Monday, February 22, 2010

2.3.4 Rational Unified Process

Rational Unified Process (RUP) [51, 63] is another iterative process model that was designed by Rational, now part of IBM. Though it is a general process model, it was designed for object-oriented development using the Unified Modeling Language (UML). (We will discuss these topics in a later chapter).

RUP proposes that development of software be divided into cycles, each cycle delivering a fully working system. Generally, each cycle is executed as a separate project whose goal is to deliver some additional capability to an existing system (built by the previous cycle). Hence, for a project, the process for a cycle forms the overall process. Each cycle itself is broken into four consecutive phases:
  • Inception phase
  • Elaboration phase
  • Construction phase
  • Transition phase
Each phase has a distinct purpose, and completion of each phase is a welldefined milestone in the project with some clearly defined outputs. The purpose of the inception phase is to establish the goals and scope of the project, and completion of this phase is the lifecycle objectives milestone. This milestone should specify the vision and high-level capability of the eventual system, what business benefits it is expected to provide, some key illustrative use cases of the system, key risks of the project, and a basic plan of the project regarding the cost and schedule. Based on the output of this phase, a go/no-go decision may be taken. And if the project is to proceed, then this milestone represents that there is a shared vision among the stakeholders and they agree to the project, its vision, benefits, cost, usage, etc.

In the elaboration phase, the architecture of the system is designed, based on the detailed requirements analysis. The completion of this phase is the lifecycle architecture milestone. At the end of this phase, it is expected that most of the requirements have been identified and specified, and the architecture of the system has been designed (and specified) in a manner that it addresses the technical risks identified in the earlier phase. In addition, a high-level project plan for the project has been prepared showing the remaining phases and iterations in those, and the current perception of risks. By the end of this phase, the critical engineering decisions regarding the choice of technologies, architecture, etc. have been taken, and a detailed understanding of the project exists. Outputs of this milestone allow technical evaluation of the proposed solution, as well as a better informed decision about cost-benefit analysis of the project.

In the construction phase, the software is built and tested. This phase results in the software product to be delivered, along with associated user and other manuals, and successfully completing this phase results in the initial operational capability milestone being achieved.

The purpose of the transition phase is to move the software from the development environment to the client’s environment, where it is to be hosted. This is a complex task which can require additional testing, conversion of old data for this software to work, training of personnel, etc. The successful execution of this phase results in achieving the milestone product release. The different phases and milestones in RUP are shown in Figure 2.7.

Figure 2.7: The RUP model

Though these phases are consecutive, each phase itself may have multiple iterations, with each iteration delivering to an internal or external customer some well-defined output which is often a part of the final deliverable of that phase’s milestone. Generally, it is expected that the construction phase will be broken into multiple iterations, each iteration producing a working system which can be used for feedback, evaluation, beta-testing, etc. Though iterations in construction are done often and it is clear what an iteration in this phase delivers, iterations may be done meaningfully in other phases as well. For example, in the elaboration phase, the first iteration may just specify the overall architecture and high-level requirements, while the second iteration may be done to thrash out the details. As another example, there may be multiple iterations to transition the developed software, with each iteration “making live” some part or some feature of the developed software.

RUP has carefully chosen the phase names so as not to confuse them with the engineering tasks that are to be done in the project, as in RUP the engineering tasks and phases are separate. Different engineering activities may be performed in a phase to achieve its milestones. RUP groups the activities into different subprocesses which it calls core process workflows. These subprocesses correspond to the tasks of performing requirements analysis, doing design, implementing the design, testing, project management, etc. Some of the subprocesses are shown in Table 2.1.

One key difference of RUP from other models is that it has separated the phases from the tasks and allows multiple of these subprocesses to function within a phase. In waterfall (or waterfall-based iterative model), a phase within a process was linked to a particular task performed by some process like requirements, design, etc. In RUP these tasks are separated from the stages, and it allows, for example, during construction, execution of the requirements process. That is, it allows some part of the requirement activity be done even in construction, something the waterfall did not allow. So, a project, if it so wishes, may do detailed requirements only for some features during the elaboration phase, and may do detailing of other requirements while the construction is going on (maybe the first iteration of it). This not only allows a project a greater degree of flexibility in planning when the different tasks should be done, it also captures the reality of the situation—it is often not possible to specify all requirements at the start and it is best to start the project with
some requirements and work out the details later.

Though a subprocess may be active in many phases, as can be expected, the volume of work or the effort being spent on the subprocess will vary with phases. For example, it is expected that a lot more effort will be spent in the requirement subprocess during elaboration, and less will be spent in construction, and still less, if any, will be spent in transition. Similarly, the model has the development process active in elaboration, which allows a project to build a prototype during the elaboration phase to help its requirements activity, if needed. However, most of the implementation does happen in the construction phase. The effort spent in a subprocess in different phases will, of course, depend on the project. However, a general pattern is indicated in Table 2.1 by specifying if the level of effort for the phase is high, medium, low, etc.

Table 2.1: Activity level of subprocesses in different phases of RUP.


Overall, RUP provides a flexible process model, which follows an iterative approach not only at a top level (through cycles), but also encourages iterative approach during each of the phases in a cycle. And in phases, it allows the different tasks to be done as per the needs of the project.

1 comment:

  1. Useful information like this one must be kept and maintained so I will put this one on my bookmark list! Thanks for this wonderful post and hoping to post more of this!

    IT Infrastructure Relocation checklist

    ReplyDelete

Pages