flowcept 0.1.2

A system to integrate data from multiple workflows.

FlowCept

FlowCept is a runtime data integration system that aims at empowering any data generation system to capture workflow provenance data using data observability, with minimal (often no) changes in the target system code. Thus, it is able to integrate data from multiple workflows, enabling users to understand complex, heterogeneous, large-scale data coming from various sources in federated environments.

FlowCept is intended to address scenarios where multiple workflows in a science campaign or in an enterprise run and generate important data to be analyzed in an integrated manner. Since these workflows may use different data generation tools or can be executed within different parallel computing systems (e.g., Dask, Spark, workflow management systems), its key differentiator is the capability to seamless integrate multi-workflow data from various sources using data observability. It builds an integrated data view at runtime of these multi-workflow data following W3C PROV recommendations for its data schema. It does not require changes in user codes or systems (i.e., instrumentation). All users need to do is to create adapters for their systems or tools, if one is not available yet.

Currently, FlowCept provides adapters for: Dask, MLFlow, TensorBoard, and Zambeze.

See the Jupyter Notebooks for utilization examples.

See the Contributing file for guidelines to contribute with new adapters. Note that we may use the term 'plugin' in the codebase as a synonym to adapter. Future releases should standardize the terminology to use adapter.

Install and Setup:

1. Install FlowCept:

pip install .[full] in this directory (or pip install flowcept[full]).

For convenience, this will install all dependencies for all adapters. But it can install dependencies for adapters you will not use. For this reason, you may want to install like this: pip install .[adapter_key1,adapter_key2] for the adapters we have implemented, e.g., pip install .[dask]. See extra_requirements if you want to install the dependencies individually.

2. Start MongoDB and Redis:

To enable the full advantages of FlowCept, the user needs to run Redis, as FlowCept's message queue system, and MongoDB, as FlowCept's main database system. The easiest way to start Redis and MongoDB is by using the docker-compose file for its dependent services: MongoDB and Redis. You only need RabbitMQ if you want to observe Zambeze messages as well.

3. Define the settings (e.g., routes and ports) accordingly in the settings.yaml file.

4. Start the observation using the Controller API, as shown in the Jupyter Notebooks.

5. To use FlowCept's Query API, see utilization examples in the notebooks.

Performance Tuning for Performance Evaluation

In the settings.yaml file, the following variables might impact interception performance:

main_redis:
  buffer_size: 50
  insertion_buffer_time_secs: 5

plugin:
  enrich_messages: false

And other variables depending on the Plugin. For instance, in Dask, timestamp creation by workers add interception overhead.

Plugins-specific info

You can run pip install flowcept[plugin_name] to install requirements for a specific plugin, instead of installing the whole package.

RabbitMQ for Zambeze plugin

$ docker run -it --rm --name rabbitmq -d -p 5672:5672 -p 15672:15672 rabbitmq:3.11-management

Tensorboard

If you're on mac, pip install may not work out of the box because of Tensorflow library. You may need to pip install tensorflow-macos instead of the tensorflow lib available in the tensorboard-requirements.

See also

• Zambeze Repository

Acknowledgement

This research uses resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725.