Murakami

Murakami is a tool for creating an automated Internet measurement service, running in a Docker container. A Murakami measurement container will automatically run supported tests using a randomized schedule a configurable number of times a day, and can be configured to export each test result to a local storage device, to one or more remote servers via SCP, or to a Google Cloud Storage bucket. Results are saved as individual files in JSON new line format (.jsonl).

Contributing to Murakami

M-Lab welcomes your contribution to Murakami development. We recommend forking the repository and submitting pull requests to address issues we've identified. If there is a feature you would like to contribute that isn't in our list of issues, please create an issue so that it can be discussed and integrated into our roadmap.

Supported Measurement Tests

• Network Diagnostic Tool (NDT) - NDT is diagnostic and single stream performance test, originally developed by Internet2, and is now maintained by M-Lab. The supported NDT protocol versions listed below measure a connection using different TCP congestion control algorithms, and have different port/firewall requirements.
  • ndt5-client-go - The ndt5 protocol test uses CUBIC TCP congestion control, and is backward compatible with previous iterations of the NDT server. ndt5 requires some non-standard ports to be open.
  • ndt7-client-go - The ndt7 protocol test uses BBR TCP congestion control for the download measurement and for upload uses the congestion control enabled in the client network. ndt7 uses standard ports 80/443.
• Neubot DASH - Dynamic Adaptive Streaming over HTTP, hosted by M-Lab. DASH is designed to measure the quality of tested networks by emulating a video streaming player.
• Ookla's speedtest-cli - Ookla's official command line test for testing internet bandwidth using the speedtest.net platform.
  • Both multi-stream and single-stream tests are supported
• OONI Probe - The Open Observatory of Network Interference (OONI) is a non-profit free software project that aims to empower decentralized efforts in documenting Internet censorship around the world. Collect data that can potentially serve as evidence of Internet censorship since it shows how, when, where, and by whom it is implemented.

For more information about each supported test see: Supported Tests Runners.

To add new internet measurement tools to Murakami, see: Adding New Measurement Tools to Murakami

Supported Operating Systems

Murakami supports Linux operating systems like Ubuntu, Debian, etc. Windows is not supported. Mac OS may work, but is yet untested.

Configuring Murakami

A Murakami container can be configured flexibly depending on the deployment scenario. If you simply run a Murakami container using M-Lab's latest image on Dockerhub, by default all tests are configured to run four times daily at randomized intervals, but with no data exporters enabled. To use Murakami and save test results, either using an individual device or a fleet of managed devices, you'll use one of the options below to configure the container at runtime, or build your own local container image.

The Murakami container can be configured/customized using:

• in-line environment variables on the command line when running the Docker container
• customizing the configs/murakami.config.example configuration file and passing the variables it contains on the command line when running the Docker container
• customizing the configs/murakami.toml.example configuration file and including it in a deployed Docker container
• environment variables in a Balena Cloud fleet and/or per device

See CONFIGURING-MURAKAMI for more information on all configuration variables, and in our documentation specific to each deployment scenario.

Deployment Scenarios

Murakami supports two types of Docker container deployments on supported systems:

• Standalone: a single Murakami device, configured on-device
• Fleet Managed Murakami using Balena Cloud: one or more Murakami devices, configured and managed by the Balena Cloud IoT platform

It is also possible to install Murakami directly on supported systems without Docker, however currently documenting direct system installation of Murakami is beyond our project scope. Future testing and documentation is needed to test and confirm supported systems and requirements.

Dockerhub Images and Tags

Measurement Lab publishes a pre-built Murakami Docker container image on Dockerhub tagged latest. Please visit our repo on Dockerhub for more details.

Building Murakami Images

If you are interested in building your own Murakami Docker images, please see our BUILD instructions.

Updating Your Murakami Images/Installations

Murakami is periodically updated with new measurements and features, as well as updates to existing measurement services. M-Lab recommends keeping your Murakami installations or container images up to date with changes in the main branch of this repository, or with the latest image on Dockerhub. See the appropriate section on Updating Murakami in the supported Deployment Scenarios documents linked above for more information.

Included Utility Scripts for Google Cloud and Balena Cloud Tasks

Utility scripts that provide some automation to setup Google Cloud resources and provisioning new devices for a Balena Cloud fleet are available in utilities/. These are designed to support M-Lab demo fleets, so please inspect and customize them to your liking.

• setup_bq.sh - Intended to be run once to setup GCP resources for a new measurement program or demo. The Google Cloud SDK must be installed and authorized with an account that has IAM permissions to create GCS buckets and BigQuery datasets, tables and views.
• setup_balena_device.sh - This script can be used to provision new Murakami devices in your Balena Cloud fleet. It requires the balena-cli to be installed and authorized.
• bq_load_murakami.sh - Intended to be scheduled daily using scheduled jobs or cron, this script will load murakami test results files within a GCS bucket from the previous day into BigQuery tables.

For more details, see Murakami on Google Cloud Demo.

Optional Prototype Data Visualization Service

An optional data visualization service, Murakami-Viz, can be used to receive test results from Murakami measurement devices if desired. Please see USING-MURAKAMI-VIZ for more information.

Included Convenience Scripts

Two convenience utilities are provided with Murakami:

• murakami_convert scripts/convert.py
  • A script to convert jsonl test results to CSV.
• murakami_upload scripts/upload.py
  • A script to upload previously collected test results to a newly deployed instance of Murakami-Viz.

What Does the Name Mean?

What do we talk about when we talk about running tests? The M-Lab team talks about this a lot, and the name "Murakami" pays homage to the renown author Haruki Murakami, who's book What I Talk About When I Talk About Running provided us inspiration.

Acknowledgements

M-Lab would like to thank the Institute for Museum and Library Services (IMLS), and Simmons University, whose partnership on IMLS Award #LG-71-18-0110-18 supported the development of many current Murakami features, as well as Murakami-Viz.