With the rapid advancements in the current digital era, cybercriminals are also adopting advanced techniques for attacking their targets. They can often be seen adapting to and using sophisticated techniques to target and infect their victims. For instance, when hackers going by the name Shadow Brokers had leaked NSA’s secret exploit called EternalBlue in April 2017, within the next few days, several cybercriminals had misused that exploit to create deadly cyber weapons of mass destruction. Traditional security techniques like signature-based anti-malware software are no longer capable of combating such sophisticated threats. More advanced methods of detecting and preventing malware via behavior-based analysis are not just buzzwords but have become a standard norm for several industries. This article provides some insights into the shortcomings of signature-based security and how behavior-based security helps attain the required levels of security.
Every malware has a unique signature (a unique string of bits, cryptic hash, or a binary pattern) that can be considered as a fingerprint for the unique identification of that malware. Since the inception of malware, most antivirus technologies were using signature-based malware detection as the primary weapon against malware-laden intrusion attempts. The anti-malware software would monitor all the data entering into a system and scan the contents to check if the source code or hashes in the files or packets match with any of the known malware threats. Signature-based detection methods were simple to implement and update for security vendors. For this, all anti-malware vendors maintained their library of known and identified threats. The efficiency and accuracy of these software products were measured in terms of the vendor having the maximum malware signatures, and their capabilities to include new signatures and push them to client systems. This method provides excellent and reliable protection against millions of known and active threats.
Limitations of signature-based security
Signature-based software is useful in detecting and protecting against an already known or identified threat, but they have some limitations. They cannot detect newly discovered threats like zero-day attacks, which are not known to the world before they are seen in the wild. Some of the recent examples of such attacks include WannaCry and the Petya ransomware outbreaks that inflicted maximum damage because of this basic flaw in the existing security systems. Also, to exploit this shortcoming, cyberattackers have started mutating their malicious code by making slight changes in such a way that their malware keeps generating new signatures while retaining its malicious functionality. This can be done via simple code transformation methods, like the insertion of junk code, applying code permutations, code expansion or shrinking, and register renaming. This allows attackers to strike in quick succession, causing catastrophic impact. To deal with such attacks, instead of relying on the scan of benign files, security vendors need to focus on detecting the basic nature or behavior of the malware.
Behavior-based security software is developed with embedded intelligence to consider deviations from the malware signatures and is capable of identifying if the incoming files may pose any threat to the networks or systems. This provides an effective way to secure end-user devices, network elements, and servers from any malicious or even potentially malicious activities.
In behavior-based detection, the software is programmed to analyze and evaluate every single line of code and analyze all the potential actions that may be performed by that code, like access to any critical or irrelevant files, processes, or internal services. Execution of OS-level instructions and rootkit level low-level code is also included in this analysis. The software tries to detect all malicious or potentially malicious activities that may have any adverse impact and notifies the concerned people to take necessary actions.
There can be multiple dimensions of malicious behavior, that are required to be scanned. This includes behavior-based intrusion detection, behavior-based threat analysis, and user behavior analytics products. Most behavior-based programs follow a policy-based control mechanism. A standard set of policies based on the vendor’s experience and expertise helps define the behaviors that can be allowed to be executed. In addition, this software also allows administrators to create or modify the policies to allow or disallow any specific requirements which may be specific to that organization or industry.
Most of the behavioral detection solutions are equipped with advanced technologies like machine learning, advanced correlation engine, and behavioral biometrics that allows mapping of typical malicious behavior like rootkit installation, attempts for detection of the sandbox environment, or attempts to disable security controls.
Limitations of behavior-based security
Signature-based detection uses a static analysis mechanism, which can be performed in real-time. But this is not the case with behavior-based security. A dynamic analysis across multiple dimensions introduces some latency, negatively impacting the performance.
Besides, there is a category of malware that first tries to detect if it is running in a sandbox. Such malware equipped with an anti-sandboxing technique may be able to avoid detection by preventing any malicious activities. Also, several behavior-based security solutions are exclusively cloud-based, which may not go well with the policies and compliance norms.
Adopting behavior-based security: Quick tips
Every industry and organization is unique and has its own definitions and parameters for security. For instance, in several financial organizations, having two to three failed attempts may be an acceptable norm, but in some critical industries like energy or oil, having a single failed attempt may also be considered as an alarming situation.
Due to the nature of behavior-based security, it is almost certain that some level of tuning will be required when you want to implement it in your environment. Customizations based on your business needs cannot be expected out-of-the-box. Initial baselines set by the vendor and the recommendations by the industry-specific leaders can help get started, but eventually, you will be required to do the fine-tuning to optimize any behavior-based security software for your organization.
When implementing behavior-based security, organizations must consider the following:
- Start early: Instead of waiting for any tipping point or major incident to happen, start exploring all the possible options for the adoption of behavior-based security. Try out the available options to get some know-how before making an actual investment.
- Collect all possible data: When more data is available for analysis, the machine learning-based system will have better chances of identifying anomalies.
- Use the optimal tools: There are several machine learning tools available in the market, both open-source as well as commercial. Explore them and choose the one for which you will have the right combination of expertise, compatibility, and budget in your organization.
- Self-enhance yourself regularly: The criminals keep changing their attack methods and keep trying new and different ways to penetrate inside the targeted networks. So you must also keep reviewing and enhancing your behavior-based security policies to find and block these adversaries.
Signature-based security and behavior-based security: Use the right combination
Both signature-based and behavior-based malware detection techniques have their own advantages and disadvantages. Using the right combination of the two helps organizations achieve an advanced level of protection. For instance, while behavior-based security can help dodge any new zero-day malware threat, a quick look back of relevant parameters (indicators of compromise) into the existing signature-based firewall and anti-malware software can instantly help prevent massive floods or waves of these attacks, providing extra layers of security across the networks. Optimal security can be obtained by having the right combination of both technologies.
Featured image: Freepik.com