Sydney suburb achieves fastest Akamai connections

Internet subscribers in the south-west Sydney suburb of Riverwood were connecting to Akamai's content servers faster than any other Australian city, according to a new report.

Akamai's state of the internet report for the third quarter of 2009 said subscribers in Riverwood connected at average speeds of 3204 Kbps - second only to the New Zealand city of Wellington in the Oceania region.

Riverwood was also the only place in Oceania to have over 80 percent of connections above 2 Mbps. Sydney as a whole could only manage 43 percent at that speed.

The author of the report, David Belson, told iTnews Riverwood's status on the list "would likely be due to our EdgeScape geo-location database having associated one or more blocks of end-user IP addresses specifically" with the suburb.

"We strive to provide data that is as granular as possible with EdgeScape," he said.

Each Akamai server has the EdgeScape data-gathering technology. It collects geographical and bandwidth-sensing information for every routable IP address on the Internet, the company said.

Belson also partially confirmed a theory that the suburb's proximity to an Akamai server could have led to the high-performance results.

Akamai has a server presence in about 70 locations across Australia.

"There may be an Akamai server deployment that is located optimally for users living in Riverwood, which would account, in part for their position on the list," Belson said.

He added: "This also speaks to the value provided by Akamai's distributed server deployment model - by delivering content from Akamai servers that are closer to more end users, we can help them maximise the speeds at which they can download content, mitigating performance-impacting issues like network congestion and distance-induced latency."

The only statistic that Riverwood did not top was the percentage of connections to Akamai at speeds above 5 Mbps. Here it scored 8 percent, behind Sydney as a whole (9.9 percent) and Canberra's massive 19 percent score.

Australia as a whole finished the third quarter with the highest level of internet penetration in the Asia Pacific. Akamai recorded 0.37 unique IPs per capita - placing it 11th globally.

Although penetration was higher than Asia Pacific internet powerhouses like South Korea, those countries continued to thump Australia on average speeds.

Korea's average connection speed was 14581 Kbps, giving it the number one rank globally. Australia came 47th in the world with a speed of 2284 Kbps.

Interestingly, that represented a 15 percent drop in average connection speed compared to the previous (consecutive) quarter.

But Stuart Spiteri, Akamai's managing director for Australia and Asia, believed the drop would spell good news for Australia in the long-run.

"We had this in Singapore last time [we ran the report]," Spiteri said.

"What you're seeing is new people want access to the internet but are signing up to moderate or low-speed plans to get into the game. So new entrants are causing a rise in unique IPs but mostly at a low-connection level.

"As more people do that over time they start to upgrade as they discover movies, Skype and so on. It can be very positive when you look at it like that."

Australian stats

Average connection speeds to Akamai (placed in top ten for Oceania region):
Riverwood 3204 Kbps (2nd)
Canberra 2990 Kbps (4th)
Sydney 2660 Kbps (6th)
Adelaide 2234 Kbps (9th)
Brisbane 2132 Kbps (10th)

Cities with the highest percentage of connections to Akamai at speeds over 5 Mbps:
Canberra 19 percent (1st)
Sydney 9.9 percent (4th)
Riverwood 8.0 percent (7th)
Melbourne 6.7 percent (8th)
Brisbane 6.1 percent (9th)
Adelaide 6.0 percent (10th)

Cities with the highest percentage of connections to Akamai at speeds over 2 Mbps:
Riverwood 82 percent (1st)
Canberra 45 percent (7th)
Sydney 43 percent (=8th)
Adelaide 43 percent (=8th)

How the average speed is calculated

"Akamai services billions of content requests per day. For each of these requests, we record, among other pieces of data, the user's IP address, the size of the file, and the amount of time it took to serve. From this, we can calculate a download speed for that request. This data is averaged over the course of the quarter, and aggregated by geography - at country/state/city levels, as appropriate." -- David Belson.