Skip to content
This repository has been archived by the owner on Apr 21, 2023. It is now read-only.

Latest commit

 

History

History
73 lines (50 loc) · 4.38 KB

ML_OPS.md

File metadata and controls

73 lines (50 loc) · 4.38 KB
Raw

The ml_ops.sh script

ml_ops.sh is a helper script for invoking the suspicious connects analysis in the common case of its use in Spot.

It loads environment variables from /etc/spot.conf, parses some arguments from the CLI, invokes the Suspicious Connects analysis from the spot-ml jar stored in target/scala-2.10/spot-ml-assembly-1.1.jar.

Usage

The ml_ops script is meant to be used as part of a full spot deployment. Data is expected to reside in the locations used by Spot ingest and in a schema extending the schema used by the suspicious connects analyses.

To run a suspicious connects analysis, execute the ml_ops.sh from the directory where the spot-ml jar is stored in the relative path target/scala-2.10/spot-ml-assembly-1.1.jar

./ml_ops.sh YYYMMDD <type> <suspicion threshold> <max results>

  • YYYYMMDD The date of the collected data that has been ingested and stored into the Spot's data repository. Eg., 20170129

  • type is one is of flow, proxy or dns.

  • suspicion threshold is an optional floating-point value in the range [0,1]. All records that receive a score strictly greater than the threshold are not reported. It defaults to 1 (no filter).

  • max results is an optional integer value greater than or equal to 0. If provided, results are ordered by score from least to greatest and the max results smallest are returned in an ascending list. If not provided, all results with a score below the given threshold are returned in an ascending list.

Example Usage

./ml_ops.sh 19731231 flow 1e-20 200

If the max results returned argument is not provided, all results with scores below the threshold will be returned, for example:

./ml_ops.sh 20150101 dns 1e-4

As the maximum probability of an event is 1, a threshold of 1 can be used to select a fixed number of most suspicious items regardless of their exact scores:

./ml_ops.sh 20150101 proxy 1 2000

ml_ops.sh output

Final results are stored in the following file on HDFS Depending on which data source is analyzed, spot-ml output will be found under the HPATH at one of

 $HPATH/dns/scored_results/YYYYMMDD/scores/dns_results.csv
 $HPATH/proxy/scored_results/YYYYMMDD/scores/results.csv
 $HPATH/flow/scored_results/YYYYMMDD/scores/flow_results.csv

It is a csv file in which network events annotated with estimated probabilities and sorted in ascending order.

Parameters taken from the /etc/spot.conf file

The ml_ops.sh script takes its values for the following parameters from the /etc/spot.conf file:

  • All spark settings Among them driver memory, number of executors, spark.driever.MaxResultSize, etc.
  • Paths to storage locations for Spot ingested data
  • USER_DOMAIN The domain name for the network being analyzed. Used to denote "internal" URLs during proxy and dns analyses.
  • TOPIC_COUNT Number of topics used for the topic modelling at the heart of the Suspicious Connects anomaly detection. Roughly, the analysis attempts to generate TOPIC_COUNT many profiles of common traffic in the cluster.
  • DUPFACTOR Used to downgrade the threat level of records similar to those marked as non-threatening by the feedback function of Spot UI. DUPFACTOR inflate the frequency of such records to make them appear less anomalous. A DUPFACTOR of 1 has no effect, and a DUPFACTOR of 1000 increases the frequency of the connection's pattern by a factor of 1000, increasing its estimated probability accordingly.
  • TOL The default value for the suspicion threshold described above. In particular: If no third argument is provided to ml_ops.sh, the suspicion threshold is filled in with the TOL value from /etc/spot.conf. If a third argument is provided to ml_ops.sh, that is the suspicion threshold used.
  • LDA_OPTIMIZER The LDA implementation to use. Set equal to "em" to execute LDA using EMLDAOptimizer or "online" to use OnlineLDAOptimizer. See LDA Spark documentation for more information.
  • LDA_ALPHA Document concentration. See LDA Spark documentation for more information.
  • LDA_BETA Topic concentration. See LDA Spark documentation for more information.