Dishwashers on the rampage: LG IoT security bug highlights risks of home automation

LG’s SmartThinQ is a framework for communication between devices

The discovery this week of a security vulnerability within SmartThinQ, a technology touted by LG for automating communication with its range of home appliances and devices, has reinforced the risks of remote Internet of Things (IoT) takeover as attackers progressively master new methods of attacking increasingly smart devices.

LG’s SmartThinQ is a framework for communication between devices that enables them to, among other things, be controlled by smartphone apps or by voice through integration with emerging smart-home devices such as Amazon Alexa and Google Home. It has been built into a range of LG refrigerators, ovens, dishwashers, air purifiers, washing machines, dryers, and robotic vacuums.

Weaknesses in the core software, which were named ‘HomeHack’ and disclosed by Check Point Software Technologies researchers to LG on 31 July, allowed an attacker to create a fake LG account, then use this to take over the account of a legitimate user that would provide access to all of their appliances.

This presents, among other things, security concerns that the remote-monitor camera on automatic vacuums could be used to surveil a target home or office.

The risk of poor IoT security isn’t limited to renegade vacuums, however: even as home video cameras and other appliances are relentlessly probed for vulnerabilities, the risk of vulnerabilities in medical devices or industrial-control sensors – which could potentially cause injury or death – have let to prioritisation of the need for urgent improvements as IoT reshapes the future of device security.

LG Electronics responded quickly to the Check Point alert, remotely updating what LG manager of smart development team Koonseok Lee said in a statement was “an advanced rooting process designed to detect security issues”. The patched and updated SmartThinQ kernel has been running “smoothly and issue-free” since then, Lee said.

This seemingly smooth patch process belies the experience of most IoT device makers, which have been notoriously lax on security in their devices. The issues are compounded as increasing development of IoT standards like MQTT improves inter-device communications and opens new channels for wholesale takeover of large numbers of devices.

“Internet connected IoT devices are now the attacker’s choice for launching DDoS attacks,” Arbor Networks Steinthor Bjarnason noted in a recent blog that blamed the devices’ limited storage, functionality and cost pressures for their endemic insecurity.

That insecurity had made IoT devices sitting ducks for attacks such as last year’s Mirai, an offshoot called Persirai capable of infecting 1250 different models of security camera; and a Windows Trojan called Windows Mirai Spreader, detected in February, that not only infected Windows computers but used them to scan corporate networks for vulnerable IoT devices.

A large-scale takeover would allow attackers to generate massive volumes of DDoS data using a large enterprise’s significant networking resources. These attacks might be directed at external targets or, Bjarnason warned, used to launch “devastating internally facing DDoS attacks against vulnerable internal resources including data centres and WAN/LAN network infrastructures. These resources are, in almost all cases, NOT protected against DDoS attacks originating from the inside.”

Earlier this year, Trend Micro’s 2017 Midyear Security Roundup warned that IoT threats like Persirai were only the tip of the iceberg, with formal tests of industrial robots showing that they were vulnerable to remote takeover. To fix the problem, developers, operators, and security firms should “make attacks on industrial robots expensive to the point of being impractical”, the firm warned while advising “stringent software engineering practices to improve code robustness, harden underlying platforms, and implement strong authentication resources”.

Regulators are working to improve confidence in IoT security, as with an EU IoT security labelling scheme and European Cybersecurity Agency whose remit includes IoT. Other security specialists report IoT vendors trying, at least, to improve their security practices with the help of outside consultants.

Yet even as some diligent vendors review their security practices with an eye to facilitating in-place upgrades – as LG was able to do to defeat HomeHack – many others are continuing to push out devices that are insecure by design and function. Nascent SecDevOps programs promise to improve overall security practices at the developer level, but time and delivery-pressured developers have proven slow to change.

Show Comments