Basic features: Malware/antivirus protection: Signature-based detection of known malware. Malware engines should be able to check not only executable files, but also other code, such as malicious JavaScript found on websites.
Application Blocking: Prevent malicious application behavior, such as using a Microsoft Office document that installs another application and starts it.
Behavior Monitoring Systems/Intrusion Prevention Systems (HIPS): This foundational technology protects computers from unknown viruses and suspicious behavior. It should include both pre-execution behavior analysis and run-time behavior analysis.
Web Protection: Search URLs and Block Known Malicious Websites: Blocked sites should include sites that can run JavaScript to perform cryptomining, as well as sites that collect user authentication credentials and other sensitive data.
Web Control: Endpoint web filtering allows administrators to control what types of files a user can download from the Internet.
Data Loss Prevention (DLP): If an attacker can go undetected, DLP capabilities can detect and prevent the last stage of some attacks when an attacker tries to exfiltrate data (get data out). This is achieved by monitoring various types of sensitive data.
Modern features: Machine Learning: There are several methods of machine learning including deep learning neural networks, Bayesian method, clustering, etc. Regardless of methodology, machine learning malware detection mechanisms should be built to detect both known and unknown malware without relying on signatures. The benefit of machine learning is that it can detect malware that has never been seen before, ideally increasing the overall rate of malware detection. Organizations must evaluate detection rates, false positive rates, and the performance impact of machine learning solutions.
Anti-exploit: Exploit protection technology is designed to protect against attackers by preventing the use of the tools and methods they rely on in the attack chain. For example, exploits such as EternalBlue and DoublePulsar have been used to launch NotPetya and WannaCry ransomware. Exploit protection technology stops a relatively small set of methods used to spread malware and carry out attacks, fending off many zero-day attacks that no one knew about before.
Special work against ransomware: Some solutions contain methods specifically designed to prevent ransomware from maliciously encrypting data. Often, special ransomware methods also fix any affected files. Ransomware solutions should stop not only file ransomware, but also disk ransomware used in destructive wipe attacks that spoof master boot record.
Credential Theft Protection: Technology designed to prevent theft of authentication passwords and hash information from memory, registry, and hard drive.
Process Protection (Elevation of Privilege): A protection designed to detect when a privileged authentication token is inserted into a process for privilege escalation as part of an active attack by an attacker. This should be effective regardless of which vulnerability, known or unknown, was used to steal the authentication token in the first place.
Endpoint Detection and Response (EDR): EDR solutions must be able to provide detailed information when looking for elusive threats, keeping IT security operations in tip-top shape, and analyzing detected incidents. It is important that the size and skill set of your team match the complexity and ease of use of the tool in question. Choosing a solution that provides detailed threat intelligence and recommendations makes it quick and easy to respond to a threat.
Advanced Detection and Response (XDR): XDR goes beyond the endpoint and server to include other data sources such as firewall, email, cloud, and mobile devices. It is designed to give organizations a holistic view of their entire environment, with the ability to drill down where needed. All of this information must be collated in a centralized repository, commonly known as a "data lake", where the user can make and receive business-critical queries.
Incident Response/Synchronized Security: Endpoint security tools should at a minimum provide an understanding of what happened to help avoid future incidents. Ideally, they should automatically respond to incidents without the need for analyst intervention to prevent threats from spreading or causing more damage. It is important that incident response tools interact with other endpoint security tools as well as network security tools.
Managed Threat Response (MTR): MTR provides 24/7 threat discovery, detection, and response by a team of experts as a fully managed service. Analysts must be able to respond to potential threats, look for indicators of compromise, and provide detailed analysis of what happened, where, when, how, and why.