RFID BUSINESS PROCESSES
5.9 RFID BUSINESS PROCESS LIFE CYCLE
5.9.1 Older Life-Cycle Models
Waterfall, incremental, and spiral models are briefly discussed, so we can compare them to the newer models in the next section.
5.9.1.1 Waterfall Life Cycle
The following steps show the stages are seen as sequential in the waterfall;
the model does not allow for iteration of previous stages. As shown in Figure 5.5, all prior stages must be completed before the model proceeds to begin the next stage [5].
We do not see working versions of the product until late in the project.
All objectives, requirements, and constraints must be clearly stated through- out the project; no ambiguity is allowed before the project begins. Because the waterfall model does not treat risk management as an important element or stage, the risks must be perceived to be low in order for the model to succeed. One example of low risk is when the company has experience in
Figure 5.5 Waterfall Life-Cycle Model
Reengineering requirements
Design
Maintenance
Requirements analysis
Integration
building RFID infrastructure in one aspect of the supply chain, then use existing designs to build another such RFID infrastructure in another aspect of the chain. The resources do not change.
The linearity of this model is not practical in the real world of RFID business process reengineering. The projects are rarely sequential, and the resources, including the RFID designs and system interoperability are rarely constant. All objectives, requirements, and goals need to be reengineered at various times in a pilot study or in the SCM Logistics Maturity Model. The reengineering of RFID business processes on a large scale may require several iterations of prior stages of the model for continuous improvement.
The waterfall model does not allow iterations of the stages.
It does not consider the impacts of the emerging RFID technologies on business processes or the impacts of reengineering business processes on each stage. The model does not consider that we need to manage risks of integrating RFID infrastructure with SCP, SCE, and SCM as a way of reengi- neering and improving RFID business processes.
For these reasons the waterfall model is out of the question. Let’s take a look at the incremental life-cycle model.
5.9.1.2 Incremental Life Cycle
The incremental model does not require a complete set of requirements or well-defined objectives and constraints as the waterfall model does. The incremental model can begin with, say, one-third of the requirements in the first project (see Figure 5.6). For our own purpose, we restrict to three elements of the model: requirements, reengineering, and implementation.
When the first project moves to the second stage of reengineering, the next third of the requirements begin in the second project, as the first project moves to the implementation stage and the second project to the design stage.
The last third of the requirements begin in the third project. When the second project moves to the implementation stage, the implementation stage from the first project is incorporated into the implementation stage in the second project and the third project moves to the reengineering stage.
When the third project moves to the implementation stage, the implemen- tation stage from the second project is incorporated into the implementation stage in the third project.
Although the advantage of the incremental model is that it shortens the project time by providing usable functionality earlier, there may be a ten- dency to push difficult problems to the future just to show success to the management. This is something for which the executives have to watch out. They should ask if all requirements have been addressed no matter what time they have started or whether the objectives and constraints need to be well defined at the beginning of the project. Another area that should be of concern to the executives is whether the interfaces between the
implementation stages will be free of bugs or did contain bugs but were not included in the report to the management. The model also assumes the resources are constant and do not contain a risk assessment element such as the Boehm Spiral Model [5] does.
In other words, if the incremental model is perceived to be risky, do not use it. Even if it is perceived not to be risky, it is not well suited for RFID Business Process Reengineering on a large scale as the reengineering effort may require several iterations of prior stages of the model for con- tinuous improvement of business processes. The incremental model does not allow iterations of the model; it allows only the incremental of the implementation stages. Like waterfall models, incremental models do not treat risk management as an important component. They do not consider that we need to manage risks of integrating RFID infrastructure with SCP, SCE, and SCM as a way of improving RFID business processes.
Figure 5.6 Incremental Life-Cycle Model First project
Second project
Third project
Requirements
Requirements
Requirements Reengineering
Reengineering
Reengineering Implementation
Implementation
Implementation
5.9.1.3 Spiral Life Cycle
The spiral model is like the incremental model but with a difference. Unlike the incremental model, the spiral model allows the system to grow in size whether the resources are constant or expanding. Like those for the incre- mental model, objectives, alternatives, and constraints for the spiral model are not always well defined. The spiral model does not require a complete set of requirements, and there is a tendency to push difficult problems to the future to demonstrate to the management that the project was a success when it was not. Unlike the incremental model, the spiral model requires clean interfaces between modules: no bugs and no problems. This is not practical in the real world as interoperability problems between applications in a heterogeneous world are more likely to happen.
In Barry Boehm’s proposed spiral model, prototyping is used to control costs and risks. Prototyping is used when risks are minimized in each of the four phases of the process from requirements to implementation. Can this spiral model be applied to the RFID business processes? Let’s take a look at the steps of the spiral model [5]:
1. Determine objectives, alternatives, and constraints.
2. Plan next phases.
3. Evaluate alternatives, identify, resolve risks.
4. Develop, verify next level product.
5. Risk analysis.
6. Risk mitigation.
7. Concept of operation.
8. Requirements analysis.
9. Design implementation and test.
The spiral model does not allow iterations of risk analysis and risk management stages. It only allows phase iteration of the process. In other words, when the spiral model proceeds to risk analysis and risk manage- ment stages and then reaches the implementation stage, it repeats the process at the second step. The iteration ends when the model “spirals” to the fourth phase. For each phase, risk analysis and mitigation depend on whether the prototyping is well defined.
The spiral model is not well suited for RFID business process reengi- neering. The spiral model does not address business processes in any life- cycle phases. On the other hand, the adaptive and predictive models allow iterations of any stages for continuous improvement until the organization is mature to reach fully the business goals of the models. The strategies, objectives, and goals, the inner core of the adaptive and predictive models, are well defined. If the goals change in response to emerging technologies
and new regulations, then the process of iterative improvements continues.
For each improvement effort, we have a new set of well-defined strategies, objectives, and goals.
The difference between the adaptive linear and dynamic models is that the latter model can interact with other models for a different set of business process reengineering requirements. The former model is a stand-alone.