|Novel Approaches to the Monitoring of Computer Networks|
|Prev||Chapter 8. A Coalescence of Systems||Next|
There are three fundamental differences between the systems examined in this work and traditional network monitoring systems.
Most network monitoring solutions try to solve the most general case to increase their usefulness and hence their marketability. This makes sound business sense, but this generalisation is often done at the expense of little details. The result is often a system that almost does what is required of it, but is not quite correct. This is the major difference between an off the shelf network management package and a purpose-developed system — an off the shelf solution does most of what is required, and a number of other things at the same time; a purpose built solution does exactly what one wants it to do, and nothing more. The disadvantage of purpose-developed solutions is that they are not usually portable to other situations.
The four applications developed as part of this project fall somewhere between these two extremes. They are network-independent enough to allow them to be usefully applied to other situations, but they attempt to solve a very specific problem rather than the general case. This network independence is achieved partly through the use of an abstraction layer that provides vendor independence.
Vendor independence is another difference between the systems developed in this project and the vast majority of commercial network monitoring applications. Most commercial applications are limited to a single manufacturer's products (or at best, a small subset of manufacturers). The solutions presented in this work are, on the other hand, almost completely vendor independent, as is noted in Section 8.3.
The third difference between these solutions and traditional network monitoring approaches is that, to a large extent, the solutions presented in this work attempt to answer the question at hand, rather than provide the network administrator with the information required to answer the question. This difference is perhaps best illustrated by the intelligent reporting system presented in Chapter 7. Traditional approaches to reporting provide symptomatic information, whereas by employing artificial intelligence techniques, the system in Section 7.3 provides information on the cause of the problem.
To a large extent, these differences are the defining feature of the solutions presented in this work, and are what make the approaches used in this project novel.