Written by Hussam Sidani, the Regional Vice President for the Middle East and Turkey at Cybereason

How secure is your organization if you can only stop attacks that have already been detected in other environments based on Indicators of Compromise (IOCs)? Secure enough, if those were the only attacks you needed to be concerned with. But what about targeted attacks with bespoke tactics, techniques, and procedures (TTPs) that have never been documented because they were designed only to be used against your organization?

In today’s threat landscape that’s what’s happening: zero-day exploits, never-before-seen malware strains, and advanced techniques developed specifically for high-value targets are plaguing security teams. Most security solutions do a pretty good job of detecting and preventing known threats, but they continue to struggle with detecting and preventing novel threats. But the issue run even deeper than that — how can security teams detect malicious activity on the network earlier if the actions and activities of the attacker are not outwardly malicious because they are typical of activity we expect to see on a network?

The diminishing value of IOCs
Following a security incident, investigators scour for the evidence and artifacts left behind by the attackers. These can include IP addresses, domain names, file hashes, and more. Once these Indicators of Compromise (IOCs) have been documented, they can be shared so that security teams at other organizations can search their environments for similar threats, and security solutions can be tuned to better detect and prevent them from being used in subsequent attacks. That’s great for everyone, except the initial victims of the attacks, of course — for them, the damage has already been done.

Bur IOCs are constantly changing and more often are unique to a specific target, so leveraging IOCs for proactive defense in another environment is unlikely to result in earlier detections. Even the assumption that IOCs are somehow uniformly applicable in every instance, for a given attack campaign in the same environment, has proven to be demonstrably false.

Furthermore, the more advanced attackers engaged with a high-value target often change their TTPs within the same kill chain when moving from one device to the next in a target environment, making early detection based on already-known IOCs nearly impossible. IOCs are still quite valuable for detecting known TTPs, just as outmoded signature-based detections are still effective for detecting common malware strains, and they will continue to be an important aspect of our security toolkits for the foreseeable future.

But given the limitations of their application in surfacing highly targeted and novel attacks as described above, the question remains as to how we can detect more reliably and earlier in the kill chain. That’s where Indicators of Behavior (IOBs) come into play.

Defining Indicators of Behaviour
IOBs describe the subtle chains of malicious activity derived from correlating enriched telemetry from across all network assets. Unlike backward-looking IOCs, IOBs offer a proactive means to leverage real-time telemetry to identify attack activity earlier, and they offer more longevity value than IOCs have ever been able to deliver.

IOBs describe the approach that malicious actors take over the course of an attack. They are based on chains of behavior that can reveal an attack at its earliest stages, which is why they are so powerful in detecting novel and highly targeted operations. Sooner or later, an attacker’s path diverges from the paths of benign actors.

But IOBs is not about just looking for anomalies or a key indicator of malice at a particular moment in time, although that’s also part of it. IOBs are about highlighting the attacker’s trajectory and intentions through analysing chains of behaviors that, when examined together, are malicious and stand out from the background of benign behaviors on the network.

IOBs can also be leveraged to detect the earliest signs of an attack in progress that are comprised of “normal activity” one would expect to see occurring on a network, such as we see with techniques like living off the land (LotL/LOLBin) attacks where legitimate tools, processes, and binaries native to the network are abused by the attacker.

Operationalising IOBs for Operation-Centric security
Today’s alert-centric approach to security puts too much focus on the generation of uncorrelated alerts and remediating the individual elements of the larger attack campaign; a process that has proven to be inefficient given the typical resource constraints security operations are subject to.

Conversely, an Operation-Centric approach leveraging IOBs can reorient the detection and response cycle by consolidating otherwise disparate alerts into a single, content-rich correlated detection that serves to comprehensively disrupt the attack progression earlier than is possible with our current reliance on IOCs alone.

Leveraging IOBs to achieve an Operation-Centric approach also presents the opportunity to create a repository of detectable behavior chains that can surface even the most novel of attacks earlier, as well as support automated response playbooks that can better disrupt attacks at their onset.

More work to be done
Understanding attacker intentions and likely pathways based on early-stage actions and activities enable defenders to proactively predict and disrupt subsequent stages of an attack, as well as provides an avenue to develop fully autonomous security operations. In order to achieve a truly Operation-Centric posture and move closer to autonomous security operations, a future-ready standard that universally defines and operationalizes IOBs is required.

To be truly useful, there needs to be a common definition, language, and expression of IOBs that is completely independent of any particular security tool or vendor. The wide array of solutions available can provide the raw telemetry as well as the color and context required to collectively interpret observable behaviours.

But, as it stands today, security tools themselves don’t provide a standardized language that can accurately describe and operationalise the chains of behavior that will enable us to detect and respond to attacks faster than the adversary can adapt. Operationalising IOBs will require standardization that will deliver the full potential value of the entire security stack to quickly and autonomously deliver the necessary context and correlations across diverse telemetry sources.

But achieving an Operation-Centric approach that leverages IOBs will ultimately empower security operations to predictively respond to changing TTPs more swiftly than attackers can modify and adjust them to circumvent defenses, which is key to finally reversing the adversary advantage and returning the high ground to the Defenders.