Morpheus

Morpheus SDK

The Morpheus developer framework allows teams to build their own optimized pipelines that address cybersecurity and information security use cases. Morpheus provides development capabilities around dynamic protection, real-time telemetry, adaptive policies, and cyber defenses for detecting and remediating cybersecurity threats.

Getting Started

Prerequisites

• Pascal architecture or better
• NVIDIA driver 450.80.02 or higher
• Docker
• The NVIDIA container toolkit

Installation

Pre-built runtime Docker image

Pre-built Morpheus Docker images can be downloaded from NGC. The runtime image includes pre-installed Morpheus and dependencies:

docker pull nvcr.io/nvidia/morpheus/morpheus:22.06-runtime

There is also a Helm chart for deploying the Morpheus SDK container as a pod into a Kubernetes cluster.

Note: You must log into the NGC public catalog to download the Morpheus image. For more information see here.

Kafka Cluster and Triton Inference Server

Morpheus provides Kubernetes Helm charts for deploying a basic Kafka cluster, a single Triton Inference Server, and an MLflow server. These are also available in NGC.

Configuration

The Morpheus pipeline can be configured in two ways:

1. Manual configuration in Python script.
2. Configuration via the provided CLI (i.e. morpheus)

Starting the Pipeline (via Manual Python Config)

See the ./examples directory for examples on how to configure a pipeline via Python.

Starting the Pipeline (via CLI)

The provided CLI (morpheus) is capable of running the included tools as well as any linear pipeline. Instructions for using the CLI can be queried with:

$ morpheus
Usage: morpheus [OPTIONS] COMMAND [ARGS]...

Options:
  --debug / --no-debug            [default: no-debug]
  --log_level [CRITICAL|FATAL|ERROR|WARN|WARNING|INFO|DEBUG]
                                  Specify the logging level to use.  [default:
                                  WARNING]
  --log_config_file FILE          Config file to use to configure logging. Use
                                  only for advanced situations. Can accept
                                  both JSON and ini style configurations
  --version                       Show the version and exit.  [default: False]
  --help                          Show this message and exit.  [default:
                                  False]

Commands:
  run    Run one of the available pipelines
  tools  Run a utility tool

Each command in the CLI has its own help information. Use morpheus [command] [...sub-command] --help to get instructions for each command and sub command. For example:

$ morpheus run pipeline-nlp inf-triton --help
Configuring Pipeline via CLI
Usage: morpheus run pipeline-nlp inf-triton [OPTIONS]

Options:
  --model_name TEXT               Model name in Triton to send messages to
                                  [required]
  --server_url TEXT               Triton server URL (IP:Port)  [required]
  --force_convert_inputs BOOLEAN  Instructs this stage to forcibly convert all
                                  input types to match what Triton is
                                  expecting. Even if this is set to `False`,
                                  automatic conversion will be done only if
                                  there would be no data loss (i.e. int32 ->
                                  int64).  [default: False]
  --use_shared_memory BOOLEAN     Whether or not to use CUDA Shared IPC Memory
                                  for transferring data to Triton. Using CUDA
                                  IPC reduces network transfer time but
                                  requires that Morpheus and Triton are
                                  located on the same machine  [default:
                                  False]
  --help                          Show this message and exit.  [default:
                                  False]

CLI Stage Configuration

When configuring a pipeline via the CLI, you start with the command morpheus run pipeline and then list the stages in order from start to finish. The order that the commands are placed in will be the order that data flows from start to end. The output of each stage will be linked to the input of the next. For example, to build a simple pipeline that reads from kafka, deserializes messages, serializes them, and then writes to a file, use the following:

$ morpheus run pipeline-nlp from-kafka --input_topic test_pcap deserialize serialize to-file --filename .tmp/temp_out.json

You should see some output similar to:

====Building Pipeline====
Added source: <from-kafka-0; KafkaSourceStage(bootstrap_servers=localhost:9092, input_topic=test_pcap, group_id=custreamz, use_dask=False, poll_interval=10millis)>
  └─> morpheus.MessageMeta
Added stage: <deserialize-1; DeserializeStage()>
  └─ morpheus.MessageMeta -> morpheus.MultiMessage
Added stage: <serialize-2; SerializeStage(include=[], exclude=['^ID$', '^_ts_'], output_type=pandas)>
  └─ morpheus.MultiMessage -> pandas.DataFrame
Added stage: <to-file-3; WriteToFileStage(filename=.tmp/temp_out.json, overwrite=False, file_type=auto)>
  └─ pandas.DataFrame -> pandas.DataFrame
====Building Pipeline Complete!====

This is important because it shows you the order of the stages and the output type of each one. Since some stages cannot accept all types of inputs, Morpheus will report an error if you have configured your pipeline incorrectly. For example, if we run the same command as above but forget the serialize stage, you will see the following:

$ morpheus run pipeline-nlp from-kafka --input_topic test_pcap deserialize to-file --filename .tmp/temp_out.json --overwrite

====Building Pipeline====
Added source: from-kafka -> <class 'cudf.core.dataframe.DataFrame'>
Added stage: deserialize -> <class 'morpheus.pipeline.messages.MultiMessage'>

Traceback (most recent call last):
  File "morpheus/pipeline/pipeline.py", line 228, in build_and_start
    current_stream_and_type = await s.build(current_stream_and_type)
  File "morpheus/pipeline/pipeline.py", line 108, in build
    raise RuntimeError("The {} stage cannot handle input of {}. Accepted input types: {}".format(
RuntimeError: The to-file stage cannot handle input of <class 'morpheus.pipeline.messages.MultiMessage'>. Accepted input types: (typing.List[str],)

This indicates that the to-file stage cannot accept the input type of morpheus.pipeline.messages.MultiMessage. This is because the to-file stage has no idea how to write that class to a file, it only knows how to write strings. To ensure you have a valid pipeline, look at the Accepted input types: (typing.List[str],) portion of the message. This indicates you need a stage that converts from the output type of the deserialize stage, morpheus.pipeline.messages.MultiMessage, to typing.List[str], which is exactly what the serialize stage does.

Pipeline Stages

A complete list of the pipeline stages will be added in the future. For now, you can query the available stages for each pipeline type via:

$ morpheus run pipeline-nlp --help
Usage: morpheus run pipeline-nlp [OPTIONS] COMMAND1 [ARGS]... [COMMAND2
                               [ARGS]...]...

<Help Paragraph Omitted>

Commands:
  add-class     Add detected classifications to each message
  add-scores    Add probability scores to each message
  buffer        (Deprecated) Buffer results
  delay         (Deprecated) Delay results for a certain duration
  deserialize   Deserialize source data from JSON.
  dropna        Drop null data entries from a DataFrame
  filter        Filter message by a classification threshold
  from-file     Load messages from a file
  from-kafka    Load messages from a Kafka cluster
  gen-viz       (Deprecated) Write out vizualization data frames
  inf-identity  Perform a no-op inference for testing
  inf-pytorch   Perform inference with PyTorch
  inf-triton    Perform inference with Triton
  mlflow-drift  Report model drift statistics to ML Flow
  monitor       Display throughput numbers at a specific point in the pipeline
  preprocess    Convert messages to tokens
  serialize     Serializes messages into a text format
  to-file       Write all messages to a file
  to-kafka      Write all messages to a Kafka cluster
  validate      Validates pipeline output against an expected output

And for the FIL pipeline:

$ morpheus run pipeline-fil --help
Usage: morpheus run pipeline-fil [OPTIONS] COMMAND1 [ARGS]... [COMMAND2
                               [ARGS]...]...

<Help Paragraph Omitted>

Commands:
  add-class     Add detected classifications to each message
  add-scores    Add probability scores to each message
  buffer        (Deprecated) Buffer results
  delay         (Deprecated) Delay results for a certain duration
  deserialize   Deserialize source data from JSON.
  dropna        Drop null data entries from a DataFrame
  filter        Filter message by a classification threshold
  from-file     Load messages from a file
  from-kafka    Load messages from a Kafka cluster
  inf-identity  Perform a no-op inference for testing
  inf-pytorch   Perform inference with PyTorch
  inf-triton    Perform inference with Triton
  mlflow-drift  Report model drift statistics to ML Flow
  monitor       Display throughput numbers at a specific point in the pipeline
  preprocess    Convert messages to tokens
  serialize     Serializes messages into a text format
  to-file       Write all messages to a file
  to-kafka      Write all messages to a Kafka cluster
  validate      Validates pipeline output against an expected output

And for AE pipeline:

$ morpheus run pipeline-fil --help
Usage: morpheus run pipeline-fil [OPTIONS] COMMAND1 [ARGS]... [COMMAND2
                               [ARGS]...]...

<Help Paragraph Omitted>

Commands:
  add-class        Add detected classifications to each message
  add-scores       Add probability scores to each message
  buffer           (Deprecated) Buffer results
  delay            (Deprecated) Delay results for a certain duration
  filter           Filter message by a classification threshold
  from-cloudtrail  Load messages from a Cloudtrail directory
  gen-viz          (Deprecated) Write out vizualization data frames
  inf-pytorch      Perform inference with PyTorch
  inf-triton       Perform inference with Triton
  monitor          Display throughput numbers at a specific point in the
                   pipeline
  preprocess       Convert messages to tokens
  serialize        Serializes messages into a text format
  timeseries       Perform time series anomaly detection and add prediction.
  to-file          Write all messages to a file
  to-kafka         Write all messages to a Kafka cluster
  train-ae         Deserialize source data from JSON
  validate         Validates pipeline output against an expected output

Note: The available commands for different types of pipelines are not the same. And the same stage in different pipelines may have different options. Please check the CLI help for the most up-to-date information during development.

Container Security

NVIDIA has observed false positive identification, by automated vulnerability scanning tools, of packages against NVD security bulletins. This can happen due to package name collisions (e.g., Mamba Boa with GPG Boa, python docker SDK with docker core). At the time of publishing this content, no known High/Critical CVE exist in these images. NVIDIA is committed to providing the highest quality software distribution to our customers.

Morpheus License

NVIDIA Morpheus is licensed under the Apache Software License 2.0.

NGC Deep Learning EULA

By pulling and using the container, you accept the terms and conditions of this End User License Agreement.