[ll Ag8arwa.land RajeshM irani, “Macro Issuesin the
Development of Grganizational Decision Support
Systems,”Proceedingsof the 28th Ammal Hawaii
InternationalConferenceon SystemSciences,vol III,
1995,pp. 917 - 926.
[14] M ichael Hammer, “ReengineeringWork: Don’t
Automate,Obliterate,”HarvardBusinessReview,JulyAuguat,1990.
[15] M iller, L.L., and Nilakanta, S., “Design of
organizationaldecisionsupport systems:the use of a
data extraction scheme to facilitate model-database
commmkation,” Proceedingsof the 24th Ammal
Hawaii InternationalConferenceon SystemSciences,
vol. 4, 1991,pp. 65 - 72.
[2] ETRI, A Study of the Integrated Strategic Investment
Management, 1993.
[3] Hisham El Sherif and Omar A. El Savvy,“IssueBased Decision Support Systems for the Egyptian
Cabinet,”M IS Quarterly,vol. 12, no. 4, 1989,pp 551 -
[16] Pagani, M . and A. Belluci, “An Organizational
decisionSupportSystemfor teletra’sTop Management,”
In Lee, RM., A.m. McCosh and P. M igliarese(Eds),
570.
[4] InformationBuilders,FOCUSfor Windows, 1993.
Organizational Decision
[S] Information Builders, FOCUSEIS for Windows
Release 3.3, 1993.
Support Systems, North-
Holland,1988,pp. 3 - 13.
[17] Philippakis,A.S. and G.I. Green,“An Architecture
for Organization-WideDecision Support Systems,”
Proceedingsof Ninth International Conferenceon
InformationSystems,1988,pp. 257 - 263.
[6] ITUT(International TelecommunicationUnion)-T(Telecommunication),
white Book, 1992.
[7] Joey F. George, “The Conceptualizationand
Development of Organizational Decision Support
Systems,”Journalof ManagementInformationSystems,
vol. 8, no. 3, 1991- 1992,pp. 109- 125.
[8] KAIST, Building an Investment-Performance
Management DSS for supporting KT O&M Strategy
Development, ProjectreportRL4DOO1,1994.
12.
[18] Sam Illiai&, SreeNilakanta, and G.M. Prabhu,
“TechnologyImperativesof BPR and their Effect on
OrganizationalDecisionSupport,”Proceedingds
of the
28th Ammal Hawaii International Conference on
SystemsSciences,
vol. 4, 1995,pp. 941 - 946.
[19] Swanson,E.B., “Distributed decision support
systems: a pempective,”Proceedingdsof the 23rd
Ammal Hawaii InternationalConferenceon Systems
Sciences,
vol. 3,1990,pp. 129- 136.
[9] KoreaTelecom,OMAYL User’s Manual, 1994.
[lo] KoreaTelecom, O&M Budget Allocation Manual,
1994.
[20] Walker, W.E., “Differencesbetweenbuilding a
traditional DSS and an ODSS: lessonsfrom the Air
Force’s Enlisted Force Management System,”
Proceedingsof the 23rd Annual Hawaii International
Conferenceon SystemsSciences,vol. 3, 1990,pp. 120128.
[l l] Korea Telccom ResearchLab., A Study on
Network Service Quality, 1993.
[12] KTA ResearchCenter,A Study on the Economy
Analysis Model for QOS management and the Planning
for the Implementation of Integrated Operation and
Maintenance Network, 1988 - 1989,
[21] Watson,RT. “A designfor an i&astructure to
supportorganizationaldecision-making,”
Proceedings
of
the 23rdAmmal HawaiiInternationalConferenceon
SystemsSciences,
vol. 3, 1990,pp. 137- 142.
[13] Leslie L. M iller, “Organizationaldecisionsupport
71
Proceedings of the 1996 Hawaii International Conference on System Sciences (HICSS-29)
1060-3425/96 $10.00 © 1996 IEEE