If you ask 10 consultants to define what security testing is today, you will more than likely get a variety of responses. Here is the Wikipedia definition:
"Security testing is a process and methodology to determine that an information system protects and maintains functionality as intended."
In my opinion, this is the most important aspect of penetration testing. Security is a process and not a product. I'd also like to add that it is a methodology and not a product.
Another component to add to our discussion is the point that security testing takes into account the main areas of a security model. A sample of this is as follows:
Authentication
Authorization
Confidentiality
Integrity
Availability
Non-repudiation
Each one of these components has to be considered when an organization is in the process of securing their environment. Each one of these areas in itself has many subareas that also have to be considered when it comes to building a secure architecture. The lesson is that when testing security, we have to address each of these areas.
It is important to note that almost all systems and/or networks today have some form of authentication and, as such, it is usually the first area we secure. This could be something as simple as users selecting a complex password or us adding additional factors to authentication, such as a token, biometrics, or certificates. No single factor of authentication is considered to be secure by itself in today's networks.
Authorization is often overlooked since it is assumed and not a component of some security models. That is one approach to take, but it's preferred to include it in most testing models, as the concept of authorization is essential since it is how we assign the rights and permissions to access a resource, and we would want to ensure it is secure. Authorization enables us to have different types of user with separate privilege levels coexist within a system. We do this when we have the concept of discretionary access, where a user can have administrator privileges on a machine or assume the role of an administrator to gain additional rights or permissions, whereas we might want to provide limited resource access to a contractor.
Confidentiality is the assurance that something we want to be protected on the machine or network is safe and not at risk of being compromised. This is made harder by the fact that the protocol (TCP/IP) running the Internet today is a protocol that was developed in the early 1970s. At that time, the Internet consisted of just a few computers, and now, even though the Internet has grown to the size it is today, we are still running the same protocol from those early days. This makes it more difficult to preserve confidentiality. It is important to note that when the developers created the protocol and the network was very small, there was an inherent sense of trust regarding who you could potentially be communicating with. This sense of trust is what we continue to fight from a security standpoint today. The concept from that early creation was and still is that you can trust that data received is from a reliable source. We know now that the Internet is at this huge size and that is definitely not the case.
Integrity is similar to confidentiality, in that we are concerned with the compromising of information. Here, we are concerned with the accuracy of data and the fact that it is not modified in transit or from its original form. A common way of doing this is to use a hashing algorithm to validate that the file is unaltered.
One of the most difficult things to secure is availability, that is, the right to have a service when required. The irony of availability is that a particular resource is available to one user, and it is later available to all. Everything seems perfect from the perspective of an honest/legitimate user. However, not all users are honest/legitimate, and due to the sheer fact that resources are finite, they can be flooded or exhausted; hence, is it more difficult to protect this area.
Non-repudiation makes the claim that a sender cannot deny sending something after the fact. This is the one I usually have the most trouble with, because a computer system can be compromised and we cannot guarantee that, within the software application, the keys we are using for the validation are actually the ones being used. Furthermore, the art of spoofing is not a new concept. With these facts in our minds, the claim that we can guarantee the origin of a transmission by a particular person from a particular computer is not entirely accurate.
Since we do not know the state of the machine with respect to its secureness, it would be very difficult to prove this concept in a court of law.
All it takes is one compromised machine, and then the theory that you can guarantee the sender goes out the window. We won't cover each of the components of security testing in detail here, because that is beyond the scope of what we are trying to achieve.
The point I want to get across in this section is that security testing is the concept of looking at each and every one of these and other components of security, addressing them by determining the amount of risk an organization has from them, and then mitigating that risk.