RIG, Constitución y Leyes

The two alternatives were considered and discussed, and pros and cons were evaluated. Alternative A is regarded as a very important area for research and it is believed that there is a great deal of scope for it in commercial organisations. However, this is a rather long-term proposition, well beyond the time scales for this project. We also felt that the work carried out so far and the theories proposed needed to be tested. This requires some experimental results to support the theoretical arguments. Hence, it was decided to take alternative B, and model a particular case study.

It is important to remark that the modelling work would suffer from some lim itations that we were aware of at this stage. Tim e constraints would force us to focus on one specific case, and the models have to be tailored to this particular scenario. In addition, due to a limitation in the resources available, it is not feasible to build the model based on generic and reusable components, as should ideally be done. These limitations will become apparent in subsequent chapters, when the modelled case study is described. In order to minimise the impact of some o f these limitations, the case study to be modelled m ust be carefully designed, so that although specific, it contains relevant features that make it good representative of real problems. The description of the selected case study is included in chapter 4, followed by detailed discussions of the m odels in chapters 5 and 6. The main outcome o f the modelling work and its connection to the ideas put forward in

Chapter 3. Proposed approach fo r the design and support o f high integrity systems 120

early chapters are included in chapter 7. Additionally, as a consequence of this modelling exercise we also developed an insight into the construction of generic modelling frameworks, i.e. alternative A. A discussion of some ideas in this area is also included in chapter 7.

3.6.

Conclusions

This chapter contains the author’s proposed approach to the problem of maintain integrity in modern telecommunication networks. Firstly, an integrity definition to categorise the integrity of a system in different regions has been proposed. The use of such a framework is to permit identification of the degree of integrity at which a network is operating at any time, and to quickly detect degradation of integrity so that necessary action can be taken to avoid failure. The construction of such a framework requires a quantitative representation of network integrity. The requirem ents of such a measure have been discussed.

Telecommunication organisations need to improve their working practices in order to cope with the dynamic, diverse and complex modern networks and services and the increasing difficulty of preserving the integrity o f these systems. A proposed overall integrity framework containing a number of key activities that must be undertaken to ensure the integrity of networks and services has been described in this chapter. The framework has been divided in two categories, namely static and dynamic. Dynamic actions are taken in live systems, whereas static actions occur ‘off­ line’, i.e. in a testing or laboratory environment, and mainly before the introduction of a change in the network.

The improvement o f integrity practises in an organisation would require an initial analysis of current practises and comparison with the activities proposed in the framework described earlier, and the integrity factors described in section 3.2.2. The cross-relationships can then be documented and used to identify further actions that need to be taken [Monton et al. 97 b]. An example with typical results of such an analysis for a new service process is illustrated in Table 3. The content of each table co-ordinate indicates whether the integrity factors on the top of the table are being applied to the process areas on the left. To facilitate the investigation, the development of a new service is considered in terms o f five sub-processes, namely; requirements capture, service specification, design, implementation and testing. Application of the analysis methodology

Chapter 3. Proposed approach fo r the design and support o f high integrity systems 121

to the design process of a ‘T elco’ revealed a number o f areas requiring action. In particular, the lack of predictive modelling to identify areas of potential risk to network integrity was identified.

Population

models

User

models

QoS Performance Signalling Monitoring Automated

records Risk analysis M odelling Req. ! V ? ? - y ? î - Spec. ! / ? ? - - ? ! y Design j / ? ? y - ? ! y Implem. ! y ? 7 y - ? ! ! Testing ? y ? y y y ! y !

^ Activity carried out ! Activity not carried out

? Lack o f information

_ Not applicable

Table 3. New service process

From the discussions in this chapter, the following conclusions can be highlighted:

a) The need for automated techniques to prevent severe degradation o f network integrity, and permit rapid recovery from failure situations;

b) The need to use formal techniques for the specification, design and testing o f services, i.e. FDTs.

c) New working practises are required, involving more formal and structured approaches and better automated docum entation techniques, i.e. KBSs

d) New service provision paradigm s require modelling techniques to aid understanding and identify problems early in the design cycles.

Chapter 3. Proposed approach fo r the design and support o f high integrity systems 122

M odelling is one key area in the proposed overall integrity framework. The importance of developing consistent, structured and generic modelling methodologies in telecommunication organisations has been discussed. This is an extremely complex task, and identification of requirements for such generic methodologies becomes a necessary and a very interesting topic of research.

A good modelling activity requires careful planning. This includes developing a modelling methodology, achieving an appropriate level of abstraction, identifying the inputs required to the models, identifying the basic components in the models and choosing the language and tools to implement the models. These factors constitute the proposed modelling framework and they have been discussed in this chapter.

As a result o f the analysis undertaken, two possible directions o f research opened up, namely investigate requirements and propose an architecture for the construction of generic, re-usable models, or to put into practise the proposed modelling strategy by modelling a particular service. These two alternatives, their pros and cons, have been discussed, and it was decided to focus on modelling a specific case study. This is the topic of the following chapters.