Despite enterprises’ best efforts to shore up their security information and event management (SIEM) postures, most platform implementations have massive gaps in coverage, including missing more than three-quarters of the common techniques that threat actors use to use to deploy ransomware, steal sensitive data, and execute other cyberattacks.
Researchers from CardinalOps analyzed data from production SIEM platforms from companies such as Splunk, Microsoft Sentinel, IBM QRadar, and Sumo Logic, and found that they have detections for just 24% of all MITRE ATT&CK techniques. That means that adversaries can execute about 150 different techniques that can bypass SIEM detection, while only about 50 techniques are spotted, the researchers said.
This is despite the fact that current SIEM systems actually do take in sufficient data to cover potentially 94% of all these techniques, CardinalOps revealed as part of the company’s “Third Annual Report on the State of SIEM Detection Risk,” released today.
Moreover, organizations are largely deluded about their own security postures and “are often unaware of the gap between the theoretical security they assume they have and the actual security they have in practice,” the researchers wrote in the report. This creates “a false impression of their detection posture.”
MITRE ATT&CK is a global knowledge base of adversary tactics and techniques based on real-world observations that’s aimed at helping organizations detect and mitigate cyberattacks. The report’s data is the result of analysis of more than 4,000 detection rules, nearly 1 million log sources and hundreds of unique log source types used in SIEM across a range of diverse industry verticals — including banking and financial services, insurance, manufacturing, energy, and media and telecommunications.
A Lack of SIEM Fine-Tuning to Blame for Detection Fails
The key issue contributing to the current state of SIEM efficacy (or lack thereof) seems to be that even though resources exist for organizations to use knowledge, automation, and other processes to detect adversaries and potential attacks on their environments, they still largely rely on manual and other “error-prone” processes for developing new detections, the researchers noted. This makes it difficult to reduce their backlogs and act quickly to fill gaps in detection.
Indeed, SIEMs themselves “are not magic” and rely on the organizations deploying them to do so correctly and efficiently, notes Mike Parkin, senior technical engineer at Vulcan Cyber, a security-as-a-service provider of enterprise cyber-risk remediation.
“Like most tools, they require fine-tuning to deliver the best results for the environment they’re deployed in,” he says. “These report results imply that many organizations have gotten the basics working but haven’t done the fine-tuning necessary to take their detection, response, and risk management strategies to the next level.”
In addition to the need to scale detection-engineering processes to develop more detections faster, one key issue that seems to be tripping up detection in enterprise SIEM deployments is that on average, they have 12% of rules that are broken, which means they will never file an alert when something is amiss, according to the report.
“This commonly occurs due to ongoing changes in the IT infrastructure, vendor log format changes, and logical or accidental errors in writing a rule,” the researchers noted in the report. “Adversaries can exploit gaps created by broken detections to successfully breach organizations.”
Why MITRE ATT&CK Matters
MITRE ATT&CK, created in 2013, has now “become the standard framework for understanding adversary playbooks and behavior,” the researchers noted. And as threat intelligence has advanced, so has the wealth of knowledge the framework provides, currently describing more than 500 techniques and sub-techniques used by threat groups such as APT28, the Lazarus Group, FIN7, and Lapsus$.
“The biggest innovation introduced by MITRE ATT&CK is that it extends the traditional intrusion kill chain model to go beyond static indicators of compromise (like IP addresses, which attackers can change constantly) to catalog all known adversary playbooks and behaviors (TTPs),” the CardinalOps researchers wrote.
Organizations clearly see the value in using MITRE ATT&CK to help them in their security efforts, with 89% currently using the knowledge base to reduce risk for security-operations use cases — such as determining priorities for detection engineering, applying threat intelligence to alert triage, and gaining a better understanding of adversary TTPs, the researchers noted, citing environment, social, and governance (ESG) research.
However, using the framework to support SIEM efforts and using it well appear to be two very different scenarios, the report found.
Closing the SIEM Gap
There are steps that organizations can take to help close the gap between what a SIEM is capable of in terms of cyberattack detection, and how they currently are using it, researchers and security experts said.
One key strategy would be to scale SIEM detection-engineering processes to develop more detections faster using automation, something that companies already use widely to great effect in “multiple areas of the SOC, such as anomaly detection and incident response,” but not so much in detection, they noted in the report.
“The detection-engineering function remains stubbornly manual and typically dependent on ‘ninjas’ with specialized expertise,” the researchers wrote.
Indeed, having a focus on automation is critical to achieving goals with limited human and financial resources, agrees one security expert.
“This includes expanding automated detection to include Internet of things (IoT) and operational technology (OT) attack vectors, as well as having plans already in place for automated threat remediation,” says John Gallagher, vice president of Viakoo Labs at Viakoo.
One key challenge that organizations continue to face is that the current attack surface — which now includes large numbers of vulnerable network-connected devices as well as the typical enterprise network — has grown well past what the IT organization is currently capable of supporting or managing, Gallagher says.
“To defend and maintain the integrity of those assets requires IT working closely with other parts of the organization to ensure those assets are visible, operational, and secure,” he says.
Indeed, Parkin observes, until organizations can get a clear picture of their threat surfaces, manage their risk, and prioritize events to focus on what matters most, there will be problems.
“We have the tools to make it happen,” he says. “But it can be a challenge to get them deployed and configured for best effect.”