business modeling

316 FUJITSU Sci. Tech. J., 42,3,p.316-322(July 2006)

Use of Business Modeling in RequirementsDefinition Phase

V Naritoshi Tsugane V Takeo Asakura(Manuscript received April 25, 2006)

For many enterprises, how to anticipate various changes in the business environmentand promptly respond to those changes are important issues. However, in manycases, the IT systems that support the basis of business are extremely complicatedand difficult to respond to changes. To address these problems, Fujitsu recentlyreviewed its SDAS system development methodology from the requirements defini-tion phase by reconsidering its original concept. First, enterprise businesses andrelated IT systems can be correctly visualized and commonly understood by everyoneby improving a few key modeling techniques for the requirements definition phase(e.g., business process modeling and business data modeling) based on our previouspractical experience. Secondly, the total optimization of systems by taking newapproaches such as Service-Oriented Architecture (SOA) becomes possible byproviding the means to redesign complex IT systems into flexible, mutually separatedones. This paper describes the concept of modeling and then examines businessprocess modeling and business data modeling. Lastly, this paper describes how thesebusiness modeling techniques are used to realize SOA systems.

1. IntroductionIs your company free from these problems?

• Although a business flow was created for aprevious development project, it cannot beused because it has not been appropriatelymaintained.

• Specifications cannot be defined because thepeople who are familiar with the business aretoo busy.

• Only one person knows about the system(i.e., the system is personalized).

• Furthermore, the people who know a lotabout the system are nearing retirement (the“baby-boomer generation” problem).

• Although your company has decided how tomodify its business, little is known aboutwhere and how the system must be modifiedand what influences the modifications willhave.

It is now clear, as suggested in the paper,“Toward Realization of Service-Oriented Architec-ture (SOA),”1) elsewhere in this special edition,that one of the most important issues for acompany is to quickly identify and respond tochanges in the business environment and that itis very important for IT systems to be able to flex-ibly respond to changes by using SOA concepts.As discussed in the SOA paper, the introductionof a service bus alone is insufficient to make acompany capable of flexibly responding to chang-es. After all, the key points are to understandwhat is required to flexibly respond to changesand to design a system that can flexibly respondto changes. Changes occur not in a system, but ina business. To become capable of quickly respond-ing to changes, it is necessary to visualize thefundamental structure and behavior of the busi-ness in which the changes occur, quickly identify

317FUJITSU Sci. Tech. J., 42,3,(July 2006)

N. Tsugane et al.: Use of Business Modeling in Requirements Definition Phase

the changes, and quickly adapt the system to thechanges. These objectives cannot be achievedwithout a business model that accurately mapsthe reality of the business. In addition, the mod-el must be implemented in a system using alanguage the system understands. In other words,the system must closely represent the businessworld, and the relationship between the businessworld and the computer world must be carefullymaintained. The SOA development project gaveFujitsu an opportunity to recognize again the im-portance of upper processes, and we thereforestarted reviewing them. We believe the followingthree points (Figure 1) are especially importantin the requirements definition phase:• Reach a consensus among the stakeholders.• Visualize and model the business, and run

plan-do-check-act (PDCA) cycles to improvethe business (business process modeling).

• Construct the system so it closely reflects thereal world (business data modeling) in an ar-chitecture that is responsive to changes(service architecture modeling).Although these three points are all impor-

tant in the requirements definition phase, thispaper focuses on the last two business modelingpoints. It outlines the concept of business model-ing and then discusses business process modelingand business data modeling. Lastly, this paperdescribes how these business modeling techniques

are used to realize SOA systems.

2. Concept of business modelingA business consists of two elements: proce-

dures (processes) and information (data). Toreflect this in a system, business modeling re-quires the two perspectives of business processmodeling and business data modeling.

Figure 2 outlines the concept of businessmodeling. The business process model in thefigure shows that order acceptance starts whenan order arrives and that shipment starts afterorder processing ends. An order has two sections(header and details), there are two types of order-ing, and multiple products can be ordered together(details section).

The figure shows that the essence of busi-ness modeling is to clarify the business flow(procedures) and the structure of the informationto be handled.

3. Business process modelingBusiness process modeling is a technique

that visualizes business in the real world, mainlyusing a business flow chart. Sharing existing busi-ness knowledge and information about futurechanges in a business helps to improve business.

3.1 Problems with conventional businessprocess modelingAlthough business process modeling depends

mainly on a business flow chart, notations are notstandardized. In fact, when we collected businessflow chart examples, we found that their infor-mation was inconsistent. In addition, because theanalysis items to be investigated and the steps tobe taken were not defined, the modeling resultstended to be personalized. This sometimes ledto failure in reaching a consensus with thecustomer.

Another issue is that a business problem thatis already known can be addressed at the time ofmodeling, whereas an unknown problem cannotbe addressed if the problem handling method is

Management/business [Key activity]

Consulting

Business process modeling

Business data modeling

Service architecture modeling

Planning

Business

Business improvement

modeling

IT system

IT improvementmodeling

DesignConstruction

Operation

Figure 1Three points in upper processes.

318 FUJITSU Sci. Tech. J., 42,3,(July 2006)


limited to modeling and does not include contin-ued monitoring using PDCA cycles. Unknownproblems are detected, for example, when thereis a sudden workload increase.

3.2 Features of SDAS-provided businessprocess modelingTo handle these issues, we developed a new

business process modeling technique calledValuevision Modeling Methodology (VMM).

The features of this technique include thefollowing:1) It uses Unified Modeling Language (UML)

so that notations can be standardized.2) It uses UML profile extensions to improve

the readability of chart symbols, the appear-ances of which are shown using extensionsso business and information types can be in-tuitively distinguished.

3) It predefines the basic analysis items, forexample, the time required to complete a pro-cess and the number of backlogged slips to

be processed.4) Its modeling procedures are standardized.5) It clarifies the points of modeling

(e.g., business activities are hierarchicallyrepresented).

6) It provides a modeling tool (in-house useonly).These features allow us to construct a busi-

ness process model that contains the informationrequired to reach a shared recognition and mini-mize the level of personalization.

Figure 3 shows a model that was construct-ed using this business process modeling technique.This model was constructed by observing thepredefined rules (e.g., the business must berepresented in a hierarchical structure andthe sections in charge of the business must beclarified).

We also note the importance of continuedefforts to find potential business problems topromote non-stop improvement. For this purpose,we provide a middleware package that combines

Order

Orderdetails

E-mail

Order reception: When an order is received from a customer via facsimile or e-mail, the order data is registered in the system via a terminal.

OrderCustomer: A (Ota-ku, Tokyo)1. Choc-bean: 102. Gum-ball: 53. Candy-ball: 20

What’sthat?

Modeling can simplify things and eliminate ambiguity.

Business = information + procedure A business is defined by clarifying 1) the information used for the business and 2) the procedures used to handle the information.

Each ordercontains at

least one linefor details.

FacsimileThere are two

ordering methods:facsimile and e-mail.

Orderacceptance

Shipment

Order

Facsimile ore-mail

1

Information modeling = business data modeling Procedure modeling = business process modeling

Figure 2Concept of business modeling.



monitoring technology with modeling techniques.This package can measure rating factors such asthe workload and time required for businessactivities and visualize them on a business flowchart (Figure 4). This function can visualize busi-ness problems in a more specific representation,allowing us to identify changes in the businessenvironment in a timely manner and respond tothem quickly. This package, therefore, providesimportant monitoring technologies for ensuringPDCA cycles work as required for businessimprovement. The area to which this monitoringtechnology belongs is called Business ProcessManagement (BPM).2) This area is attractingmuch attention in the industry because it isrequired for enterprise internal control.

4. Business data modelingBusiness data modeling visualizes entities in

the real world, mainly by using a class diagram.By focusing on these entities to identify the

<<System-assisted>>Order form entry

<<System-assisted>>Purchase order form entry

<<System-assisted>>Arrival confirmation

<<Human action>>Acceptance & Send-out

<<System-assisted>>Product stock inquiry

Order information input screen

Stock informationconfirmation screen

<<System>>Work instructions

Business process

Sales department Shipping department (warehouse)

Business flow

Business data

Stock insufficient

Purchase order input screen

Shipment information input screen

Shipment

Terminal entry 1

Delivery form Shipment form

Purchase order form

Relationshipwith

businessprocess

Supplier

Details (outline, activation conditions, required time, cost, processing amount, etc.) are confirmed.

Business process flow

Stock sufficient

Figure 3Image of business process modeling.

Data collected through monitoringRunning system

A0010-01 Request Sept. 1A0010-01 Reply Sept. 2A0010-02 Request Sept. 1....

Quotationrequest service

Quotationreply

service

...

Figure 4Visualization with monitor tool.

essential structure of a business, we canconstruct a change-responsive IT system in whichessential and non-essential elements are clearlydistinguished.



4.1 Problems involved in conventionaldata modelingData modeling lies at the heart of the Data

Oriented Approach (DOA), which assumes thatinformation is the essence of a company. Compa-nies typically handle a huge amount of complexinformation. With DOA, a single attempt is madeto faithfully map this complex world onto a mod-el. Therefore, the IT systems constructed fromthese models tend to be as complex as the models.

Another problem with this conventional tech-nique is that it is time-consuming and requires alarge amount of human resources because itattempts to include all information (data) handledby the company in the model.

4.2 Features of SDAS-provided businessdata modelingTo overcome the problems mentioned in the

previous section, we developed a new businessdata modeling technique that is based on the con-cepts of kaname entities and events.3)

One of the features of this technique is thatdata modeling is divided into two stages: the busi-ness analysis phase and the system design phase.The business analysis phase focuses on KanameAnalysis to grasp the essence of the business. Inthe system design phase, the business ismodeled based on the results of Kaname Analysisusing conventional DOA with an emphasis oncomprehensiveness.

Kaname Analysis provides a simple modelthat enables the essence of real-world situationsto be easily grasped. It does this by focusing onreal-world kaname entities and events, which aredefined below.1) Kaname entities

Kaname entities are things of interest (e.g.,people, goods, and money) that are connected witha company activity (business) and whose state canchange.2) Kaname events

Kaname events are activities that trigger astate transition of a kaname entity.

Because Kaname Analysis classifies all enti-ties and all events as being essential ornon-essential, the framework of an IT system canbe well defined. This allows us to design a change-responsive IT system without spending much timeor human resources.

Figure 5 indicates the differences betweenconventional DOA and Kaname Analysis.

Figure 6 shows the flow of Kaname Analysis.In this new business data modeling tech-

nique, we first represent the structure of theinformation to be handled for the business as adata model using a UML class diagram or similarmeans. During the modeling, we focus only onthe kaname entities at the core of the business.Next, we analyze the relationships among theentities to confirm they have been correctlyextracted. Then, we represent the state changesthat could occur for each entity as an eventsequence model. These two models represent theessence of the business in the simplest way.

After completing the above steps, we can findthe points where the relationships betweenkaname events and entities are the loosest. Thisenables us to identify the segmented units thatcompose an IT system and then make the systemcapable of flexibly responding to changes. Thefunctions provided by the segmented system arecalled services in the SOA world (describedlater).

5. Relationship betweenbusiness modeling and SOAFujitsu announced the establishment of its

system of SOA technologies in July 2005 underthe slogan of “Build a change-resistant IT sys-tem.” In Japan, the system provides an SOAdevelopment technique called SDAS/ServiceModeling. This technique combines business pro-cess modeling and business data modeling(described in other papers in this special issue)with the service modeling described in the SOApaper. Fujitsu provides industry-specific SOAdevelopment solutions in Japan that are based on



Stock confirmation

Purchase order

Order confirmation

Service bus

Quote Logistics Payment

Purchase order and stock confirmation

Quote/orderLogistics

Payment

Real world(complex)

Conventional technique

Faithful mappingusing ER model, etc.

DOA-handled data

Kaname Analysis

Kaname Analysis

Real world(kaname entities/events)

Mapping the essence represented by kaname

entities/events

Kaname world

Loosely arranged IT system centered around kaname entity/eventChaotic IT system with

unclear boundaries

Service X Service Y Service M

Service bus

Service

Order Shipment

Shipment

Service N

Order

Order

Product

Product

Customer

Customer

Orderacceptance

....

....

Kaname event

Kaname event

1

11

Kaname entity

Trigger of entitystate transition

Figure 5Difference between conventional DOA and Kaname Analysis.

Figure 6Image of Kaname Analysis.



this SOA development technique and includeindustry-specific know-how. More specifically,Fujitsu’s SOA development solutions includeValuevision, a solution for industy and distribu-tion focusing on business and processimprovement; FUTURITY, a solution for socialinfrastructures based on reviewed images of howsystems should be; and EVOLUO, a solution forfinancial institutions focusing on business tie-upsthat satisfy needs related to industry reorganiza-tions and increasingly diversified distributionchannels.

The purpose of SOA is to construct an ITsystem architecture that allows IT systems to flex-ibly and quickly respond to changes in business.Before a flexible IT system can be constructed,business modeling is required. To conduct busi-ness modeling appropriately, it is necessary torecognize again the importance of upper process-es. SOA truly provides an opportunity for thisrecognition.

6. ConclusionThis paper described business modeling,

which is an important technique used in phaseswhere requirements are defined through discus-sions with the customer.

The need for business modeling is increas-ing year by year in many areas, for example, SOAand EA (total optimization planning). We willmake further efforts to improve our techniques,educate and train modeling staff, and accumulatemodeling results and make them available asreference information.

References1) I. Morita: Toward Realization of Service-Orient-

ed Architecture (SOA). Fujitsu Sci. Tech. J., 42,3, p.306-315 (2006).

2) Excellences of “BPM Infrastructure.” (in Japa-nese), Fujitsu Journal, June 2004.http://jp.fujitsu.com/about/journal/272/newtechnology/02.html

3) Z. Nakamura: A Planning Method of SuccessfulSimple Work by Kaname Entity/Event Analysis.(in Japanese), The Nikkan Kogyo Shimbun, Ltd.,2003.

Naritoshi Tsugane, Fujitsu Ltd.Mr. Tsugane graduated from AichiCommercial High School, Japan in1983. He joined Fujitsu Ltd., Tokyo,Japan in 1983, where he has beenengaged in development of technologyfor business requirement analysis andsystem requirement analysis.

E-mail: [email protected]

Takeo Asakura, Fujitsu Ltd.Mr. Asakura received the B.S. and M.S.degrees in Informatics from ShizuokaUniversity, Japan in 2002 and 2004,respectively. He joined Fujitsu Ltd.,Tokyo, Japan in 2004, where he hasbeen engaged in development oftechnology for business requirementanalysis and system requirementanalysis.

E-mail: [email protected]

business modeling

Documents