Of everything that we learned up to now, we can already define the first "two dogmas of faith" of a CIO.

• Its main new function will be to generate value using the best available technologies, it has not been feeling you to enter a company or Government and the things continue in the same way without new generations of value.

• When entering a company or a Government, identify its current Closed Loops and all its pertinent guidelines. Analyze the Closed Loops already existent and its failures inside of the principles of a Feedback Control System, and research the new Closed Loops that are necessaries.

The best technical definition of the Reliability is the following:

• Reliability is the probability of a System to execute, without failures , a certain mission, under certain conditions, during a certain period of time.

If you well designed a System of Control inside the principles of a Feedback Control System, the minimum that you want for any link in whole the involved Closed Loops, it is that he has a good one (calculated) Probability of working well in a certain period of time. The wanted period and the wanted Reliability are obviously defined by you, in agreement with the characteristics of the Process, of the company, of the desired Set-Point, etc.

Let us return to our old graph:

See that the Reliability is part of all the links (of any link of any Loop or Sub-Loop), in the case above equal or larger than 8 (8 = a number arbitrary).

Imagine a race with four athletes, of those in that each one gives to the following a bar to take until the end. You need, to win the race, that its total time is less than four minutes. Let us examine the Reliability now of each one of the four athletes, because each one of them has a past (statistical data) that can be measured. The first three athletes have a past Reliability of 100% in the time of one minute, that is to say, in run of up to 1 minute none of them failed. But the fourth athlete, has a Reliability of only 70% in run of one minute.

As the final result (Set-Point = to win the race) it depends on a sequential work, the weakest link is the fourth athlete, and the Reliability of that chain, as we already know, it is determined by its weaker link: 70%.

In the above graph the Reliability is uniform/harmonic in the 4 links of the Process, but if we imagine that in the link of the Measurements she is 5, that means that link has very better Probability of failing than the other three links.

If you will make a plan in a company, she will be entitled  to demand  the vendors of the involved links (of the equipment used in the links) the Reliability of each link, so that you can calculate all the involved Reliabilities. That demand can be in writing in a purchase, or in the specification of an equipment (a link) for its purchase competition, etc.

Without this, serious effects can appear for your company. We will mention a real example: 

• When defining the conditions of privatization of the phone companies of Brazil, the Brazilian Government didn't define the minimum Reliabilities for each link of communications, and free from that obligation the new phone companies didn't installed Redundancies (we will see that in the future) in most of the links of communications and only done it in some main links, to decrease its investments. The final result is that, a lot of times the connections are not completed, or they interrupt, or they work with low level, etc.

The same thing is worth for the Computer System of your company, that today is usually the spine of a company.

On the other hand, if a Feedback Control System had its unalterable mathematical treatment for about 80 years, the Reliability a lot moved mathematically in the last two decades, pushed that was for the NASA because the obvious reason that the NASA Systems  should be the more (mathematically) reliable possible. And the industry took advantage of that development and used it more and more, mainly in the electronic fields, computation and telecommunications. Besides the industry was forced, liking or not, to use the calculations of Reliability of the NASA, because this forced them by contracts in all the industrial suppliers for the space research.

And  today, even ignoring the NASA, no more is designed an electronic equipment without calculations of the Reliability during the  whole project and the execution of the prototype. In the same way, it  is not projected a System of Computation without the same calculations of Reliability. The NASA continues with its demands, but the market started to also demand them.

In spite of the differences of methods (mathematical models) as above we mentioned, Reliability is defined like this:

R(t) = 0.9998 (0.9998 are an example value).

R = Reliability
t = Time.

Therefore, if we have R(10,000)=0.9998, that means that link will work 0.9998 in 10,000 hours, and it won't work in 0.0002 in the same 10,000 hours. In other words, that link:

• In a mission of 10,000 hours he will work perfectly for 9,998 hours and it won't work for 2 hours (it will be in failure state). And is that acceptable? It will depend on all the other links, of everything besides of the Reliability of each one of the other links. And if it is not acceptable, you will learn, further on, how to transform it to be acceptable.

If a Process (for example, a System of Computation) has the Reliability of R(1,000)=0.9765, any link of that Loop (or of those Closed Loops) can must have a smaller Reliability than that, all the links should be the same or superiors than 0.9765. What involves other subject that we will study also further on, the Redundancy.

We are not talking only about equipments, but also of persons. We will see.