Presentation is loading. Please wait.

Presentation is loading. Please wait.

1 Next Generation Kernel Activity Monitoring Edward Balas, Indiana University Michael Davis, Savid Technologies IU Partners in Crime: Camilo Viecco Gregory.

Similar presentations


Presentation on theme: "1 Next Generation Kernel Activity Monitoring Edward Balas, Indiana University Michael Davis, Savid Technologies IU Partners in Crime: Camilo Viecco Gregory."— Presentation transcript:

1 1 Next Generation Kernel Activity Monitoring Edward Balas, Indiana University Michael Davis, Savid Technologies IU Partners in Crime: Camilo Viecco Gregory Travis

2 2 Agenda  Introduction to Kernel based activity monitoring (Sebek as context)  Performance evaluation  A possible solution  Revaluate performance  Roadmap

3 3 Kernel based Activity Monitoring Motivation  Goal is to observe activity while minimizing disruptions.  Network data analysis is no longer sufficient Intruders started using session encryption  Desire to study intra-system behavior Process vs User behavior

4 4 Kernel Activity Monitoring Basics  Basic tasks Observe system call activity Record Data Export Data

5 5 Kernel Activity Monitoring Basics II  Try to make it hard to detect Fail, Try, Fail… Notice that nobody bothers looking  Implemented as Loadable Module or kernel patch  Lets not call it a Rootkit, but a system enhancement.

6 6 Sebek as Kernel Activity Monitor  Linux developed by Balas  Win32 developed by Davis  Latest versions collect Keystrokes / read data Process heritage Files opened by a process Sockets opened by a process  GenIII Honeynet design based in part on this type of data.

7 7 Illustration of Data’s Value

8 8 Illustration of Data’s value

9 9 Sebek as Firehose  Sebek defies one of the claimed values of honeynets “They only collect data of value”  Rudimentary of control on what it collects Makes more analysis work Limits use in operations Provides potential for detection

10 10 Delay imposed by Sebek  Micro RDTSC based, Macro “dd” based  1 bytes Reads from /dev/zero 1,000,000 times  Linux 2.6 Sebek

11 11 Quantity of “uninteresting” data  Per hour record generation rates: Idle Keystroke only ~5,212 Idle Full Read ~98,000 Slowly surfing porn w/ Full Read ~750,000

12 12 What is the problem?  Delay could be used as basis for detection Create per CPU profiles determine if test data exceeds delay threshold Generalize distribution and look for deviations in curve shape  Unintended data capture/sensitive data  Data Volume

13 13 Obvious Solution: don’t record uninteresting  Depending on use case, a prior sense of what is of interest is present Is /dev/zero ever of interest? If we are watching for remote connections, do we care about unrelated local activity  Recall that recent version of Sebek records process tree, sockets and file activity…  Applicable for any Kernel-based Activity Monitoring

14 14 Filtering approach  Make it feel like firewall filters  Make decisions based on Process name Socket parameters File names User names  Use process tree to make filters dynamic

15 15 Example: Monitor a remote user  Action=keystrokes user=ebalas sock=(proto=tcp local_port=22) opt=(follow_child_proc ) Keystroke monitor ebalas when he logs in remotely via ssh. Also monitor any processes decendant from the process that serviced the socket.

16 16 More Examples  action=ignore file=(name=/var strict inc_subdirs) Ignore any activity associated with files in the /var sub directory  action=keystrokes user=bob file(name=/var/sekrit/squirel.txt opt=(follow_child_proc) Silly honeytoken.

17 17 Okam prototype  Other Kernel Activity Monitor(OKAM)  Based on Sebek  Binary flags are added to inode and process data structure to control if we record.  Tagging occurs when A process forks File is opened Socket activity is observed

18 18 Filtered Performance  Volume of data at idle reduced dramatically  Performance when not recording similar to stock  Depends on good filter selection for improvement but control is good.

19 19 Future Direction  Benchmark other monitored system calls  Explore HTB / fair queuing as a way to prevent local process level logging Dos.  Release Okam or integrate into Sebek? Trademarks,Copyright Oh my! Hopefully have something out in May


Download ppt "1 Next Generation Kernel Activity Monitoring Edward Balas, Indiana University Michael Davis, Savid Technologies IU Partners in Crime: Camilo Viecco Gregory."

Similar presentations


Ads by Google