US critical infrastructure providers are failing to implement adequate information security measures, according to the US Industrial Control System Cyber Emergency Response Team (ISC-CERT).
The security team, set up in 2009 under the Department of Homeland Security, received 198 security incident reports in 2011, a massive increase in from the 41 it received in 2010.
Despite the surge in incident reports, ICS-CERT's investigations uncovered just seven instances in 2011 where an organisation suffered an intrusion, which was actually down from the eight it confirmed in 2010.
The report suggests there is an increased willingness within critical infrastructure to submit a security incident report to the ICS. In its inaugural year ICS-CERT received just nine reports in its first year, and discovered two intrusions.
ICS-Cert’s report canvasses confirmed intrusions energy, nuclear, manufacturing, water and chemical plants.
The year 2010 proved to be interesting for the response team. Responding to a report at one unnamed manufacturer, the team discovered Stuxnet on “all their engineering workstations” and several machines connected to its manufacturing control systems network.
The same year ICS-CERT investigated a Mariposa botnet outbreak on a nuclear facility’s enterprise network and discovered it occurred after an employee inserted an infected USB into their laptop which they acquired from an instructor at an industry event.
The team discovered the USB had been shared by other nuclear industry personnel but when it identified the instructor and requested he provide a list of attendees, he declined the offer, saying he would contact attendees himself.
ICS-CERT said it was unable to verify if the potentially impacted companies were ever notified.
The report includes a number of cases where the compromised provider was the victim of a targeted attack.
Seven out of 17 incidents (where it confirmed an intrusion) were the result of a spear-phishing attack, while 11 cases involved “sophisticated threat actors”. However, it notes that “no intrusions were identified directly into control system networks”, whereas in “several” cases the attackers had exfiltrated data.
In 12 out of the 17 cases, ICS-CERT said login limitations and network segmentation with properly configured firewalls could have prevented or minimised the impact of the attacks.
Ten organisations could have detected the intrusion had they implemented ingress and egress IP and domain filtering.
A common problem amongst the organisations where it has investigated an incident was insufficient logging capabilities, leaving them unable to provide log data for analysis.