Designing converging corporate networks

When designing converging corporate networks, quality process methodologies cannot be disregarded.

The cyclical evolutionary model of a continuous improvement cycle includes the following phases:

Let us look at each one of these phases:

Planning phase

The requirements of the network to be installed based on the objectives and needs of the system are identified by determining what one wishes to obtain in terms of connectivity, time of response, applications that will be run and contingencies.

Design phase

On the basis of the previous phase, different alternatives of architecture are modeled, taking advantage of available connectivity conditions, optimizing various criteria, which will be examined throughout the present chapter.

Implementation phase

Although it is assumed that the process is linear, in this important phase, the assumptions of the design are checked, and eventually a new design is made to adjust what is planned and designed to reality.

Operating phase

Once implemented, the conditions are there to carry out different actions ensuring the use of all the resources allocated, facilitating their measurement.

Measurement and optimization phase:

Without a doubt, one of the most important phases is the follow-up and verification of all the assumptions adopted in the previous phases.

Measurement permits obtaining adjustment and improvement criteria in the design, starting up another new cycle of improvement.

The specific methodology for designing networks is therefore also based on this model, and the following phases can be identified, materializing the more operational aspects listed below:

• Identifying business needs that the network must meet

Initial business needs are surveyed so as to give priority to the different criteria affecting the design.

• In case the design does not start at zero, the current network should be assessed if relevant.

In the case of redesigning a current network, it is necessary to audit and review its performance, the soundness of the data that are carried must be identified, as well as the quality of service, performance regarding the loss of packages, delays, and response times of the computer applications.

Bottlenecks of the current network should also be identified in this phase.

• Designing the network’s architecture

The architecture, topology, service, quality, routing protocols, routing, etc. should be determined on the basis of the data gathered in the two previous phases.

• If possible, pilot network to check the design hypotheses should be built.

Whenever possible, it is advisable to simulate the real conditions in a model or pilot that checks the general hypotheses before implementing the complete solution.

In this phase, it is reasonable to suppose that there will be new variants leading to an early redesigning of the system, thus resolving many problems and saving costs in subsequent stages.

Translation of information in figure:

N errores detectados=N errors detected
Costo de la corrección=Cost of correction
Diseño=Design
Implementación=Implementation
Entrega=Delivery
Uso=Use

In the chart, the importance of early detection of errors through adequate testing prior to implementation and delivery of the system for use can be appreciated. The cost of correcting errors grows in a non-linear fashion over time; it is extremely costly to correct an error in those stages when the system is already being used.

If tests have to be conducted in the phase immediately after the design, although costs may reach a peak during testing, the cumulative cost of all the corrections is clearly lower because, under controllable conditions, it is quicker and more efficient to introduce changes.

Planning implementation

Implementation phases must be clearly established so as to secure the material and human resources needed in each implementation phase; project follow-up with adequate methodologies such as those developed by the PMI Institute is indispensable.

Documenting the design

Of no less importance is the availability of the complete documentation of all the designs, test procedures, and results expected from all the measurements to be made.

Implementing the design

The methodology that is recommended is the same as that for any other project; the PMI Institute provides powerful tools to monitor implementation of a project.

Checking the design hypotheses in the real system

Although most of the design errors should be found in the testing phase, there is always a group of errors that can only be detected in real operating conditions.

Measuring system performance

As an essential part of providing criteria for redesigning, it is finally necessary to measure each variable affecting performance of the system, as we mentioned earlier, delays, jitter, quality of service of the different applications and the voice/video, recovery time in the face of contingencies, alternative connectivity plans, behavior in the face of changes in topology, etc.

With real measured values, it is possible to decide if it is necessary to redesign part of the system, the contingency plan or the optimization of the quality of service implemented.

Redesigning optimizing the solution

The design cycle’s last objective is finally to introduce the changes that have not been detected in the testing phase and ensuring that the measurements in the real system justify its optimization.

Luis M. Peluffo Cánepa
Academic Director Programme CITEL - ITU
Catholic University of Uruguay
E-Mail: lpeluffo@ucu.edu.uy