Presentation is loading. Please wait.

Presentation is loading. Please wait.

Jeremy Kackley, James Jacobs, Paulus Wahjudi and Jean Gourd.

Similar presentations


Presentation on theme: "Jeremy Kackley, James Jacobs, Paulus Wahjudi and Jean Gourd."— Presentation transcript:

1 Jeremy Kackley, James Jacobs, Paulus Wahjudi and Jean Gourd

2  What are they?  Code that migrates from machine to machine  How are they utilized?  Examples  Searching  Visiting several resources that contain data.  Sorting the data, and combining it into a payload.  Computation done remotely.  Communication  Can also be used to deliver data.

3  Advantages:  Reactive/Adaptive  Reliability  Autonomous  Efficient  Disadvantages  Nontraditional  Lack of Standards  Complexity  Security

4  Trustworthiness  Agent trustworthiness  Sandbox  Fairly good solution  Agency trustworthiness  Encryption  Keep 'payload' secure.  Difficult  Focus of this work.

5  System for monitoring network data for the purpose of detecting compromised resources.  Four threat levels organized by severity  Level 1: Observation  Situation normal  CAN monitors network passively via Probe agent dispatches  Level 2: Investigation  Anomalous data observed by the passive monitoring system.  Actively monitor the anomalous nodes by dispatching team of Commander and Detective agents  Level 3: Confirmation  Active monitoring has also detected anomalies.  Attempt to confirm state of the nodes in question.  Takes the form of a Secret agent  Level 4: Resolution  System has detected compromise.  Attempt to resolve:  Alert Human  “Log” activity but permit  Block activity  Shut down node (DDOS, out of band signal…)

6

7  MAIDs relies upon anomaly detection, what if a node is entirely passive?  Pollination is a scheme to detect passive, ‘mole- like’ attackers.  Inspired by Bee:  Bee’s visiting flowers to get nectar  Incidentally, they gain pollen  They also deposit pollen  Pollen on the bee’s provides a roadmap of where they’ve been

8  Agent Pollination  Agents visit nodes in the course of activities  Agents gain pollen  Against leave pollen behind  Amount of pollen represents the time spent at nodes  Sequence of pollen represents road-map of where the agent has been  Implications  Incorrect or missing sequences are new anomalies and represent ‘issues’ that require investigation  Amount of pollen can represent the types of data an agent is interested in when cross-referenced with the types of data stored at various nodes  Nodes with practically no pollen might indicate a node that has no resources and is sniffing passing agents  Standard inference models can be utilized to generate even more anomalous triggers for MAIDS

9

10  Manipulate Open System Interconnection OSI transport layer by either  Appending additional packets containing pollen information to the sequence representing the agent  Manipulating the packets themselves via packet tagging  Pollination does not need to be active everywhere; can only pollinate ‘sensitive’ nodes and thus track ‘important’ data  Degree of pollination can vary depending on threat level, as can consequences to agents with suspicious pollen patterns  Pollination patterns can be periodically changed to make it more difficult to spoof

11

12

13  Situation normal.  Probes distributed  Record communication.  Do not move.  Agents visit network.  Normal agent behavior.  During this process, they pick up data from the probes.  Central Authority Node  Compares data from the probes as it arrives naturally.  Mines for anomalies.

14  Anomalies detected.  Could be nothing; 'lag.'  Deploy a set of agents  Detective agents  Actively monitor  Commander Agent  Takes information from detective agents and analyzes it for anomalies

15  Anomalies still detected.  Deploy a “Secret Agent”  Designed to appear externally as a regular agent.  Executes predetermined series of actions, reports observed results, if possible.  Detective agents observe the 'actual' results  Commander agent analyzes results  Agency exonerated  Elevation of threat level.

16  Level 4 assumes compromise has occurred  This situation must be resolved.  Possible avenues of resolution:  Human Intervention  Redirect output to a 'vault' for later analysis  Attempt to fool agency into thinking it is still actually part of the network.  Blockade output of node.  Protect the network, and agents, by preventing access to or from the suspected node.  Automated attack on the node.  The appropriate response depends upon the network.

17  Simply ask for human aid.  This can be thought of as raising an alert.  No automated action taken by the system.  This step is implied in all other possible resolutions.

18  “Saves” the output of the node for later analysis.  Limited action against node is taken.  Attempts to obscure the fact that the compromise is detected until a human decides what action to take.

19  This response takes active steps to protect the network by preventing communication with the affected node.  This could itself be detrimental to the network; leading to bottlenecks or failure.

20  If data is of an especially sensitive nature; it might be desirable to attempt to remove the affected device from the network by offensive means.  Again, this could damage the network.


Download ppt "Jeremy Kackley, James Jacobs, Paulus Wahjudi and Jean Gourd."

Similar presentations


Ads by Google