Endpoint Protection

From Embedded Lab Vienna for IoT & Security
Jump to: navigation, search


This article describes the importance of endpoint protection and the types of threats an endpoint can be a victim to. It also gives insights on data leakage, the concept of sandboxing and offers possible solutions for protecting an endpoint.

Types of Threats

The CNNIC “Statistical Report on Internet Development in China” defined 20 types of internet usages, then classified internet users into 7 groups based on their habits, time and activities done while surfing on the web or using some applications connected to the internet. Some of these categories are Gamer, Network Businessman or Socializer. People in all of these groups may encounter various threats attempting to harm their systems or steal valuable information from them. Such threats can be fraud, attacks by virus, system overhead, false positive or prevalent malware. In general, threats can be divided into two main types:

  • Internal threats
  • External threats

Endpoint Protection Suites primarily focus on external threats. The five major endpoint threats in 2018 were:

  • Phishing
  • Unpatched vulnerabilities
  • Malvertising
  • Drive-by-downloads
  • Data loss and theft

Data Leakage

Trusting third party companies (agents) with handling confidential data has several advantages: partner companies may gain access to customer data, that can be done by using the same service as the distributor (data owner). This way not only reading, but also updating and altering data in any way can be more efficient, but can also lead to worse data quality, due to perturbation. One of the main goals of the agents is to keep the data quality as high as possible while also preventing data leakage via endpoints. This means the protection of computer networks, that are remotely bridged to client devices, such as mobile phones, laptops or tablets. Many companies are trying to ensure endpoint security by using a single, multi-purpose software solution, that enables administrators to setup a centrally managed security system, but there are also several specialized products on the market. 53% of the companies, that took part in the 2018 Insider Threat report have been victims of insider attacks more than once a year, which highly emphasises the importance of the usage of proper cyber security systems.


Malware (malicious software) is by far the most dangerous type of threat and even though EPPs using the most updated signature databases used to detect such programs, they have widely spread across systems in the globe. The best way to eliminate malware in time is to use behavior-based detection, which is why EDR systems are most suited for this task.


A virus has the capability to spread from host to host and is designed to replicate itself. It is a type of malicious code or program written to alter the way a computer operates. A virus attaches itself to a legitimate program in order to execute its code and in the process, it has the potential to cause unexpected or damaging effects. This can be anything from harming the system software by corrupting data to fully destroying it.

False Positive

This happens, when an antivirus program or any other security system identifies a legitimate code as illegitimate by mistake.

Endpoint Protection Systems

Endpoint Protection Solutions are categorized into two main approaches:

  • Endpoint Protection Platforms (EPP)
  • Endpoint Detection and Response (EDR)

Endpoint Protection Platform

One of the best solutions to achieve endpoint protection is to use an Endpoint Protection Platform (EPP), which consists of several tools, that can protect our systems from unwanted intrusion, such as firewalls, antivirus systems and data encryption. Such an Endpoint Protection Platform has several main characteristics:

  • It is deployed on endpoint devices
  • It focuses on preventing file-based attacks
  • It can detect malicious activity
  • It can provide investigation & Remediation capabilities

Endpoint protection platforms are software using signatures to ensure negative endpoint protection. They also have access to an own database, which stores signatures of collected threats and use it to match them with other signatures, which helps them detecting possible threats. A signature of a virus can be compared to a human fingerprint, is it a set of unique data or bits of code that makes it possible to identify it. An EPP also uses different services to monitor processes running in networks for the sake of malfunction, virus detection. Even though most attacks can be stopped by monitoring and signature matching, some viruses are still able to get in the system and that’s where the so-called Host-based Intrusion Prevention System (HIPS) takes action. It does not only wipe out malicious activities, but is also capable of deep behavioral inspection. The annihilation of threats, that already got into the system needs to be done as fast as possible, since the virus can spread really fast and take over the whole system.

According to a 2019 study the average protection rate of an Endpoint Protection Platform is 99.3% with an average false positive rate of 17 in 3000, which equals to 0.0057%.

Advantages of Endpoint Protection Platforms

  • It provides a single, central endpoint security management solution
  • It simplifies security management
  • It protects and endpoint against most well-known threats

Disadvantages of Endpoint Protection Platforms

  • It cannot deal with the complete network
  • It reduces efficiency & productivity because it needs a lot of resources
  • It cannot protect against insider threats


Endpoint Detection and Response

This is one of the newest cybersecurity technology, which combines endpoint management solutions and endpoint antivirus systems to wipe out malicious activities. EDR are complex systems also designed to monitor the state of multiple devices at the same time in bigger security infrastructures. Since this technology uses a moderner approach to endpoint detection, machine learning also plays a huge part in the functionality of EDR systems using calculations and well-tested algorithms to catch several kinds of suspicious activities from ruining the infrastructure without the need for manually created signatures. When relying on this method, attacks can be detected not by analyzing the file system, but by interpreting and monitoring the system, focusing on its malicious activities providing stable, real-time capture and alerting solutions.

Advantages of EDR

  • It provides greater visibility of the network and the state of each individual endpoint
  • It is able to scan entire networks
  • It can detect unusual activity and respoind by isolating a possibly compromised endpoint

Disadvantages of EDR

  • It produces a high volume of false alarms
  • It creates mountains of backlogs and each alarm has to be assessed individually to identify the threat level
  • Log retention is a burden to the system and its operators so often logs get deleted before they are assessed

Some examples of EDR technology are FireEye Endpoint Security and Symantec Endpoint Protection.


FireEye is a US cyber security company based in California, which offer solutions for cloud, network, email, or endpoint systems. Their Endpoint Protection Platform uses a signature-based engine and also has the capability to use EDR through a behavior-based analytics engine. It is able to find threats, for which signatures do not exist yet using MalwareGuard, which is based on machine learning algorithms, that help fulfilling this task. FireEye has various different features, that can be downloaded as modules on the FireEye Market. Some features include:

  • Investigating known and unknown threats on tens of thousands of endpoints in minutes
  • Identifying vectors an attack used to infiltrate an endpoint
  • Determining whether an attack occurred (and persists) on a specific endpoint and where it spread
  • Establishing timeline and duration of endpoint compromises and follow the incident
  • Clearly identifying which endpoints and systems need containment to prevent

further compromise

It also comes with a malware guard, a machine learning engine to protect against more complex threats, as well as an exploit guard, a behavior analysis engine, which protects against software exploits.

Symantec Endpoint Security

Symantec Endpoint Security provides security at the endpoint for both traditional and mobile devices across the three attack phases—Pre-Attack, Attack, and Post Attack. Some of its features include Attack Surface Reduction, Attack and Breach Prevention, and Post Breach Response and Remediation. Attack Surface Reduction is based on advanced policy controls and technologies, that continuously scan for vulnerabilities and misconfigurations on the endpoint. Its Attack Prevention System uses Machine Learning and Artificial Intelligence on devices and in the cloud to detect schemes to identify evolving threats across device types. Breach prevention is used to detect suspicious activity as early as possible at the endpoint—before they have any opportunity to persist on the network.


Sandboxing is an approach Endpoint Protection Platforms use to prevent intrusions rather than detecting them. It is defined as a method of isolating untrusted data or software that is executed in the same address sapce as the trusted software. By executing suspicious software in an isolated envronemnt its behavior and output can be closely monitored and tested without having to worry that it will impact the entire system or network.

Sandboxing Weaknesses

  • Sandbox environments look slightly different to real environments and can therefore be detected by the malware
  • A Sandbox can only handle and process files up to a certain size
  • The attackers' software can also act as a trojan horse, act innocent during the sandboxing process and then infect the system when triggered afterwards.