Collective Threat Analysis FAQ
What is collective threat analysis?
Collective threat analysis enables users to share select data with ExtraHop to improve the accuracy of detections, such as Command-and-Control (C&C) Beaconing.
By default, any data sent to the ExtraHop Cloud Service that might uniquely identify a network participant (such as an IP address or username) is encrypted with a key that is stored on the sensor and to which ExtraHop has no access.
Reveal(x) Enterprise users can send data to the Machine Learning Service by enabling ExtraHop Cloud Services in the Administration settings. For example, the system can send external plaintext IP addresses, domain names, and hostnames that are associated with detected suspicious behavior. This setting is enabled in Reveal(x)360 by default and can not be disabled. For a full list of data types sent to the ExtraHop Machine Learning Service, and to see how the data is applied to improve threat detection, see the Machine Learning section of the ExtraHop Security, Privacy and Trust Overview.
By opting in to share this plaintext data, you contribute to a large community dataset that can be analyzed for everyone's benefit—especially your own. This dataset includes both plaintext data and de-identified metadata associated with threats detected by ExtraHop.
How secure is my data?
When you opt-in to send ExtraHop the external plaintext IP addresses, hostnames, and domain names observed on your network, the ExtraHop sensor sends this metadata to the Machine Learning Service through TLS 1.2 or TLS 1.3 connections and perfect forward secrecy (PFS). Both data in transit and data at rest is stored securely in an encrypted highly-protected datastore.
You can learn more about how ExtraHop secures your data in the ExtraHop Security, Privacy, and Trust Overview.
Why should I opt-in?
Here are the ways that you benefit from contributing to collective research and analysis.
- Improve context about your detections
- ExtraHop cloud-based machine learning can take advantage of plaintext data
when analyzing suspicious behavior. Rich data surfaces detections with
For example, take the website of a local coffee shop that has poorly configured web analytics. This website frequently reaches out to an external analytics server with performance statistics. The website traffic might be detected on your network for 30-second rapid beaconing—a behavior that is also commonly observed in malicious command-and-control (C&C) beacons. However, with access to the external plaintext hostname and IP address of the analytics server associated with the detection, the ExtraHop system can better determine whether the rapid beaconing is tied to a known malicious source. Improved context helps ExtraHop tell you when traffic is malicious and reduces false positives.
- Help stop novel attacks on your network
- ExtraHop performs big-data analytics to hunt for stealthy and advanced
attacks that individual organizations might overlook. The entire customer
base is automatically and immediately protected from each newly identified
For example, ExtraHop might observe that devices across multiple networks are establishing reverse SSH tunnels to a suspicious IP address. Upon further analysis, the suspicious IP address appears to be hosting a C&C server that is exhibiting behaviors previously associated with a known threat group. ExtraHop immediately updates all deployed sensors with detections to protect all cloud-connected deployments from the newly identified threat.
- Improve machine-learning models in your detections
- ExtraHop leverages community-sourced data when training machine-learning algorithms and developing new machine-learning models, which are designed to find attacks on user networks. We also refine our understanding of benign behavior patterns by monitoring how behaviors manifest across the networks of different industries, sizes, and geographic locations.
Thank you for your feedback. Can we contact you to ask follow up questions?