Crap computers in a crap box: Smart-meter blackouts risk to UK
Sniff a device's wireless, pwn a power plant, warns Brit biz
Analysis You'd be forgiven for thinking this is the plot of a Saturday night BBC2 drama: hackers tinkering with smart electricity meters deliberately cut the power to whole neighbourhoods.
But, according to a UK computer security biz, weak authentication checks and a lack of other security controls on said equipment could allow just that.
Greg Jones, director at security consultancy and penetration testing firm Digital Assurance, discovered numerous shortcomings in three samples of two types of smart meters, which are installed in homes to monitor and control the use of gas and electricity. Typically, these devices are wirelessly connected to the supplier so data and upgrades can be easily transmitted over the air - some even using the mobile phone network.
A hacker could therefore crack the wireless communications between the meter and the supplier, and send his or her own mischievous messages back to base, or commands to other meters, it is claimed.
After buying the test samples through online auction sites, Jones wanted to find out what made the electronics tick. He said he discovered the protection against hardware tampering was far less stringent than that applied to comparable consumer security products, such as an Xbox game console.
"Smart meters are essentially crap computers in a crap box," Jones claimed.
Attention in the past has predominantly focused on how power plants' critical systems can be compromised over traditional wired computer networks.
Little consideration has been given to how they might be hacked wirelessly, it seems. The proprietary protocols used by power equipment can be intercepted and analysed using readily accessible software-defined radio (SDR*) equipment and a PC.
With up to 53 million smart meters (essentially miniature SCADA devices) installed in 30 million homes and businesses in Blighty between 2014 and 2019, the number of potential access points on suppliers' networks is set to increase dramatically.
The data relayed between these devices and their headquarters can be intercepted, captured, jammed or replayed using SDR equipment, providing the hacker with network-wide access to individual home meters, control stations, generating stations and transmission facilities.
Armed with nothing more than a soldering iron and some basic programming, Jones explored how smart meters can be exploited, controlled and manipulated. He found security flaws in both the design and implementation of several devices.
'We extracted all of its passwords'
"The meter is manufactured by a significant vendor who specialises in smart metering. The specific meter is MID/Ofgem certified (can be used in the UK on the grid) and is in use in the UK and extensively abroad," Jones explained. "This meter on which nearly all of the work has been done supports the International Electrotechnical Commission's protocol standards and currently uses the GSM mobile phone network for wide-area network communications."
"We extracted all of its passwords from EEPROM [programmable read-only memory chip] and can use these to communicate with other meters from the meter supplier - and the vendor, as some of these passwords are factory defaults."
Flipping the switch to disconnect or reconnect supplies is "fairly trivial" once you have the super-user password for the device and the necessary connectivity. This could be via several means, including local connection via wired serial, or GSM to run a fake base station attack.
Exploits could include remotely disconnecting a home or office building's power supply; something that could even be applied across an entire neighbourhood.
"Some devices do feature authentication," Jones explained. "But it's a mixed bag and down to the manufacturer."
Components of the electricity grid previously relied upon their relative obscurity to protect them but this is changing with the rollout of smart meters. Defences need to be built into the system, according to Jones.
"The only way of protecting a wireless device from an SDR attack at present is to ensure that it has been designed, configured and deployed to resist over-the-air attacks. Very few vendors of such equipment will give this type of assurance, so independent testing is currently the only option until the industry applies itself to developing a solution. Understanding exactly what radio systems have been deployed and ensuring adequate risk assessments have been conducted is an essential first step,” he explained.
Sponsored: IBM FlashSystem V9000 product guide