Security experts use advanced data mining and artificial intelligence techniques to perform deep analysis on several years of cyberattacks in Japan, South Korea, the United States and Europe – and were able to learn a great deal about the attackers and their sponsors
When the top minds at a security company leverage their talents and AI tools to dig deep into more than five years of persistent, multi-attack campaigns against industrial and government interests, patterns emerge. Scary patterns.
Like, for example, that many of the attacks are linked – despite their originating at different IP addresses. Despite the wide variations in targets, from infrastructure concerns like oil and gas, electric utilities, finance, transportation and construction. And despite the attacks growing in sophistication over time.
The research was conducted by SPEAR, the security research team employed by Cylance, a Southern California company focused on the use of AI techniques to proactively protect against persistent threats and malware. In February 2016, Cylance was positioned in the Visionaries quadrant of the latest Gartner Magic Quadrant report for endpoint protection platforms.
The SPEAR team has released a new report that details the extensive multi-year initiative to gather data about cyberattacks, and subject them to rigorous analysis. Called “Operation Dust Storm,” the report, written by Cylance’s Director of Threat Intelligence, Jon Gross, explains that many — if not all — of these attacked were either conducted by the same hackers, or supported by the same group of backers. (SPEAR means “Sophisticated Penetration, Exploitation, Analysis and Response.”)
“Attack telemetry in 2015 indicates the Dust Storm group has migrated from more traditional government and defense-related intelligence targets to exclusively seek out organisations involved in Japanese critical infrastructure and resources” the report explains, using the phrase “Dust Storm” to describe the threat group. The report continues, “SPEAR’s current research indicates the group’s present focus has shifted to specifically and exclusively target Japanese companies or Japanese subdivisions of larger foreign organisations.”
Many of the initial hacks exploited a number of weaknesses in Microsoft Windows. For example, an attack in April 2011 is described this way: “The attack was initiated by a spear phishing email that contained a Word document embedded with a zero-day Flash exploit (CVE-2011-0611).”
As time went on, the nature of the attacks changed, as in one attack a few years later: “Beginning in February 2014, there was definitive evidence to suggest the group used a watering hole attack on a popular software reseller to deliver an Internet Explorer zero-day, CVE-2014-0322, to a number of unsuspecting targets."
Why does SPEAR believe that these attacks are linked? As the report explains, the evidence began to be compelling: “Activity in 2015 was significantly more interesting, and prompted SPEAR to begin studying Operation Dust Storm’s other activities. SPEAR identified a number of second-stage backdoors with hardcoded proxy addresses and credentials. These proxy addresses revealed the attacker had compromised a number of Japanese companies involved in power generation, oil and natural gas, construction, finance, and transportation.”
A common factor appeared to be an archaic set of Microsoft software development tools used by the hackers to program backdoors. “The second-stage implants were also programmed and compiled using Microsoft Visual Studio 6; an archaic version of Visual Studio that seems to be preferred by malware authors. Despite using an old version of Visual Studio, the backdoor is well designed by comparison and provides a full suite of functionality to the attack.”
There were additional common factors as well. Many of the exploits studied by the SPEAR team linked back to a small group of remote servers, accessed by IP address or by a relatively small group of fully qualified URIs. The command language that the hackers could use to communicate with compromised systems, such as to download data, was consistent across this time period, even as the attack vectors themselves changed.
Even as the attacks progressed, there were clear signs that the cybercriminals were specifically looking at Japanese companies, as well as the Japanese subsidiaries of foreign companies based in South Korea. For example, many of the backdoors would attempt to detect whether or not the victim was using a Japanese keyboard via a call to the Windows API “GetKeyboardType.” There was some evidence to suggest Operation Dust Storm leveraged an Ichitaro zero-day “CVE-2013- 5990” exploit to target Japanese victims. Ichitaro is a popular Japanese word processing program designed by a company called JustSystems.
Evolution of Hacking Prowess
The initial samples of vulnerability exploits and backdoor code recovered by the SPEAR team were extremely blunt, unsophisticated and easy for the security industry to detect. That was in 2010. Later, the code became trickier to detect as the cybercriminals gained skills. Starting in 2011, the hackers took advantage of vulnerabilities in Internet Explorer 8, Yahoo, Windows Live, and Flash.
That’s not all. The hackers evolved how they broke into systems. At first, they targeted their victims via spear-phishing – that is, sending target emails to employees of specific companies or government agencies. The emails appeared credible – but opening links in those messages either directly installed malware, redirected the end user to a website that installed malware, or opened documents that were themselves carriers of malware.
As the attackers became more savvy, they also deployed watering holes, where the hackers take over websites commonly used and trusted by their intended victims, and turn them into malware delivery systems. As industry workers visited those websites — which it would be perfectly appropriate to do – their computers became infected.
And then there’s Android, which the cyberhackers targeted for backdoors and data theft.
According to the report: “Perhaps even more interesting was the fact that the group adopted and eventually customised several Android backdoors to suit their purposes in the beginning of 2015. The group rapidly expanded their mobile operations in May 2015. The initial backdoors were relatively simple, and would continually forward all SMS messages and call information back to the C2 servers. Later variants became much more complex, and included the ability to enumerate and exfiltrate specific files directly from the infected devices. All of the identified victims for the Android Trojans resided in Japan or South Korea. The infrastructure to support the Android campaigns was massive in comparison to previous operations. More than two hundred domains have been identified to date.”