Smart Irrigation Systems Expose Water Utilities to Attacks
13.8.2018 securityweek IoT

A team of experts has analyzed smart irrigation systems from several vendors and found vulnerabilities that can be exploited to cause potentially serious disruptions to urban water services.

Researchers from the Ben-Gurion University of the Negev in Israel recently published a paper describing what they call a “piping botnet,” a botnet of smart sprinklers that can be used to quickly empty water towers and even anti-flood water reservoirs.

Based on an analysis of popular smart irrigation systems from RainMachine, BlueSpray and GreenIQ, they determined that a bot running on a device in the same local area network (LAN) can detect an irrigation system within 15 minutes by analyzing outgoing traffic. They also showed that the bot can initiate the watering process and cause significant damage.

“While previous attacks against critical infrastructure required the attacker to compromise the systems of critical infrastructure, we present an attack against critical infrastructure that does not necessitate compromising the infrastructure itself and is done indirectly by attacking attacking client infrastructure that is not under the control of the critical infrastructure provider,” the researchers explained in their paper.

Smart irrigation systems rely on sensors and online services for improved efficiency. Users can control the system remotely from a mobile phone or computer, and they can configure it using a dedicated cloud service. These products can also adapt the watering schedule based on data obtained from weather forecast services.

While many smart irrigation systems communicate via Wi-Fi, some also have a GSM component that gives them direct access to the Internet. Researchers conducted an Internet search using Shodan and discovered tens of exposed devices from one of the targeted vendors.

According to the experts, malicious actors can create a botnet of smart irrigation systems by infecting various types of Internet-connected devices with malware (e.g. routers, laptops, smartphones). The malware searches the local network for irrigation systems and takes control of them using various security flaws. The attacker can then manipulate the compromised system via command and control (C&C) servers.

Once in the network, hackers can launch man-in-the-middle (MitM) attacks and spoof the input of the irrigation system. Researchers found that attackers could spoof the system’s configuration, the weather forecast, and various sensors (rain, water flow and soil moisture sensors) to manipulate the sprinklers.

In addition to spoofing attacks, hackers can launch replay attacks, where they send arbitrary instructions to the targeted device in the form of legitimate data. Specially crafted HTTP packets containing watering plan updates are sent to the system so that the sprinklers are activated as specified by the attacker.

Replay attacks can also be used to open the valves of smart irrigation systems and initiate the watering process whenever the hacker wishes. In their experiments, researchers got the master valve of a system to open and close every ten seconds.

Launching malicious attacks against irrigation systems can have a significant impact on water utilities and their customers, the researchers warned.

For example, threat actors can activate sprinklers and keep them running until areal reservoirs and water tanks have been emptied, which can result in temporary water outages or at least a reduction of the water flow. This can be particularly problematic in regions where there is a shortage of water.

Based on the calculations performed by the researchers, a botnet of roughly 1,300 irrigation systems can empty a standard water tower in an hour. A larger botnet of nearly 24,000 sprinklers can empty an anti-flood water reservoir overnight.

Increasing the water consumption also leads to financial damage, which can be significant, especially in areas where water is expensive.

Each type of attack was demonstrated against one of the targeted products. All impacted vendors, including one weather forecast service abused in the tests, were notified in June and some of them have started implementing measures to prevent potential attacks.