SoS Musings #44 - Industrial Robots and Cybersecurity
SoS Musings #44 -
Industrial Robots and Cybersecurity
Industrial robots are continuing to grow in use and sophistication in the realm of manufacturing, but are they secure against cyberattacks? An industrial robot is described as a complex cyber-physical system or mechanical device used for manufacturing that is automated, programmable, and capable of movement in three or more axes. Industrial robots are used in place of humans in manufacturing operations to perform tasks considered highly dangerous or repetitive in an accurate manner. These types of robots are typically applied in manufacturing areas that require high endurance, speed, and precision, such as welding, painting, ironing, assembly, palletizing, product inspection, and testing. The global Industrial Robotics market value is expected to reach more than $70,000 million by 2023, exhibiting a Compound Annual Growth Rate (CAGR) of 9.4% during the forecast period. The major catalyst behind this market growth is the significant increase in labor charges, which has pushed manufacturers into replacing human labor with machines, especially during the COVID-19 pandemic. The growth in the use of industrial robots must be accompanied by increased security for such technology.
Several studies have shown the vulnerability of industrial robots to cyberattacks that could significantly impact safety and production activities. Research conducted by Trend Micro found more than 83,000 industrial robots from Belden, Eurotech, Moxa, Westermo, Sierra Wireless, Digi, and more, vulnerable to remote cyberattacks due to their exposure via FTP servers and the exposure of industrial routers. Over 5,100 of these vulnerable industrial robots did not have authentication in place. Trend Micro’s report, titled “Rogue Robots: Testing the Limits of an Industrial Robot’s Security,” also details five types of robot attacks that could inflict harm to human operators and damage equipment, significantly reducing safety for factory workers and the quality of products. The Trend Micro researchers have shown how attackers can abuse software security flaws to carry out such attacks. Two of the demonstrated attacks involve manipulating robot status information to reduce a human operator’s awareness of a robot’s true status and increase the likelihood of the operator losing control and getting injured. The other demonstrated attacks allow malicious actors to alter control-loop parameters, calibration parameters, and production logic. These attacks can cause robots to move inaccurately or unexpectedly, or manipulate the programs used by the robots into introducing a flaw into the workplace, posing a threat to operators’ safety in addition to the integrity and accuracy of manufacturing operations. The researchers pointed out several vulnerabilities that could be lead to the execution of such attacks against industrial robots, which include the use of outdated software components, default credentials, weak authentication, poor transport encryption, insecure web interfaces, unencrypted storage, inadequate software protection, and the ease at which industrial routers can be found and recognized using easily accessible technical materials. IOActive researchers pointed out the potential execution of attacks on industrial robots by insiders such as the robot operators themselves. Malicious robot operators have the potential to be major insider threats in that they can use their direct access to a robot’s hardware or manual interface to alter its behavior, possibly causing operation failures and injuries to others. A malicious robot operator can tamper with exposed connectivity ports using special USB devices and Ethernet connections. A joint study by researchers at the Polytechnic University of Milan and Trend Micro brought further attention to legacy programming languages such as RAPID, KRL, AS, PDL2, and PacScript that were designed decades ago without security in mind and how they leave industrial robots vulnerable to being hijacked by attackers to disrupt production lines and steal intellectual property. The researchers analyzed 100 open-source automation programs developed with these languages and found that many of them contained vulnerabilities that could be exploited by hackers to control and interrupt industrial robot activities. Another study by researchers at IOActive demonstrated how robots could be hijacked by ransomware through the exploitation of vulnerabilities that can allow attackers to execute commands on the robot remotely, potentially crippling factories and businesses. Such attacks and vulnerabilities must be addressed by increased security development and research efforts.
There are several different factors in need of more research and collaborative efforts to improve cybersecurity for industrial robots. The Robotic Industries Association (RIA) calls on robot manufacturers, integrators, and operators to be held more accountable for the security of these robots. As manufacturers design and make robots, robot controllers, and devices such as machine vision cameras, laser scanners, and robot end-of-arm tooling that support the activities and operation of these products, they must ensure that they are implementing security measures throughout the design process and writing secure firmware. Implementors and systems integrators must ensure that manufacturers’ robotic products are applied and configured in a way that doesn’t leave the products susceptible to tampering and remote access by unauthorized entities. Manufacturers should force robot operators to change default user names and passwords at set up. Robot operators need to make sure that the physical environment in which the robot operates remains secure and that cyber risks can be mitigated as quickly as possible. The RIA encourages the adoption of a defense-in-depth approach to robot security. This approach refers to building security into each layer of a robot’s control system architecture. Security defenses must be implemented for a robot’s embedded operating system, application code, communications code, cloud servers, and more. The development of guidance and tools aimed at bolstering robot security must also be continued. For example, Trend Micro Research, in collaboration with the Robotic Operating System Industrial Consortium, created guidelines to help Industry 4.0 developers securely write the task programs that rely on legacy programming languages and are used to control industrial robots’ automatic movements in order to reduce the risk of attacks on such robots. These guidelines cover secure configuration and deployment procedures, authentication for communication between systems, the implementation of access control policies and proper error handling, as well as the performance of input validation and output sanitization. The Polytechnic University of Milan, together with Trend Micro researchers, also developed a tool for the detection of malicious code in task programs used by industrial robots, helping to prevent damage at runtime.
As industrial robots continue to grow in use and complexity, it is essential for the security community, as well as robot manufacturers, integrators, and operators, to take further steps towards developing and implementing better security mechanisms or practices for robotics.