What Can CAL™ (Collective Analytics Layer) Do For You?

CAL™ has billions of data points that it can bring to bear to power its analytics—and we’re adding more every day

It’s been two years since we announced CAL™ , our Collective Analytics Layer. Since then, we’ve made fantastic strides in leveraging some of the latest big data technology to make our users’ lives easier. We launched CAL with two intentions: to solve problems that were so computationally intensive that they required separate technology, and to distribute the answer rather than the solution. Fast forward to today. CAL has billions of data points that it can bring to bear to power its analytics—and we’re adding more every day.

In this article, we’ll cover the insights that CAL provides and then go deeper into how to use that intelligence in your day-to-day analysis, with instructions for both novice and advanced users.

What can CAL do for me?

At this time, CAL provides insights in three main forms: reputation, Classifiers, and contextual fields. They can each be used to help your analysts and orchestration processes make better decisions faster. Let’s take a look at each of these insights to see how they may make your day easier. Figure 1 shows an example of CAL Insights on the Indicator Analytics card on the Details screen for an Indicator.

Figure 1

Reputation

CAL generates its own reputation score on a 0–1000 scale, similar to the ThreatAssess algorithm. We have a writeup on how ThreatAssess and CAL play together but the takeaway is pretty simple: CAL uses its massive data set and our analytics to help provide a baseline reputation score. There are a few things to note about CAL’s reputation analytics:

  1. They are not actually presented in the ThreatConnect Platform. Let’s face it, information overload can be a very real thing. To simplify the user experience, we’ve designed ThreatAssess to combine CAL’s opinion with those of your analysts and tailored processes. In other words, CAL’s reputation score is factored  into an object’s ThreatAssess score. Of course, this calculation is customizable: System Administrators can configure ThreatAssess to weigh CAL’s opinion a lot, a little, or not at all.
  2. Reputation scoring isn’t one size fits all. There are elements of relevance and risk to your organization. Our goal with CAL’s reputation algorithm is to provide the best baseline that we can for Indicators.
  3. Our reputation score is based on lots of data. CAL manages the dynamic collection, curation, and aggregation of lots of data that you simply don’t want to do yourself. It pulls in massive whitelists to help clear the noise out of your workflow. CAL also aggregates all of the reported observations on Indicators to prioritize the threats that are active now.
  4. Reputation goes beyond the score you see on the Indicator Analytics card when you view the Details screen for an Indicator. CAL also uses ThreatConnect’s Indicator Status system to help you maintain uninteresting IOCs for the sake of thoroughness without having them inundate you with false alarms.
Making Use of Reputation

If you’re participating in CAL, it’s already making your life easier! Still, here are a few things to consider if you want to step up your game by incorporating CAL’s reputation insights into your workflow:

Simple

If you have sufficient permissions, you can leverage CAL’s score by weighing it more heavily in the ThreatAssess configuration page, as detailed in the “Configuring ThreatAssess” section of the ThreatConnect Account Administration User Guide. For developing teams, this is extremely helpful while you start to marshall your intelligence processes: CAL can give you some kind of score to start with, and you can focus on triaging the universally critical threats before worrying about creating intelligence of your own.

Intermediate

The Indicator Status feature (circled at the top of Figure 2) gives you a way to remove a lot of noise from your system. Again, if you’re participating in CAL, you’re already leveraging its insights on hundreds of millions of Indicators! Keep the CAL Status Lock box unchecked to let CAL set the flag on whether each Indicator should be enabled or disabled as far as piping it to your integrations. Of course,  you can adapt your processes to manually set (and lock) that flag for Indicators that you know are or aren’t of interest.

Figure 2

 

Advanced

To empower our security ninjas, we do actually expose the CAL score via Playbooks. You can build your Playbooks to bin your Indicators (or avoid creating them at all) based on CAL’s score right off the bat. This can be especially helpful when it comes to removing noise from your system or firing off alerts and triage workflows. If CAL has decided something is universally good or bad, take that step out of the equation for your analysts!

Classifiers

Our analytics apply a series of labels called Classifiers to Indicators that meet certain conditions. These labels are designed to give you a clear, concise vocabulary to understand some of the salient data points about an Indicator. The Classifiers are similar to Tags in ThreatConnect, except that they’re applied by CAL using the totality of its data set and statistical models. Figure 1 shows some Classifiers in the Classification section of the Indicator Analytics card.

As we add more data collection and analytical models, we will continue to expand our vocabulary of Classifiers and fine-tune the conditions that apply them.

Using Classifiers

If you’re not sure how you’d use Classifiers in your day-to-day processes, here are some examples:

Simple

Something as simple as the Executable.Android or Executable.iOS Classifier may help you quickly identify binaries that run on platforms that are outside of your area of responsibility. If your organization doesn’t use Android or iOS devices, then you can easily move along!

Intermediate

If you stumble across a host that CAL identifies as having the IntrusionPhase.C2.Current Classifier, then you may have an active breach on your hands! These Indicators have been classified based on the findings of the ThreatConnect Research team, and you can head on over to the ThreatConnect Intelligence Source to learn more about the associated Threat to determine your next steps.

Advanced

CAL’s DNS monitoring system can let you know about the resolution patterns of certain hosts. If you see an IP address with the DNSHosts.Malicious.Current Classifier, then you can follow it in ThreatConnect and you’ll get notifications when additional hosts in the system start resolving to it.

Contextual Fields

CAL also provides a series of contextual fields surrounding an Indicator to help you decide what to do next. These fields may come from a variety of sources:

  • Aggregated, anonymized data. CAL takes telemetry information from all of our participating instances and aggregates it after removing any identifying information. This allows CAL to provide global counts on key data points, such as how many observations have been reported on the Indicator or how many false positive votes it’s gotten and when.
  • Enrichment data. To power its analytics, CAL has access to all sorts of data it’s collected. We want you to have this information, too! You may get certain information as appropriate, such as where a hostname is ranked in the Alexa Top 1 Million domains list or what OSINT feeds reported it and when.

Making use of Contextual Fields

If you’re not sure how you’d use these contextual fields in your day-to-day processes, here are some examples:

Simple

If you’re looking at an IOC that has a high score, but has a lot of false-positive votes, you may have stumbled into the twilight zone of bad intel! It happens sometimes—our feeds and partners occasionally let benign Indicators slip into our discussions. Sometimes Indicators were bad and then get their act together and get clean. CAL’s global false-positive data can help you better isolate bad Indicators that have gone good.

Intermediate

If an IOC in question has a high number of global observations, then it may be active across the ThreatConnect user base. You may be able to identify the ebb and flow of adversary activity before you’re in the adversary’s sights, benefitting from the anonymized reporting of your peers. Trendline data can help you pinpoint where in time to look if you’re doing retroactive analysis as well.

Advanced

If you’re triaging phishing emails and see an unknown SMTP server, CAL may be able to tell you that it’s owned by Google and is part of the GSuite Mail Server. Understanding who owns infrastructure—specifically free or rented infrastructure—can help you quickly determine your next steps. Whether you’re picking up the phone to request a takedown from a hosting provider or simply blacklisting an IP address, these are the insights that start to make a difference at scale.

In Conclusion

CAL has come a long way in making sure that we are answering questions our users have about intelligence, sometimes before they even know to ask about it. By combining our unique data set and domain expertise, we’re starting to discover novel Indicators at a high rate and a high confidence level. By leveraging CAL, you can, too.

Keep in mind that these CAL insights aren’t just available to your human analysts, but also to your Playbooks! Stay tuned as we start to showcase ways that CAL can drive your orchestration processes automatically, using its reputation and classification analytics to help you move faster and smarter.

 

 

About the Author
Drew Gidwani

Drew Gidwani is the Director of Analytics at ThreatConnect. He drives the data modeling, collection, and analytics both within the core ThreatConnect platform and in CAL. Previously, Drew worked for the Department of Defense where he leveraged his varied analysis experiences to scale growing intelligence teams in the face of the ever-changing threats we face today. Drew holds a B.S. from Carnegie Mellon University and an M.S. from Johns Hopkins University. He currently resides in Maryland with his fierce warrior dog named Gimli.