Enterprise Unified Process, The: Extending the Rational Unified Process [ペーパーバック]

In the interests of disclosure, I must share that I was one of the technical reviewers for this fascinating book. But my recommendations of it are based upon its sensible content. Here is my take on how the Enterprise Unified Process can improve your organization.

Enterprise thinking is not about thinking "bigger", it's about thinking differently. The Enterprise Unified Process is a qualitative leap forward that will change the scope of your project from simply a development lifecycle to a SYSTEM lifecycle. The authors aptly recommend that if you are not yet using an iterative process, then you should first adopt the Rational UP, Scrum, FDD or other project-focused development process. But if your needs go beyond merely managing or executing development projects, you will need to move to the EUP.

Structurally, the EUP builds on the RUP, and augments RUP to address the needs of an enterprise rather than just software development. Structurally, the EUP adds two new phases to RUP: the Production phase and the Retirement Phase. The Production phase begins after RUP's Transition phase, and this is where each release of software spends its entire life after it is developed. The activities of the Production phase focus on keeping your software and systems running, backed-up, and that defect reports are being addressed. The Retirement phase defines the activities involved in removing software, or a system, from production. Retirement is a complex process that can be accomplished all-at-once, or incrementally over time, but retirement always requires coordination with other processes in your organization.

EUP adds a new support Discipline to RUP: the Operations and Support Discipline. This discipline includes disaster planning and recovery, service-level agreements between operations and support teams, and addressing end-user questions and problem reports for systems in production.

But EUP goes far beyond vanilla RUP with the addition of "Enterprise Management Disciplines" that are specific and unique to the enterprise context. Here, EUP introduces the disciplines and associated roles and activities for:

* Enterprise Business Modeling

* Portfolio Managment

* Enterprise Architecture

* Strategic Reuse

* People Management

* Enterprise Administration

* Software Process Improvement

Each discipline is explained clearly with discussion of the workflows that are executed, sample case studies, tools that can assist you in the discipline, and--my favorite--"anti-patterns" for each discipline. For me the anti-patterns helped me to better understand each discipline's goals.

If you are familiar with RUP, you should have a feeling now for how much RUP does not address for the enterprise, and how much the EUP has to offer. But how can an organization move to EUP? Luckily, the authors conclude the book with sound recommendations on adopting the EUP. If you are involved with, or responsible for, software development to support your company's business goals, this book will help you understand the many dimensions that RUP does not begin to address.

Scott Ambler and the crew at Ronin have helped define the nature of Agile development (see www.agilemodeling.com) and that is the task set here : make RUP, a very structured IT development process more agile and flexible. But also along the way, Scott and crew also help to make it more complete - they bring the single process up into a larger framework of an organizational viewpoint. This is aided by extending the Zachman framework slightly plus adding three key components to the RUP development cycle.

First, the the RUP Transition is expanded into Transition, Production and Retirement. By explicitly bringing a complete cycle view of development key IT asset issues are addressed which otherwise might be omitted. Ditto for adding Strategic Reuse and Software Process Improvement - these are tough issues to handle. For example, many eXtreme Programming adherents are dead set against any amounts of time spent on Reuse and Strategic Improvements; but Scott and company set some disciplined guidelines for both processes.

But the most important contribution of the book is to set reasonable goals for modeling and planning- in effect, EUP strives to make these activities deliver clear benefits measured better team communications, consensus building and measurable targets and process goals. Agile uses models as tests of design.

So what is missing, since we rated this 4 on 5. First, although the book does treat risk management it does not use it as a guide to process effectiveness - those models and plans and every step of the EUP process better help reduce the risks of a project or one of its phases going awry or they have failed to deliver. See Boehm and Turner's Balancing Agility and Discipline for the risk based approach to project management. Second, SOA-Service Oriented Architectures and Web Services are changing the nature of development profoundly because they up the value of workflow and integration between systems (long neglected in IT projects). As noted EUP catches some of this but unfortunately does not look at the specific ramifications.

But overall this book is conundrum challenging - its a disciplined and fruitful look at making a structured planning process more agile.

The Enterprise Unified Process (EUP) unites diverse disciplines, including development, enterprise architecture, operations, production and portfolio management, reuse and business process modeling, under an easy to follow framework. It was refreshing to find a book that recognizes the need to accommodate the installed base of existing software as part of the planning, development and deployment process. This is an excellent guide for any manager who wants to ensure that essential IT disciplines are addressed.

The focus of EUP is to enhance the commonly accepted Rational Unified Process (RUP). The authors have added new disciplines to RUP that include business modeling, portfolio management, enterprise administration, reuse, enterprise architecture and process improvement. The introduction of business modeling into the overall process is essential to weave IT processes and disciplines into the most essential driver of any systems initiative - the business. The enterprise architecture discussion was also refreshing given that many organizations have forgone this discipline and have created redundant, stovepipe applications and data structures that significantly stifle business agility.

The "Reuse" chapter raises the rarely deployed reuse strategy. It is critically important to not replicate business processes, models, systems, data structures, source code and interfaces. The costs and risks of trying to keep parallel assets synchronized have been written about extensively. This book promotes the idea that reuse is just another aspect of the enterprise unified process. It is also one of the few discussions about reuse that recognizes the value of harvesting existing assets.

Also of note is the portfolio management discussion that focuses attention on the need to incorporate project management with application management. It should be noted, however, that portfolio management has much less focus on applications than the traditional industry definition as promoted by Gartner, Inc.

Finally, this book makes great use of tips, tool references and citations to books or papers that readers can use to expand on their understanding of a given topic. The last chapter of the book takes a realistic and honest look at deploying the enterprise unified process, including its possible retirement.