mPlane – an Intelligent Measurement Plane for Future Network and Application Management




Decentralized and diverse, the Internet is resilient and universal. However, its distributed nature leads to operational brittleness and difficulty in identifying and tracking the root causes of performance and availability issues. The first step to improve this fragmentation is measurement: illuminating the currently obscure dynamics of the Internet. To address this, we advocate a measurement plane, or mPlane alongside the Internet’s data and control planes.


mPlane consists of a Distributed Measurement Infrastructure to perform active, passive and hybrid measurements; it operates at a wide variety of scales and dynamically supports new functionality. A Repository and Analysis layer collects, stores, and analyses the collected data via parallel processing and data mining. Finally, an Intelligent Reasoner iteratively drills down into the cause of an evidence, determining the conditions leading to given issues, and supporting the understanding of problem origins.

Recent Blog Entries

mPlane SDK for Python 3 released

The mPlane Software Development Kit for Python 3 is now available from the Python Package Index. From any reasonable Python 3 installation, you can now simply type:

$ pip install mplane-sdk

(note that you may need to be root to do this, depending on your Python configuration).

To learn more about the SDK, start with the README file

Developing new components with the SDK is relatively straightforward: you simply have to wrap the code performing your measurement or query in an instance of mplane.model.Service, define capabilities to describe the specifications that code implements, and create a configuration file pointing at this Service, defining the certificate identifying the component and configuring access control. Then start the component runtime by running:

$ mpcom --config <path-to-config-file>

On the Rationale for the mPlane Measurement Architecture

mPlane is architected around three principles, derived from the challenges presented by the distributed, multiscalar, heterogenous environment in which it must operate: schema-centric measurement definition, explicit iterative measurement support, and weak imperativeness.

Why We Developed DBStream

Systems for processing network monitoring data need to offer high performance and usability. We think the Open Source system DBStream offers both.

1 of 3