AARNet's CDN Mirror

About AARNet's Mirror

AARNet's mirror archive was established as a pilot project in 1998 as recommendation for a Large Object Cache from the AARNet2 Caching and Mirroring working group. The initial digital curator was Jason Andrade

It aims to provide cost savings on international traffic and improved performance to frequently accessed archives to AARNet's members.

The initial deployment was a 4 processor Sun SS1000 with 256M of ram and 50G of disk, based at the University of Queensland within the Questnet regional network.

Over two years, the server was upgraded to 1G of ram (via a donation) and approximately 500 Gigabytes of disk space from UQ and AARNet. This was reaching the limits of the base hardware and a decision was made to upgrade the system.

In 2001, Sun Microsystems donated a new Sun Enterprise 450 and AARNet purchased two A1000 disk systems for the mirror. It was physically located at the AARNet Point of Presence (PoP) at the ACT Regional Network Organisation (ACTRNO), The Australian National University in Canberra.

In 2006, a significantly upgraded hardware solution was deployed with a dedicated connection to the AARNet3 backbone. Additional disk space provided a larger software archive coverage that has been populated with up to date software and archive distributions. The mirror was relocated back to QLD as part of the upgrade process.

During late 2009 the AARNet mirror underwent a full network, storage, and server hardware refresh being upgraded to 10Gbps in the process. With a total of 84TB of storage on Hitachi Data Systems hardware, 1.5TB of fast SSD based cache for popular content, and over 100GB of RAM, the AARNet Mirror has been tested at a sustained 7.6Gbps of random content delivery. 2010 will see significantly more content brought online.

Requests to archive or point to suitable objects, questions or suggestions can be made to the mirror management team at:


The mirror now provides a large coverage of OpenSource, and academically related archives to AARNet members, including UNIX sources and applications, Linux, Windows and OS X software and internet documents such as RFCs and FAQs.

We hope you will assist AARNet in realising the potential benefits of this service. We would be pleased to receive any suggestions about content, performance, experiences or issues that will assist in the development of this service.