-
Notifications
You must be signed in to change notification settings - Fork 4
Description
The functionalities of EDR solutions vary so much between vendors, and the overall list of functionalities is growing so numerous, that I think it would be best to split EDR into two (or more) Actuator Profiles.
The way I see it, EDR can be generalized into five main categories:
- Detection
- Detection of anomalous/malicious behavior by running system events through detection engines
- Updating the signature set(s) of the detection engine with new/custom signatures
- Response
- Enforcing policy (proactive response)
- Blacklisting hashes, IPs, domains, disallowing external drives, limiting app execution
- Acting on endpoints (retroactive response)
- Port isolation, killing processes, initiating AV scans, quarantining binaries
- Enforcing policy (proactive response)
- Asset management
- Information regarding hosts and users in an organization
- Data storage and querying
- Retaining system events and asset information in a database
- Querying the data using filters, conditions, logical operators
- Security posture assessment (few solutions have this, sometimes called XDR)
- Host software inventories and OS information, aggregated with known vulnerabilities and TTPs for security recommendations
As it stands, we have the detection part covered in the AP. But as mentioned in Issue #26, facilitating queries that contain criteria and a high level of granularity will be a lot of work, and it will bloat the current AP.
So my suggestions are:
- Rename this AP to Endpoint Response, review it, and move to have version 1.0 published
- Make a new AP that focuses on advanced queries in EDR systems, @Vasileios-Mavroeidis and I have been calling it "analytics"
- Review which of the five categories other than Response fit under this new AP
In regards to suggestion 3, I think we could make a case for putting all the remaining categories under a "endpoint analytics" AP, as all categories other than Detection pertains to querying for data in some way. When it comes to detection, the signature sets of EDR systems are usually the same queries that one use for threat hunting, and so adding a "Update<signature/advanced query>" command would be all we need.
I also think this would be a good time to look at integrating/endorsing the Open Cybersecurity Alliance STIX-shifter and Kestrel projects with OpenC2, as discussed somewhat in issue #26.