Testing the Internet as a global comprehensive ICT system

Testing ICT systems represents one phase of any project, but also a measure of quality assurance for products and services of ICT technology. Unlike other industrial testing system tests are performed at three different levels and at the level of meeting the technical specifications for hardware, functional and technical requirements for software and system behavior in different environments and boundary conditions. Also, as a special type of testing is examined to prove the reality.

Most often when testing using best industry practices, but often it is necessary to expand testing beyond those areas that do know the industry or by the system designer had the idea for the system under the given conditions and in order to give way. This is the case whenever the system is operating outside the border, that is, when manufactured and installed the system designer or exceeds the expectations of the contracting authority so users are discovering the many unforeseen ways to use the system.

Standard testing system covers borderline cases, performance or design conditions do not predict, inputs and methods of use and is usually carried out for the devices directly over them while the large and spatially distributed systems commonly tested model of system in which it is very important to model as closely as possible represents your system.

Tests are carried out so that collectively cover all the complex specifications of the system as well as all possible outcomes, with the number of different tests limited to a reasonable level so that the new testing methods are introduced, and only if the existing features of there is a possibility of a new outcome by applying a given test - there is no need to prove what has already been proven.

It is worth mentioning that the tester examines and proves the reality and that the tests conducted on the given mathematical models cannot be more than the accuracy of the accuracy of the model, so that every inconsistency or error in the model automatically leads to errors in testing.

Perhaps the best recommendation for the testers to be professional skeptics that do not believe none of the pictures, scheme or information obtained from the developers of system, operatives or owners of the system. One of the reasons why you should be skeptical of everything is a phenomenon that is often necessary to simplify the ideas that is being implemented in order to fit the same in certain technical standards, and that could easily have presented to those who work with them should be introduced only at the level of ideas, not to possess a deeper knowledge in the field. Testers then have to measure and verify the reality and believe only the data that provoke real system, and that the evaluation of results trying to keep it a stronger critical line.

Simplification as we have already noted the problem drains, because the data obtained from the real system differ significantly from those obtained from the model.

Special characteristics of the testing system is the fact that no one not even the designers nor testers do not know everything about what and how system works. This is because the system has wider limits than those observe. In this sense there is a problem that is part of the system must be observed and measured, and which do not, or that are relevant to the characteristics and the relevant parts of the system interdependent relationships.

The question that arises when there are some discrepancies and anomalies system is whether these problems are important for the whole system or local, that affect other elements of the system, that is transmitted and how to escalate. Since most professionals in the ICT sector have no education from systems theory most local problems is considered isolated and local, and the fact that the system each element system affects every other element system and the system as such is often forgotten or consciously disregarded whereby only the question whether the accumulated small problems escalate enough to threaten the functioning of the whole system.

In this respect it should be noted that in theory all the of system has a context. This means that testers must evolve context before testing the system. This is particularly important when testing software on the extra functionality, ie, the gray area of application, device or software when working outside the projected value. The unwritten conditions require the use of checks and those conditions that were not originally planned in the project. In this way, the examination of real solutions expands the area in which the reliability of the device or software can be used. This type of test has so far been mainly reserved for the software components of system, but rarely for infrastructure, but due to the rapid growth of infrastructure and continuous overload due to the accelerated growth of the services that use the same this kind of testing has become a very interesting and in the area of infrastructure.

In this sense it is interesting to check the growth scenarios, which can easily be described as an attempt to understand how the project will fall apart if exaggerate any of the elements of system, ie, in this particular case the network infrastructure to fall when the network appears too many simultaneous services that attempt to benefit at the same time (the problem of lack of capacity). In this way, testing the system becomes a engine for constant deep learning about the system.

Testing system must not be so expensive that company or institution does not want to do them, but not so cheap that due to the fluctuation of knowledge on testing techniques become random. Therefore, it is important to note that many so-called automated testing technically are not testing but only characteristics proof .

Automated tests cannot be aware of the emergent problems, especially if it occurs in the time between two automatic testing. In this regard it is important to always perform system testing and application of human intelligence, because at the moment there is no automated software system or artificial intelligence service enough complex, precise and accurate which could meet the requirements of automated testing, while still maintaining critical awareness.

Also, control of the work of testers is particularly difficult in scripted tests. However, tools such as he Panaya Scenario Recorder allow you to create appropriate test scenarios that could be automatically used to replace human testers work in certain areas of freeing people for the necessary training in the field of testing. Testers need to stay sharp in order to properly do their job.

If we look at the work of testers in the ICT industry will notice that in this industry testing almost everything, every product or service, software, hardware, system elements, but never carried out extensive testing by full global infrastructure utilization. Although there is a very large number of tests that can consistently implement the elements of Internet infrastructure nobody has ever conducted a real full scale crash test of Internet. Due to the size and complexity, as well as countless technical redundancy of the parts of the software and hardware infrastructure, there is a question whether this is feasible. But how, not only our economy, government and legal system depends very much on the Internet, but you can also say that most of the lower social infrastructure (companies, families, associations, cities, personal connections) directly and inextricably linked to a large degree dependent on the Internet raises the question is whether such a test is necessary and what would be able to learn from it and learn.

If we agree that it is necessary, at least in terms of determining the limits of resistance system it is necessary to re-envision how to make this test work conducted in such a way that the consequences that its implementation and enforcement are minimized. Certainly knowing the limit of resistance system would help to more clearly the need for redundancy institutions and develop procedures for disaster recovery. The development of these procedures is necessary and it can be said is critical to minimize the effects of disasters, but also crucial for the operational management of the global network in the case until the fall of the Internet in its entirety and come.

Since it is difficult to assume that the planet could engage a natural disaster that completely destroyed the infrastructure of the Internet, but at the same time was such that the human race is able to survive come to the conclusion that the only likely scenario in this case the operation of malicious individuals or groups united around this objective.

Of course, such an ambitious goal is hardly feasible without a huge technical and human resources required to implement such a project at work. And if there is a group with this aim it is clear that this is a terrorist group or as likely cyber power some of the countries that have this kind of security forces.

The fact that from year to year on the Internet appears more and more malicious software, that frequent destructive attacks the thesis that in the foreseeable future possible hypothetical scenario in which to let the militant group could decide to attempt to overthrow the Internet in its entirety. Of course, the question is what would be motivated, and by what means and techniques make this attempt was implemented, but it is obvious that the probability of such an event is growing in the future, so it is not illogical to assume that it will nor be some kind of global testing the Internet in the future which would could be implemented at the level of physical and logical infrastructure.

For now, there are a number of ways the Internet as a global verification system using a simulation model or on their characteristics and methods of verification largely silent, mostly for security reasons, as confirmed successful scenarios are nothing but possible plans for attacks. In addition to include the observation of the real behavior may simplify the infrastructure of these models require considerable processing power, and the results obtained by them are not the most reliable.

Another way to check the resistance of a global network of testing for reduced size, but such testing also have a lower degree of complexity, ignoring the synergetic effects that increase with the size.

In general it can be said that the inquiries into real resistance indispensable types of Internet checks for a global infrastructure, but also that due to the size and complexity of the same, is not possible in reality, however, resorted to testing models that give only a vague and approximate picture of the behavior of the global network conditions close to the fall of the network.

Although extremely significant hopefully give us this kind of testing data collected will never need in reality and that the models generated by consolidation and evaluation remain the same only the necessary security protocols that we will never need to use this data in order recovered global network from disaster of such proportions.