19 Jul 2021

Zero-day Vulnerability

19 Jul 2021

A zero-day (also known as 0-day) vulnerability refers to the vulnerabilities that haven't been patched yet.

There may be holes and flaws in a software made by developers, who haven't detected these vulnerabilities and who haven't been informed about them yet. As long as vulnerability is not fixed, it can be used for network access, remote control access, data manipulation, etc.

The term "zero-day" reflects that developers don't have a single day to fix the defect, since it's unknown to them yet.

Most vulnerabilities are ordinary bugs in the application code, rather than flaws in security systems or high-level errors in logic.

The National Institute of Standards and Technology (NIST) reports that 64% of software vulnerabilities stem from programming errors and not a lack of security features.

There are so many ways to make an error in code that will cause a vulnerability. Luckily, there are certain patterns in all that. Most errors can be classified, and specific error patterns can be identified.

These patterns are well-studied and classified. The most significant classifications are:

OWASP Application Security Verification Standard (OWASP ASVS);
Common Weakness Enumeration (CWE);
SEI CERT Coding Standards.

Static code analyzers, such as PVS-Studio, are used to identify zero-day vulnerabilities. Usually, such analyzers support one or more of the standards listed above.

Note. Static analyzers are a fairly general concept. To emphasize that the analyzer is focused on preventing vulnerabilities, it is called a Static Application Security Testing (SAST) tool. See also PVS-Studio SAST.

It's not right to say that SAST solutions directly identify a zero-day vulnerability. Analyzers identify potential vulnerabilities. In other words, they indicate strange / abnormal code fragments that could be errors and security defects.

It's only a small part of the identified errors that pose a threat. Let's take, for example, a "buffer overflow" error. All standards classify this error as extremely dangerous from the security point of view. But only a small part of such errors can be exploited. Most of them cause a lot of troubles, but not fatal consequences. For example, a buffer overflow can cause an image corruption.

Only a part of the errors can cause vulnerabilities though, it makes no sense for programmers to try to separate them somehow. It would be rational to correct all the code sections with issued warnings (except obvious false positives of the analyzer). Even if it's not a potential vulnerability that is being fixed, but just a bug, it is still useful. Especially since it's often very difficult to understand whether one or another error poses a threat from the security point of view. It's better to play safe and fix it.

Additional links: