Visit the full NBR website
HomeTechnology  

TelstraClear limitless weekend degraded speed 40% to 64% - TrueNet

UPDATE: TelstraClear’s unmetered broadband weekend can be counted a success after customers took advantage in a frenzy of up and downloading, the company saidthis afternoon.

Customers used two and a half times more data overall than the previous weekend, TelstraClear said. Total usage was 359 terabytes (TB), 216TB more than the weekend before.

The greatest data benefit was to customers on the TelstraClear cable network, with total data they used being three times more than the previous weekend, jumping from 76 to 232TB, TelstraClear said.

The company also seized on TrueNet's independent testing (below) and TrueNet's analysis that, overall, the trial was a success. 

“We expected customers to take advantage of this opportunity and were upfront from the start. We said ‘With lots of people online for longer, some customers could experience slower than normal connection speeds’,” TelstraClear head of consumer markets Steve Jackson said.

TrueNet found DSL speed dropped by up to 40%, and Cable by up to 64%. At times, congestion prevented testing.

“We made substantial increases in our national and international capacity before the weekend, and put further capacity in place when the demand exceeded this. We acknowledge that some customers were unhappy with their internet experience, just as we note that many have reported they didn’t mind slower speeds because they were able to use far more data than they otherwise would, some reporting downloading more in the weekend than their entire monthly cap of 30GB," Mr Jackson said.

“We will analyse the feedback and other data before making decisions on this or future broadband promotions, however we can assure customers that we will continue to focus on giving them value,” Mr Jackson said.


7am: TelstraClear's all-you-can-eat weekend saw many complain about a degradation in speed, and here-and-there spot tests by individuals.

A number of customers told NBR their connections were unusable at times.

This morning independent testing company TrueNet has released figures that offer some empirical evidence. TrueNet uses a nationwide network of volunteers who install probes on their connections to track ISP performance in a survey carried out in partnership with the Telecommunications Users Association (Tuanz).

TrueNet principal John Butt said over the no-limit weekend, "TelstraClear national speeds reduced by almost 40% for ADSL customers and 64% for Cable customers." (Yesterday, TelstraClear said only international traffic slowed.)

Click chart to enlarge.

Given the circumstances, "the company managed the weekend well .... Overall customers remained able to surf the net and to access most services,” Mr Butt said.

Mr Butt said TelstraClear Cable customers were most affected. Their speed dropped to being on par with DSL.

“This is what we would expect. The speed has reduced to the available capacity, with both technologies having equal access to capacity and thus having equal speeds.”

Customers using TelstraClear’s unbundled DSL service (that is, accounts fed from Telecom exchanges where TelstraClear has moved in its own DSL switching gear) could well have escaped unscathed, with TrueNet’s testing showing minimal impact for unbundled customers.

Click chart to enlarge.

Download times for internationally hosted webpages were affected, but overall the service continued to operate.

Test results of international webpage download times  show that TelstraClear struggled with capacity until Sunday morning at 6am when they obviously fixed a problem, Mr Butt said. This fix started to fade again by the peak of 8pm and 10pm.

Overloading issue
Mr Butt said TelstraClear clearly had overloading issues on the Saturday, but by Sunday had largely resolved the problem.

“From 6am to 6pm, the service was operating as it normally does, but once the volume built up again after 6pm, the performance degraded.”

All in all, TelstraClear performed quite well considering the volumes of traffic being sent and received, Mr Butt said.

Average speeds
TrueNet found the average speed over the weekend was 7Mbit/s, but for some customers the speed available was potentially worse, Mr Butt said.

The average speed during the peak hour of 10pm on Friday night was 3.3Mbit/s, with the peaks on Saturday and Sunday a little earlier and with more capacity with the average speed being 4.1Mbit/s and 3.9Mbit/s respectively.

Too congested to measure at times
An interesting statistic is the number of times our probes were unable to measure - potentially due to congestion, Mr Butt said.

There were no missed measures for ADSL probes throughout the weekend, but for Cable, missed tests ranged from 7% (Saturday) to 18% (Sunday).

In the previous month, tests on Cable probes missed measuring less than 2% of the time, although on one day, the 23rd November, missed measures reached 7%, ie the same as Saturday.

Missed tests occur when volunteers' use of the Internet coincides with the TrueNet probes attempts to complete their own seperate, and independent download tests. Other missed tests would be due to time-outs if congestion impacts tests, Mr Butt said.

"TrueNet's interpretation is that TelstraClear's experiment with unlimited broadband was a success with average speeds ranging from 3.3Mbit/s to 7Mbit/s with a limited number of outages."

More by Chris Keall

Comments and questions

Well for downloading anything worth downloading (i.e. from overseas) speed degraration was >90%. On a normal day downloading a game off Steam runs at a reasonable ~500Kb/s possibly dropping to as low as 300Kb/s if there's heavy use. On the free weekend that went as low at 30Kb/s and rarely went as high as 50Kb/s. My maths might be a bit shonky but that doesn't look like a 40-64% degradation to me.

Kb (Kilobit) or KB (KiloByte)/s?

I get 1.3Mb from Steam on a ASDL and slingshot...I might stay away from Telstra..

The weekend event is not representative of your typical speed. As TelstraClear also has a locally hosted and unmetered Steam Content Server, you would be quite certain of achieving line speed downloads for available content as other TelstraClear users do.

The only time this does not apply is when http delivered titles are not available on any local Steam Content Servers and have to come from official Valve Content Servers.

If you have any more questions about it, you can hit me up directly on twitter - http://twitter.com/simcore (personal account, but happy to answer any questions you have Barry).

I'd stay away from TelstraClear simply because they think it is OK to use their commercial network and customer base as a test bed? Poor form!

International traffic was simply hopeless. I imagine the reason they think things were resolved by Sunday is because so many of their customers simply gave on up using the Internet.

If they are going to invite heavy loads, they need to be prepared to meet the demand.

I use TrustPower Kinects VDSL service and get 10MB up and 40MB down. Its awesome. My friends use their DSL service and get great speeds. His speed over doubled when switched from slingshot. You get what you pay for

According to speedtest.net, my NZ speeds dropped by over 85%, from 15 Mbps to under 2 Mbps. International was even worse, often totally unusable with sites not opening at all. iTunes repeatedly timed out and gave up trying to open the iTunes store part of the app.

I will have used far less TelstraClear data than normal, as I gave up and switched to the 4 - 6 Mbps I get on Vodafone by tethering on my iPhone. Thank goodness for the 3 GB I have there!

In Australia people get plans in the hundreds of GB and thats the norm. And they have 20million people so if TC can't handle a few customers using 2 or 3 times there data cap then their network is clearly under capacity. USA it's all unlimited every month even on your 4G phone and they have what, 300million people so, capacity needs to be improved I think!

Something must be wrong with the TC network. I am on Cable - Telstra's local NZ SpeedTest showed "local service degradation" to ~800kbs (LOCAL TRAFFIC) when I have a 15mbs/2mds. Traceroutes also showed dropped packets. What I don't understand is: If the real time gaming and YouTube video was trashed (no one would use that), so what exactly was the type of traffic that was using their net. Was it large file transfers peer to peer or vpns ? Were local customers networks using a free commercial tunnel ?

what a bunch of brain dead idiots Telstraclear are - they even admitted they knew there could be network issues and went ahead and did the offer anyhow...

This just smacks of arrogance..

Its like a guy with one barbecue opening an all you eat restaurant at an eden park game...

« Back to home page