ANGEL - Automated Network Games Enhancement Layer
Introduction
ANGEL leverages the emerging field of Machine Learning (ML) to first automate
the classification and isolation of interactive and non-interactive traffic within
a service provider's network, and then to utilise this information to dynamically
reconfigure the network to better support the current traffic flows and deliver an
improved performance to the end users. Within this scope, the project has led to the
development of a prototype system with accompanying new protocols that allow the
adjustment of Consumer Premise Equipment (CPE - eg. cable/adsl modem) in response
to real-time detection to changes in the current traffic profile within the ISP
network.
Download a video - angel_hq_h264.mp4
(53MB Download, mp4 file, h264 Video, AAC Audio) - outlining the premise behind the
design of ANGEL, and demonstrating its effectiveness.
Please choose from the links in the menu above to learn more about ANGEL or to
see ANGEL related publications and presentations.
Who is involved?
The researchers involved in this project are:
Smart Internet Technology CRC
ANGEL is part of the Smart
Networks project of the Smart Internet
Technology CRC, who's stated aim is:
"The aim is to develop a flexible network and server infrastructure based on an
open architecture paradigm. This will enable a plethora of third party development
and solutions to be deployed cost effectively and scalably. Taking advantage of
processing resources made available within the network and through integration of
network-based knowledge and capabilities into the user application, a host of new
applications and services will be enabled and customised for delivery to a diverse
range of end-user devices."
Within the scope of this project, the work being conducted at
Swinburne University includes not only the
development of this automated network management within the ISP network to better
support multi-player networked games, but also the work being carried out as part
of the SONG (Simulating Online Networked Games) database project.
|