An Enterprise Architectures Implementation Methodology

Published Date: 02 Nov 2017

Based on Agent-Oriented Technology

Introduction

Enterprise Architecture (EA) was developed to align Information Technology (IT) and business, and to reduce complexity of Information System (IS) within enterprise. EA causes that enterprise has the integrated environment in business and IT. To address those purposes EA presents three stages, As-Is architecture, To-Be architecture, and transition plan. In As-Is architecture, EA will be defined current situation of business and IT of enterprise by means of set of definitions which illustrate the current state of the enterprise's mission, business processes and technology's infrastructure. The key role of this stage is vision of enterprise. In To-Be architecture EA will be represented the desired architecture including future of business and IT based on vision of enterprise. This type of architecture is the result of enterprise's long-term strategies and plans. The key role of this stage is to identify appropriate ISs. In EA, Transition plan is the essential strategy that will be employed for migration from the As-Is (current or baseline) to the To-Be (desired) one. The key role of this stage is using the proper implementation method. In this definition, an enterprise may be a company, an institution, or a department within a company. Within EA the elements to be described may be functions, data, network equipment, software components, business locations, human roles, performance indicators, etc [28][22][16]. A good architectural practice helps a company innovate and change by providing both stability and flexibility.

Today’s enterprises are known that the efficient management and utilization of information through IT is a critical factor to business success, and an essential means to achieving competitive advantage. EA addresses this need, by providing a strategic context for the evolution of the IT system in response to the constantly changing needs of the business environment. Furthermore, a good EA enables enterprise to achieve the right balance between IT efficiency and business innovation. It allows individual business units to innovate safely in their pursuit of competitive advantage. Moreover, other advantages of utilizing EA are categorized within following types:

Business: It can be increased efficiency of business operation due to sharing business across the enterprise and improving business productivity.

IT: It can be increased efficiency of IT operation due to improving applications interoperability, developing appropriate software, and upgrading and exchanging application components

Risk: It can be reduced risk due to declining complexity of business and IT, and faster returning current investment.

EA is comprises from theories, methodologies, technologies, and practical approaches which they are supported design, build, and EA management. [7][24][36][35]. In EA, Framework is a logical structure for the classification and organization of the various descriptions of an organization. EA frameworks are given standards for explanation of architecture. An EA methodology provides method for taking diverse aspects of business and transforms them into models while framework provides a structure within which these models can systematically be placed. EA methods define the development activities in an EA project [32] [28][8][27]. Existing EA methodologies are usually presented in two parts: a framework and a method. Usually, the frameworks are quite sophisticated and the methods are rather simple. The frameworks, a term widely used in EA, provide guidelines on how to make the enterprise model [29][32][40]. EA implementation methodology is the process of defining architecture for the use of information in support of the business and the plan for implementing them.

Agent-oriented techniques represent an exciting new means of analyzing, designing and building complex software systems. They have the potential to significantly improve current practice in software engineering and to extend the range of applications that can feasibly be tackled. They have high level attributes. These attributes include independence, rational planning, intelligence, context level interaction, adaptability, and reactivity [43][46].

Background of the problem:

There are several problems that we identified in this research.

Most EA frameworks do not propose a method to implement the EA. It is obvious that no matter how complete and well-defined a framework is, it would be inappropriate without having a plan for implementing it.

Enterprise architecture is complex because it involves different types of practitioners with different goals and practices during the lifecycle of the required changes. Most frameworks and methodologies tried to simplify understanding of EA by means of several tactics and techniques, but it still is complex which means that it needs more activities and guidelines.

Enterprises need business and Information Technology (IT) integration as basic factor to achieve their competitiveness. Unfortunately, largely of the IT projects face to fail respect to they could not achieving this union. For addressing this issue EA goes beyond IT. It addresses business implementation by integration business processes of the IT resources (e.g. applications, networks, and other) and of the business resources (e.g. facilities, people, and other). Unfortunately, according to the analysts’ forecasts, the success rate of the EA projects is not much higher than that of the IT projects [24][31][33] due to they have problem in implementation phase.

Existing EA methodologies are greatly developed based on the architect's experiences which they have gained from multiple projects. The experience and the best practices are caught by means of patterns that are reused from project to project. Thus there are not appropriate theoretical fundamentals behind them.

Some of the existence EA methodologies focus on architectural artifacts such as, specific documents, reports, analysis, models, or other tangible artifacts that contributes to an architectural description for increasing general understanding of enterprise systems [32][42][35][36][37]. There are lacks of consideration for implementation phase.

All famous frameworks (such as: DODAF, FEAF, TOGAF, EAP, and others) which have particular method for implementing had several revisions since they are introduced.

None of existing EA implementation methodologies are complete due to they do not deal with all topics, such as planning, selecting tools, and estimating cost.

Most of existing EA developing methods (TEAF, FEAF, and others) only can be utilized in particular framework. So they can not useful for all enterprises due to those frameworks were developed for specific enterprises.

Although some EA developing methods like TOGAF (ADM) are comprehensive and represented better plan rather than others for implementing EA, they accepted that their owned methods are technically complex processes which they need more time and cost with involving many specialists.

Below table (Table 1) illustrate comparison of four famous (FEAF, TEAF, DODAF, and TOGAF) frameworks based on implementation methodology criterion.

Table1 – Comparing EA Frameworks [18][20][21][22][24][33]

FEAF

DODAF

TOGAF

TEAF

Framework Attitude

Matrix Like Zachman consist of 3 columns:

Data, Application and Technology

Functional, Technical and Systematic attitude

4 Layers of Attitude:

Business, Data, Application, Technology

Matrix Like Zachman consist of 4 columns:

Function, Information, Structure and infrastructure

Implementation Methodology

Spewak

6 Step without mentioned in detail

Use ADM architecture

Using Role, Responsibility and EA Concept

Problem Statement:

The problem that we would like to address in this research is: there is lack of EA implementation methodology that uses to support Enterprise Architecture implementation. The benefits of proposed methodology include:

Explicit guidelines for each level of implementing (Design, Analyze, and others )

Providing appropriate foundation for concepts and principles of implementing EA

Supporting EA implementation by way of concentration on developing apt application

Supporting EA project against future changes by means of agents automatic adaptability

We took it as specific research problem. Our research work is to develop Enterprise Architecture implementation Methodology by means of Agent-Oriented technology; we call this Methodology as EAIM.

Research Questions

Which factors can affect on EA implementation methodology?

What are the characteristics of the existing EA implementation methodology?

What are the success factors and obstacles the existing EA implementation methodologies?

How to develop EA implementation methodology based on factors and characteristics?

How to evaluate proposed EA implementation methodology?

Objective of the Research

To investigate factors that affect EA implementation methodology

To evaluate the existing EA implementation methodologies

To identify the existing EA implementation methodology success factors and obstacles

To develop an EA implementation Methodology

To evaluate proposed EA implementation methodology

Research Methodology:

6.1 Research Design

The flow of activities is described in Figure 2 which classifies the research work in four phases.

Phase 1-

Literature

Phase 2-

Method Development

Phase 3-

Collecting Data

Phase 4-

Analysis

Phase 5-

Result and Finding

Problem Statement

Literature Review and background

Analysis method

Conclusion and Future work

Use Case

Method development

Assessment

Survey

Method validation

Figure 2 – Research design

Phase

Objective

Method

Deliverables

1

To collect needed data for literature

To define research plan

To analyse collected data and restate problem statement clearly

Academic data-base

survey

Survey

2

To develop proposed methodology

To test proposed methodology

3

To gather data from use case, survey and questionnaire

Use Case, Survey

4

To evaluate for proposed methodology

Questionnaire

SPSS, Excel

Validate the proposed methodology, Analyzed data

5

To investigate the objectives and goals

To conclude research activities

Outcomes and results

Table3- research design objectives

6.2 Research method

According to this research tendency there would be several accepted research methods that can be utilized.

Action Research

Action Research (AR) will be employed slightly for development phase (Research design). According to action method researcher must develop, test and amend. Thus, using this method within the proposed work development phase is more benefit [9].

Conceptual development

Conceptual development is part of the constructive research methodology type. It will be employed in the current research in order to enable the development of the formalizations and the meta-methodology [12].

Descriptive / Interpretive research

This method has been selected for use in the critical literature review. The main reason for this choice is that it will allow the researcher to develop a cumulative knowledge of the enterprise architecture implementation problems and will ensure that the current research is relevant and builds on previous achievements [11].

Case Study

Case study is a very suitable method for this research. It could provide a detailed description of the as-is state (and relevant past events) of the organization involved in the field validation phase. It will be employed largely among of this research and play the main role [12][10].

Survey

The survey method has been utilized due to the literature review. A survey of the major enterprise architecture frameworks / methodologies and agent-oriented architecture has been perceived as potentially revealing [15][14].

6.3 data collecting method

Primary Data

Data gathered by the researcher (Primary data) use is desirable and this research will be used it.

Questionnaire: it will be used in validation phase.

Participant observation: it will be used in development and validation phases.

Secondary Data

Use of secondary data for purposes has its critics and its defenders. Secondary data is becoming more available by use of data warehousing, public databases, and automatically collected traces on organizations and their employees behavior.

The approach adopted in this research project is that secondary data originating from a reliable source may be used, observing some precautions such as the purpose of, and the methods used in the original data collection.

6.4 Research Tools

This section presents some of the tools that needed for research design, or specialized qualitative, quantitative or mixed data analysis. The research tools chosen for the current research must support the research methods actually used. Therefore, final choices for research tools will have to be made as the research progresses.

Research Design

Methodologist’s Toolchest by IdeaWorks is a research design tool that may be employed in an early phase. It is an expert system providing research strategies, scheduling and budgeting.

Qualitative

NUD*IST (NVIVO) allows management of ’Non-numerical Unstructured Data’ using indexing, searching and theorizing. Atlas.ti and HyperResearch provide visual qualitative analysis of textual, graphical or audio/video data. This kind of tool may provide additional insights on the qualitative data gathered within the current research.

Quantitative

SphinxSurvey provides for the design, administration, processing and analysis of surveys. SphinxSurvey is not likely to be used. GBStat supports statistic research methods. SPSS and Excel are another software which are employed in order that analyzing Quantitative data.

Mixed

C-I-Said and Code-A-Text are examples of software providing qualitative and quantitative analysis of text, charts, audio/video etc. Their dual approach to data analysis may readily enable the use of triangulation.

Modeling

The conceptual development in development phase of this research must be created by suitable modeling tools towards building and testing models. Tools will be selected by considering some factors such as:

- Suitability for the particular research;

- Interoperability / compatibility (e.g. possibility to import / export models);

- Availability of these tools for the researcher.

In this research System Architecture, Enterprise Architecture, and Oracle case will be employed largely.

Significance of Research

Successful implementation of EA is a vital section due to the importance and major role of enterprise architecture in developing an enterprise. Thus proposed methodology which it will be developed for describing how to implementation EA can be used as the best reference for those enterprises which they want to utilize EA.

On the other hand, based on this research's results, agent-oriented enterprise architecture is introduced, which also extends agent-oriented architecture usage in other important fields of information technology.

Research scope:

The concentration of the research is more focused on making the achievements of this attempt scientifically, theoretically, and in term of usage, more acceptable. However according to EA methodologies this research is highly focused on the steps, guidelines, requirements, and the methods of developing and implementing the architecture. Therefore it does not address the details, tools and symbols.

Related Work:

The existing work related to our research goals can be discussed from the two aspects. (1) From EA methodologies, what similar methods exist, and what difference between them and our research? (2) With regard to research methodology and methods, what work has been done, and what’s the new requirement for our proposed methodology?

10.1 Similar systems

In this Section, we’ll first present a methodology for EA based on agent oriented architecture, so that the position of our methodology can be made clearly. Then, I’ll discuss what similar methods are available both from EA and Agent-Oriented Architecture, and the main differences between the proposed AOEAM and its counterparts.

10.2 Current methodologies

First, there is no methodology we can find which can support both EA and Agent-Oriented Architecture as we planed in AOEAM. However, there are many EA methodologies and Agent-oriented methodologies available, respectively.

Approximate Time table

This section describes possible milestones in the current research project. The provisional schedule presented is subject to time, human error and financial resources availability.

Time

Activity

First Year

Preliminary research design

Advanced research design

Critical literature review

Conceptual development

Publishing

Second Year

Testing the method

Use Case

State findings / Disseminate results.

Publishing

Third Tear

Publishing

Table 3- Research time schedule