Broadband customers purchasing a 100Mbps service through a particular retail service provider (RSP) may not have been receiving the service they paid for, according to an ACCC report.
The ACCC's latest broadband performance monitoring and reporting program pilot revealed that customers of one RSP who purchased a 100Mbps HFC service were getting an average speed of 70Mbps instead. This was based on a data sample collected by the consumer watchdog.
The ACCC tested 90 home fixed-line broadband connections with various technologies over three months. Broadband performance measurement company, SamKnows, and network insight and analysis provider, Comdate, helped conduct the pilot.
The ACCC proposed a broadband performance monitoring and reporting program to address information asymmetry in the market and help consumers make more informed purchasing decisions. It also helps RSPs find where there’s a market failure and come up with ideas on how to differentiate themselves from competitors.
“An ongoing program would compare the performance of various RSPs in relation to the speed tiers they offer to help consumers verify performance claims made by RSPs about their services both pre and post-purchase,” the report said.
The report measured key factors that affect broadband performance. This includes downstream and upstream speeds, latency, video streaming, web browsing, packet loss, VoIP emulation (jitter), and Domain Name Service (DNS) response times and failure rate.
The data found some RSPs provide more stable services during peak periods than others, and how that could affect consumers’ purchasing decisions.
“For example, a consumer requiring a stable, high speed ADSL service may select RSP 2, but a consumer who wants a high speed service but does not need to use the Internet during peak periods may select RSP 9.
“On the other hand, a consumer who does not require high speeds and is more focussed on getting a cheaper service may select RSP 10 (if RSP 10 is a budget provider),” the ACCC said.
The report also looked at erratic performance during peak periods when it comes to upstream and downstream for all RSPs that use NBN's access network. One RSP stood out in its erratic performance, with the others being fairly stable, suggesting that it has not provisioned enough capacity for peak hours, rather than it being an NBN issue.
RSPs that provide both ADSL and HFC services were compared for how they perform in speed by hour of the week. It found that HFC services fluctuated significantly throughout the day.
“Of interest is that the average downstream speed of HFC service deteriorates significantly during the peak period to the extent that average speeds are sometimes below the speeds of ADSL services.”
Another interesting finding in the report is in relation to latency and comparing local and international servers that host online applications.
The report looked at servers located in Hong Kong and Melbourne for ADSL with each RSP.
“The difference in latency on international links between RSPs is likely a result of different international backhaul arrangements… These results show that if a website or game server was hosted in Hong Kong, users on RSP 2’s ADSL service would likely experience significantly better performance than users on RSP 4’s ADSL service, despite them having similar latency results within Melbourne.
“Consumers who frequently browse international websites or play online games on servers hosted internationally may want to select a provider who provides services with low latency on international links.”
Another interesting finding is in relation to video streaming. The report found that higher speeds did not perform significantly better than lower speeds when streaming video, after meeting the minimum bitrate requirement of a particular video streaming service.
Communications Alliance challenges pilot results
In response to the report, the Communications Alliance says the data sample size used for the pilot needs to be larger and more representative of all user situations.
“We need to ensure that if a monitoring program is introduced, it is cost-effective, produces reliable data and takes account of the fact that there are factors beyond the control of service providers that can influence the results,” said John Stanton, Communications Alliance CEO.
The ACCC noted in its report that collecting data on a larger scale and further ensuring it’s representative would help confirm the results truly reflect performance.
Stanton also said there needs to be a rigorous cost-benefit analysis to go with this proposed broadband performance monitoring and reporting program.
“The growing diversity of access technologies within the NBN multi-technology mix, the need to divide the results by region and the fact that there are more than 400 broadband service providers in Australia may add up to a very expensive solution – the cost of which will ultimately fall on taxpayers or internet consumers.”
The ACCC said it has not yet decided to go ahead with its proposed program as needs further stakeholder consultation.