An Architecture and Prototype for Target Process Areas at CMMI Levels 3 and 4

近年來，能力成熟度整合模型(CMMI)已廣泛地運用在軟體產業。若以CMMI作為軟體生產流程控管的標準，可提升台灣軟體製造能力，進而提昇台灣軟體產業的國際競爭力。然而，目前台灣大多數的軟體公司仍停留在CMMI層級一或是層級二。主要的原因是：一、CMMI架構太過於龐大，欲採用CMMI作為流程改善標準的公司不知道從何著手。二、目前尚缺乏支援層級三或層級四的相關工具及導入方法。因此，本研究將重心放在CMMI層級三和層級四，並提出相關工具來支援軟體業者邁向CMMI更高的層級。 首先，我們重新詮釋CMMI架構，並定義出五大焦點流程領域(組織流程定義、組織流程專注、組織訓練、組織流程績效，和數量化專案管理)。根據焦點流程領域，提出CMMI層級三和層級四的支援架構。 此架構包含流程描述與定義、流程模型、流程模擬，以及流程分析四大模組。我們實做一個系統雛型包含三大主要模組。藉由本系統，軟體業者可以模擬軟體開發流程，衡量流程績效和產品品質，並透過參數微調，評估不同的流程改善方案。這些弁鈳ㄔR分支援CMMI層級三和層級四的精神---流程定義和數量化管理。 本架構及雛形可達成大部分的焦點流程領域的特定目標和特定實務，若輔以配套的管理制度，我們相信，本研究可以協助軟體業者往CMMI層級三和層級四邁進。Capability Maturity Model Integration (CMMI) has been widely used in recent years. Adopting CMMI helps software vendors improve their software development ability, lower costs, reduce cycle time, and insure high quality products. It is also advantageous when facing competitors from China and other countries. However, most software companies in Taiwan belong to CMMI Level 1 or Level 2. The key problems are as follows: (1) CMMI is a formidable document. Its size alone makes it hard to comprehend and use for process improvement. (2) There are few tools and methodologies to support Levels 3 and 4. Therefore, we focus on process areas at CMMI Levels 3 and 4 and present related tools to support them. First, we reinterpret the CMMI maturity model and identify our target process areas (i.e. Organizational Process Definition, Organizational Process Focus, Organization Training, Organizational Process Performance, and Quantitative Project Management). Based on these five process areas, we then propose a CMMI support architecture for Levels 3 and 4. Our CMMI Levels 3 and 4 Support Architecture presents four components that support our target process areas. We implement a system prototype that has three main components, namely: process modeling, process simulation and process analysis. It enables quantitative evaluation of software processes and facilitates comparison of different process designs. These functions totally support the spirit of Levels 3 and 4, i.e. process definition and quantitative management. We believe that our architecture and prototype support target process areas because we have accomplished most of the specific goals and specific practices in each process area. That is, we provide a tool to help software vendors move to higher levels of CMMI.Chapter 1 Introduction 1 1.1 Motivation 1 1.2 Objective 2 1.3 Organization 3 Chapter 2 Literature Review 4 2.1 CMMI 4 2.1.1 The History of CMMI 4 2.1.2 The CMMI Framework 5 2.1.3 The CMMI Maturity Model and Process Areas 8 2.1.4 The Four Categories of CMMI 12 2.2 Software Process 20 2.2.1 Software Process Definition 20 2.2.2 Software Process Features 20 2.2.3 ISPW-6 Software Process Example 20 2.3 Software Process Modeling Simulation 24 2.3.1 Definition 24 2.3.2 Why Simulation? 24 2.3.3 Application Fields 25 2.3.4 Software Process Models and Simulation Approach 26 2.4 Related Works 29 2.4.1 Martin, Raffo, and Vandeville (1999) 29 2.4.2 Christie (1999) 29 Chapter 3 Research Target 32 3.1 Key Problems 32 3.2 Reinterpreting the CMMI Maturity Model 33 3.3 Identifying Target Process Areas 36 Chapter 4 Research Method 41 4.1 Terminology Definition 41 4.2 Hypothesis 42 4.2.1 Hypothesis 1 42 4.2.2 Hypothesis 2 42 4.3 CMMI Levels 3 and 4 Support Architecture 43 4.3.1 Process Description and Definition 44 4.3.2 Process Modeling 44 4.3.3 Process Simulation 47 4.3.4 Process Analysis 51 4.3.5 Data Collection Issues 51 4.4 Implementation Issues 52 4.4.1 Programming Language and Platform 52 4.4.2 Simulation Approach 52 Chapter 5 Application Scenario 55 5.1 ISPW-6 Software Process Description 55 5.2 Simulation Scenario 56 5.3 Process Modeling 56 5.4 Simulation Experiment 58 5.4.1 Key Calibration 58 5.4.2 ISPW-6 Process Simulation-Non Feedback 59 5.4.3 Alternative Resource Allocation 61 5.4.4 ISPW-6 Process Simulation-Feedback 62 5.5 Evaluation 64 Chapter 6 Conclusion and Future 68 6.1 Conclusion 68 6.2 Future Work 69 6.2.1 Visual Models 69 6.2.2 Hybrid Models 70 Bibliography 71 Appendix 7