The charting module of MS Office which is used by Excel is quite powerful but the interface is not at all intuitive and badly documented. There may be people who like it but I find it does not choose very useful defaults and I have to spend quite some time coercing the bloody thing into doing what I want. Oddly, or perhaps not, the object model that lets you manipulate charts in code is quite neat and more or less self documenting.
3 Likes
What has now been delivered?
In a nut shell after reading the report linked in the article.
98% of all NBN customer connections can achieve speeds of 25Mbps.
The Report also highlights the Fixed Line outcome of more than 90% of FL connections able to achieve speeds of at least 50Mbps.
Repeated references to this statistic tends to distract from reality. 90% of Fixed Line Services is approx 80% of all services when the Fixed Wireless and Satellite footprint is considered.
So - approx 80% of all NBN connections are able to achieve consistent speeds of at least 50Mbps when customers are using the service might be more accurate and honest. No puffery required?
The report avoids revealing any meaningful data on actual FW and Sat services. Aside from confirming the theoretical service speed possible at a site. Unlike fixed line bandwidth, there are severe capacity limitations on FW and Satellite NBN services. Neither solution is equal to any of the fixed line solutions in capacity, or speed.
It seems unlikely any government report will ever honestly detail the difference in service quality between Fixed Line and Satellite/FW, although all are equally expensive. More so for Satellite due to lower data allowances.
4 Likes
Having recently dumped nbn satellite, I donât know why I persisted with its unreliability and slow effective speeds (700ms ping time and ~15Mbps download/3Mbps up) for so long. Using my phone as a hotspot results in much more satisfactory internet performance, and we can watch iView live streaming without missing parts of programs.
Iâd suggest anyone else suffering with nbn satellite who has half decent phone signal give phone hotspotting a go, leaving nbn satellite to those with no phone signal.
6 Likes
The latest update featuring the usual âlots of data but does it mean muchâ report.
It now includes Fixed Wireless for those so encumbered to compare.
2 Likes
It seems to me that FW performance has two problems, one due to limits of the tech (mostly related to getting a strong uninterrupted signal between subscriber and base station) and the other due to implementation, that is the equipment between the base station and the larger web. My reason for saying this is that the tech seems to suffer from congestion much more than any of the wire connected techs.
As it happens my service is consistently good any time of day including the âNetflix hourâ On a 25/5 plan I consistently get 23-24/4.5; latency etc are no problem nor are outages. I realise this is somewhat atypical. It isnât as though I am right next to the base station, it is 7km away and I donât get degradation during heavy rain either.
A general observation on the new data re FW. As with wired tech the methodology is meticulous in showing error bars etc which are related to sample size and the variability of data. As with wired tech they give no indication how the white boxes are distributed so we have no idea if the sample is biased or not. The published performance figures based on sampled data only represent the true population if (and only if) the whites boxes are distributed randomly with respect to all measurement criteria.
As they are given to volunteers not premises chosen at random, the sample is only random if there are sufficient volunteers to choose between to allow stratified randomisation. If this is not possible the sample is biased through self-selection at least and maybe other ways. To take a hypothetical; if the budget for delivering and installing white boxes is a constraint then more remote areas may be under represented.
If anybody can find any information on how the white boxes are allocated I would be much obliged.
1 Like
The NBN was publicising 250Mbps and 750Mbps HFC being made progressively available from early last year with all HFC customers supposedly able to have 250Mbps by June this year.
There was some scepticism at the time.
The nbn is also offering higher speed plans on the fibre network.
The latest ACCC Measuring Broadband Australia Report No 11, Dec 2020 neglects to report on any of the higher speed plans, HFC or Fibre. IE faster than 100Mbps.
Has anyone seen any real world data on how these plans are performing?
2 Likes
After putting in a new TP-Link VR600v router I ran speedtest for the first time in a month. The new router reflects the old D-Link DVA2800 performance. I enjoyed this report where there are statistics and damned statistics. Australian average is 74 Mbps? One can guess that techies, gamers, and the few with âtop endâ services are over-represented. The mobile chart is equally funny in that my mobile is as often HSDPA as it is 4G, not considering the unresponsive networks at the back end and the difference in expectations between $500 and $1,500 platforms.
BTW, the replacement was due to a curious fault in the 2800; the main difference I see is the TP-Link WiFi signal is very weak compared to the D-Link. Otherwise it is a nice product purchased at a good sale price. The 2800 will be repaired or refunded in due time.
edit: D-Link deserves âbig tick recognitionâ for excellent product support, details not relevant here.
4 Likes
It does reflect one real life outcome in our experience. Sample of one! Mobile broadband performance in Australia is often faster than the fixed line services. Speedtest suggested and average of 109 Mbps in the report!
Admittedly our ADSL2 service since all the recent rain has gone a little mouldy and slowed down to under 10 Mbps. Not quite slow enough to convince of the benefits of a change to the NBN FW service with itâs best endeavours 6 Mbps peak time service?
5 Likes
Look on the bright side - we have leapt over Serbia in the rankings. It would be nice to see fixed Internet that fast, but maybe the NBN can yet be rescued.
4 Likes
The quarterly data pile has been released.
Sometimes one has to wonder, especially since download speeds do not seem to consistently reflect customer satisfaction, that is a mix of reliability, performance, and a capable friendly voice when one is needed, all dependent on the same opaque NBN Co for transport services.
4 Likes
The Sam Knows experience is wanting more volunteers. If someone wants to add their connection to the âpileâ (well put @PhilT 
) they can apply at
https://measuringbroadbandaustralia.com.au/
The whitebox is sent from the UK as Sam Knows is based there (ACCC use them for their expertise) and has 4 Ethernet ports and a WAN port (which you use to connect to your router or modem). Once you are connected and itâs running you can get personal results at anytime you wish to query it or run personalised speed tests (your connection).
When it tests the connection it uses some data, this amount is advised in your results pane. While it isnât much each time the unit tries several times each day (we have noted up to 8 tests in a day). The tests are run a number of times to get a good picture of what the connection is like over both peak and non peak times.
5 Likes
The quarterly report has been released. The snow seems as deep as ever even though there is some good information that can be gleaned, as well as information that begs why they bother it seems so meaningless, unless you are the NBN Executives whose bonuses depend on that data.
3 Likes
Slush and rocks as the afternoon snow melts, leaving an ice glaze the next morning. Typical Australian snow.
For the Fixed Wireless service the report is based on the data from 62 customers, (ref p41).
This represents a very insignificant number of tower sites. Each tower has multiple cells. It is not a reliable sample. It is at best representative of the specific
towers and assigned cell for those 62 customers on the 25Mbps tier or the 42 customers for the Fixed Wireless Plus option, (up to approx 50Mbps).
To be meaningful the NBN needs to provide a correlation between the sample customers and towers/cells/locations. IE a list all the NBN operational towers, the numbers of cells on each, and the assigned numbers of customers on each cell. More useful in that it can be compared to those cells and towers sampled. There is no assurance the result is without bias? Individual customers may see less performance due to circumstances, alignment to cells and distances.
4 Likes
In looking at the speeds of the Fixed Wireless NBN service, is anyone able to explain why the assessed data rates for streaming Netflix differ?
Netflix advises 5Mbps for HD and 25Mbps for UHD.
The ACCC report assesses performance based on less than half these recommended speeds.
From the report:
The ACCC report explains how the streaming speed is measured. HD and UHD are not defined in the report Glossary of terms used.
Experience previously of ADSL and Netflix SD required a 2-3Mbps connection. It was sometimes necessary to reduce quality to maintain a reliable connection.
.
4 Likes
A guess, Netflix uses the term ârecommendedâ for speeds that would make a satisfied happy Netflix customer. I suspect the NBN uses a standard that is minimally functional to make a happy satisfied Minister who might be reading the report.
Others?
7 Likes
And possibly to have uninterrupted streaming (occasional buffering is annoying to most viewers especially if it occurs in a good part of the video). The additional speed provides a buffer/bandwidth for other devices connected to the internet, along with the Netflix streaming deviceâŚas it is unlikely that a household will have one device connected at one time.
6 Likes
The Netflix recommendations are based on the calculated streaming bandwidth required for each type of stream. This incorporates the encoding formatâs compression value. Roughly a 640X480 @ 30 FPS requires 0.96Mbps if H.264 encoded for good quality, if MPEG4 it requires about 2.4 Mbps. This rate equates to roughly SD quality. 720p (HD) requires around 2 Mbps if H.264 or 6 Mbps if MPEG4. 1080p (FHD) requires 4 Mbps if H.264 or 12 Mbps if MPEG4, 1440 (QHD) requires 8 Mbps if H.264 or 24 Mbps if MPEG4. 4K (UHD) is even higher requirements again and a 25 Mbps nbn⢠plan will barely cope (it will be buffering in most cases), 8K and you need much higher bandwidth again and a 25 Mbps plan will not cope.
The NBN Co recommendations are based on a opaque calculation they have made it appears to me with no real relationship to real world data rates in my opinion. If you run nothing else but the stream then including overheads a HD stream would be all you could guarantee (without the need for buffering) on a 25/5 plan, forget a 4K or 8K stream. This accepts that 25/5 plans are not always going to achieve those speeds throughout the day. Peak periods will be much less. Remembering that Fixed Wireless is no longer a guarantee of a speed but is sold as a best effort service now.
7 Likes
The quarterly report (6 Dec 2021). Lots of data just because they can? It would be good if they included reports on the usefulness of fully charged mobiles during down times per RSP, per locality 
3 Likes
The Fixed Wireless results for the Busy Hours indicated significant drops in average speed for the faster FW Plus services to 35-40Mbps depending on the day, by my reading of the data. The slower 25/5Mbps âStandard speedâ plans more consistently averaged around 20Mbps for the Busy Hours.
As these are averages the ACCC report does not immediately reveal how many customers receiver lesser than average or by how much. The following graph offers some insight. Note 80% of FW users have a service that can reliably stream a single UHD service in the busy hours.
Conversely the graph is saying
- 20% of FW customers cannot get reliably 12Mbps,
- 15% cannot get 8.8Mbps
- 8% cannot get 6.6 Mbps
With only 41%, which is less than half those on FW able to get 24Mbps reliably, itâs a very raw deal.
One would hope they did not give up a reliable ADSL2 service for the benefits of the NBN.
In an associated release the ACCC commented,
âGiven the growing importance of upload speeds to how consumers use their broadband services, the ACCC is keeping a close eye on how clear retailers are with their customers about upload speeds, and we will consider any necessary amendments to our Broadband Speeds Claims Guidance for industry.â
It raises a further question as to whether the current design of the NBN is fit for purpose? With the exception of possibly full fat fibre FTTP the system has been designed to prioritise downloads over uploading by a large margin. The future includes increasing use of video conferencing and calls from home, work from home and cloud services, remote backup, and that all important home monitoring. In a more connected world should the ACCC be asking not only if the upload speeds are being achieved, but whether the plans available are fit for purpose?
4 Likes
As you noticed, the ACCC is keeping a list about how well the âNBN & RSPsâ delivers on claims. So far they have resulted in watering down claims, not improving service. What a surprise!
5 Likes
