Overview
CAIA has developed a number of systems which utilise
machine learning (ML) techniques to classify network
traffic based on flow statistics.
ANGEL and
more recently
DIFFUSE
(funded by the Cisco University Research Program) have
both proved to be novel and capable architectures for
providing automated QoS provisioning for IP networks based
on ML classification.
DIFFUSE's architectural approach of integrating with
FreeBSD's IPFW firewall system makes it possible to easily
integrate and deploy advanced ML capabilities with a
general purpose operating system and widely used
firewall.
DIFFUSE extends IPFW to classify traffic based on
statistical properties of flows being observed in
realtime, and instantiate network actions across a
distributed set of "action nodes" for particular flows if
required.
Network architects frequently require the ability to
classify different traffic types flowing across a network,
typically using packet inspection capabilities of base
system tools such as ipfw and pf. Traffic classification
then enables the provision of customised service levels to
different traffic types (such as priority packet queuing
and forwarding, or allocation of specific bandwidth
guarantees).
This project aims to refine our DIFFUSE prototype and
integrate all components of the architecture into
FreeBSD.
Project Goals
- Cleanup and audit the DIFFUSE prototype code to prepare it for inclusion in FreeBSD.
- Integrate the DIFFUSE kernel and userspace code into the FreeBSD Subversion "head" branch.
- Add a new capability to the classifier so that it can perform an asynchronously controlled dump of all current rules to the flow exporter so that a freshly booted action node can receive the current system state.
- Add a rule templating facility to the collector so that it can be run on any system which provides fairly standard IP firewalling capabilities, not just FreeBSD systems running IPFW.
Schedule
The project will conclude by the end of October 2011.
Program Members