Inverted Capacity Extended Engineering Experiment (ICE3)
Introduction
The recent widespread uptake of broadband access technologies has led to a
shift in how the Internet is being used. The availability of an always-connected,
high-speed Internet connection means that home users are increasingly likely to
use the Internet as an information repository and content delivery resource.
Higher content access speeds coupled with a zero connection time means that
Internet usage can become more spontaneous rather than planned for.
An always-on broadband Internet also increases the range of applications that
users are willing to try and adopt, beyond the current staple diet of web surfing
and email. Exposure to peer-to-peer applications, voice-based phone-like systems
(VoIP), and interactive multi-player games, creates further interest in these
more advanced applications involving mass data transfer and/or real-time
telecommunications.
The traditional Internet access model involves low bandwidth last-mile circuits
aggregating into higher bandwidth metropolitan, regional, and international
backbones. Consumer market last-mile access typically involves 56K dial-up, cable
modem, or ADSL technologies. Regional backbones are often measured in gigabits
per second, and many international backbones have capacities in the hundreds of
megabits per second.
Of particular interest is whether the current Internet architecture could support
an explosion in the patronage of these services and applications. This leads to
the questions of what network and applications developers should keep in mind
when designing these systems.
What would happen if we inverted this capacity hierarchy?
This question is no longer purely theoretical. Many countries are - or already
have - deploying high-speed broadband to their
citizens. Australia has joined this online revolution with the announcement
and initial deployment of the National Broadband
Network (NBN). The upshot is that in the near future we will an increasing
number of users connecting to the Internet at rates up to 100Mbps, with the
promise of up to 1Gbps.
In this environment, we could image that town- and city-wide broadband IP access
is provided as a common utility service (like electricity and gas). Every
suburb's IP customers could be milliseconds (and only a few hops) away from each
other. Every town library could run web-caches for their neighborhoods,
revitalizing their roles as 21st century information repositories.
ICE3 aims to characterize the performance and service quality impact
of inverting the content and capacity hierarchy, with particular reference to the
impact on the perceived performance of existing networked applications.
Goals
ICE3 encompasses the following topics:
- Evaluate the performance characteristics of some existing and emerging
content distribution methods (e.g. Email, Web, or peer to peer methods) as
a function of network bandwidth, latency, and hop counts (e.g. the dynamic
behavior of the underlying transport protocols, such as TCP)
- Develop plausible, alternative IP network architectures based on inverting
the existing bandwidth and service location hierarchy, including large
scale distribution of content caches around urban areas.
- Evaluate the consumer's likely experience if such alternative IP network
architectures were deployed, and the impact on wide- and local-area IP
traffic patterns and load growth.
As part of this project we will develop and release tools to assist in data gathering and analysis, and publish interim results and papers on this website.
Program Members
Jason But
Philip Branch
Sebastian Zander
Tony Cricenti
Grenville Armitage
(Alumni: David Prior, Clancy Malcolm, Wendy Vanhonacker, Claudio Favi)
|
