NBN Australia to undergo first ACCC investigation/monitoring into speed

[desertcomputer]

4 hours ago, Windspeed36 said:

I've without a doubt been shaped by Optus - chewed through 4TB+ per month when I first got it. Went down to 3/10 during peak. Complained to TIO, Optus suddenly fixed my area and I went back up to 80-100/40. 

 

I then asked for a static IP and now I'm back down to 30/35 at most times

I used to have speed issues during peak hours, but after a couple of phone calls to supports and complaining about the speed and ping (ping mostly). TPG have seem to solve the issue a day after, now during peak hours ping and download is also more consistent. I also hear pole complaining about the speed and they seem to fix the issue eventually, they could just be setting special rules/QOS for people for complain about their speed but who really knows.  

 

Backing up my NAS to crash plan (3TB down, 2tb to go), almost at 4TB mark :D. It really slow even at 30mpbs ... 9 day remaining 

[leadeater]

6 hours ago, JAKEBAB said:

I don't think anyone really denies FTTN is an upgrade, it's just a really shitty upgrade. either way what good is FTTP with these CVC prices anyway, limited to 100 - 200mbps and plagued with congestion lol.

FTTN isn't really an upgrade, not in the way it is being advertised. It's the very same topology as before, many cabinets were already fibre backhaul and there was nothing in the past stopping ISP's from upgrading the equipment to ones that support newer and better DSL protocols.

 

All the previous limitations still exist after a FTTN upgrade. To support VDSL2 and G.Fast you have to be very close to a Node and you need high quality copper phone lines which is only available in newer developments who should have been putting in ducting to allow for future upgrades, blown fibre.

 

If you are going to get the full benefits of a FTTN design you need to deploy more Nodes, to deploy more Nodes you need to lay the fibre to it and then new or move existing copper for every house to the Node. If you have to do that then FTTH/P is the exact same civil works cost which is the reason for the back out of a full FTTH/P deployment.

 

I wouldn't really complain about actually getting equipment in Cabinets/Nodes upgraded and fibre backhaul for ones that don't already have it but do hold them to account for what it actually is, the very thing ISP's should have been doing themselves without public money.

[skywake]

2 hours ago, mr moose said:

The difference in data used on wireless versus fixed line currently is a reflection of pricing and usage type.  I Don't know what the actual physical limitation is of the mobile network, But given it is more likely to be rolled out in smaller towns and country areas than inner city (where fttp is cost effective) it would be interesting to see a feasibility study.

The pricing of mobile data is a bit of a chicken and the egg sort of scenario. The pricing is high because the capacity is low which reduces the usage which improves the speeds. But you can do a guesstimate for capacity based on a bit of simple maths. From my understanding typical 4G tower has about a 5KM radius with something like 2Gbps of total capacity. So lets for arguments sake say that 5G pushes the capacity upto 10Gbps and they use a tower that covers a 500m radius.

 

If you assume the houses in the area covered are typical quarter acre blocks and there's a bit of extra space used for roads/parks ect? It works out that a 500m radius should approximately cover around 500houses. Which means that at full capacity that tower is pushing 20Mbps per house. Which isn't bad considering that people won't be hammering the tower all at once but in terms of total theoretical capacity it not better than FTTN and it's barely better than ADSL. That's with what I'd wager are fairly friendly estimates of what "5G" is actually capable of and the sort of tower densities that NIMBYs would tolerate.

 

26 minutes ago, leadeater said:

If you are going to get the full benefits of a FTTN design you need to deploy more Nodes, to deploy more Nodes you need to lay the fibre to it and then new or move existing copper for every house to the Node. If you have to do that then FTTH/P is the exact same civil works cost which is the reason for the back out of a full FTTH/P deployment.

That's the issue and to a degree that's what's happening. They are deploying more nodes, a lot more nodes. They're also actually replacing some of the existing copper in some cases. And in some areas they're actually doing FTTP despite the area being planned for FTTN because of the amount of copper they would have had to replace. All of which together has meant that the deployment of FTTN is taking about as long as the FTTP plan was going to.

But still, what we are getting is technically the second best plan for broadband infrastructure the country has had. Government or otherwise. And there were plenty of ideas floating around from the party that is now implementing it before they got hold of the keys. Including a plan to do the whole thing with LTE.....

[JAKEBAB]

10 hours ago, leadeater said:

FTTN isn't really an upgrade, not in the way it is being advertised. It's the very same topology as before, many cabinets were already fibre backhaul and there was nothing in the past stopping ISP's from upgrading the equipment to ones that support newer and better DSL protocols.

 

All the previous limitations still exist after a FTTN upgrade. To support VDSL2 and G.Fast you have to be very close to a Node and you need high quality copper phone lines which is only available in newer developments who should have been putting in ducting to allow for future upgrades, blown fibre.

 

If you are going to get the full benefits of a FTTN design you need to deploy more Nodes, to deploy more Nodes you need to lay the fibre to it and then new or move existing copper for every house to the Node. If you have to do that then FTTH/P is the exact same civil works cost which is the reason for the back out of a full FTTH/P deployment.

 

I wouldn't really complain about actually getting equipment in Cabinets/Nodes upgraded and fibre backhaul for ones that don't already have it but do hold them to account for what it actually is, the very thing ISP's should have been doing themselves without public money.

1. Yes but it still is technically an upgrade since people will experience faster speeds no matter how little.

 

2. Agreed.

 

3.Yes and now they have introduced FTTdp which is just ridiculous, with all the shit that spat like skinny fiber and being able to put multi-ports directly into the pit, no more FDH's etc, it really shows this is all just a political game of anything but FTTP.

 

4. Well yes for the price we are paying for FTTN which is very similar to the price of FTTP with all the cost blowouts, (which no one totally saw coming -_-) i'd rather us just get no "upgrade" than waste the money.

[mr moose]

8 hours ago, skywake said:

The pricing of mobile data is a bit of a chicken and the egg sort of scenario. The pricing is high because the capacity is low which reduces the usage which improves the speeds. But you can do a guesstimate for capacity based on a bit of simple maths. From my understanding typical 4G tower has about a 5KM radius with something like 2Gbps of total capacity. So lets for arguments sake say that 5G pushes the capacity upto 10Gbps and they use a tower that covers a 500m radius.

 

If you assume the houses in the area covered are typical quarter acre blocks and there's a bit of extra space used for roads/parks ect? It works out that a 500m radius should approximately cover around 500houses. Which means that at full capacity that tower is pushing 20Mbps per house. Which isn't bad considering that people won't be hammering the tower all at once but in terms of total theoretical capacity it not better than FTTN and it's barely better than ADSL. That's with what I'd wager are fairly friendly estimates of what "5G" is actually capable of and the sort of tower densities that NIMBYs would tolerate.

 

That's the issue and to a degree that's what's happening. They are deploying more nodes, a lot more nodes. They're also actually replacing some of the existing copper in some cases. And in some areas they're actually doing FTTP despite the area being planned for FTTN because of the amount of copper they would have had to replace. All of which together has meant that the deployment of FTTN is taking about as long as the FTTP plan was going to.

But still, what we are getting is technically the second best plan for broadband infrastructure the country has had. Government or otherwise. And there were plenty of ideas floating around from the party that is now implementing it before they got hold of the keys. Including a plan to do the whole thing with LTE.....

 

Here's a good real world example. South Korea has 33M mobile users (3 times our number of total internet connections) using 93 petabytes but condensed into an area half the size of victoria on their 4g mobile network.  While we consume 2.6Exabytes it is spread out of a vastly larger area, Which means the difference can be made up with more towers. 

 

https://www.mobileworldlive.com/asia/asia-news/4g-accounts-92-koreas-data-traffic/

 

I really do think that with system improvements and the spread of our population wireless does have the potential,  it may not currently supply inner suburbs, but it doesn't have to as most inner suburbs are well supported with a fiber back haul if not fttp already.  

[skywake]

5 hours ago, mr moose said:

Here's a good real world example. South Korea has 33M mobile users (3 times our number of total internet connections) using 93 petabytes but condensed into an area half the size of victoria on their 4g mobile network.  While we consume 2.6Exabytes it is spread out of a vastly larger area, Which means the difference can be made up with more towers.

That doesn't mean its a viable upgrade path from xDSL. Just because it's technically possible doesn't mean it's commercially viable. As I said, there's a point where the tower density would be so high in suburbia that the NIMBYs would get vocal. Easy to get away with that sort of thing in the middle of a city, much harder in the suburbs. You also forget that South Korea has fantastic 4G AND fantastic fixed line infrastructure. 

But you are right about one thing. We're actually ranked pretty damn high in terms of mobile speeds because of factors like those you've outlined. I'm looking at a graph now and South Korea has an average 4G speed of ~46Mbps, we're at ~32Mbps, Japan is ~22Mbps and the US is 14Mbps. If you dropped more data on our mobile network by rapidly reducing the cost of mobile data that would soon change. The bigger issue here is that we're ranked 51st in the OECD in terms of total average internet speeds. On that measure South Korea is again well up the list with ~26Mbps, Japan is ~19Mbps, US ~17Mbps.... we're ~10Mbps....

[mr moose]

1 hour ago, skywake said:

That doesn't mean its a viable upgrade path from xDSL. Just because it's technically possible doesn't mean it's commercially viable. As I said, there's a point where the tower density would be so high in suburbia that the NIMBYs would get vocal. Easy to get away with that sort of thing in the middle of a city, much harder in the suburbs. You also forget that South Korea has fantastic 4G AND fantastic fixed line infrastructure. 

But you are right about one thing. We're actually ranked pretty damn high in terms of mobile speeds because of factors like those you've outlined. I'm looking at a graph now and South Korea has an average 4G speed of ~46Mbps, we're at ~32Mbps, Japan is ~22Mbps and the US is 14Mbps. If you dropped more data on our mobile network by rapidly reducing the cost of mobile data that would soon change. The bigger issue here is that we're ranked 51st in the OECD in terms of total average internet speeds. On that measure South Korea is again well up the list with ~26Mbps, Japan is ~19Mbps, US ~17Mbps.... we're ~10Mbps....

you're right, it might change if people started using more data, or again because the population density is so incredibly less, they may not need to do more than double the towers to maintain speed.   But from this point on it is really just speculation on my part. 

[Cheddle]

what a sham... ISP's will simply QoS the taget locations for the speed testing...

Ive had multiple ISPS on my 100mbit connection and some are SCUM! Ive had speedtest.net report perfect 100mbit download speeds but then I cannt watch a 720p youtube video and my steam, origin and oculus store downloads are all at less than 100kb/s

Good luck.

[mr moose]

1 hour ago, Cheddle said:

what a sham... ISP's will simply QoS the taget locations for the speed testing...

Ive had multiple ISPS on my 100mbit connection and some are SCUM! Ive had speedtest.net report perfect 100mbit download speeds but then I cannt watch a 720p youtube video and my steam, origin and oculus store downloads are all at less than 100kb/s

Good luck.

that was my fear too.   I know telstra were experimenting with shaping for specific services, but I don't know whether they actually went ahead with it or not.  More than likely they did given the reports from a few of the members here.  

 

What would be good is if google, stan, netflix or anyone with the tech, could link a speed test up through their URL's.  That way ISP's would be caught red handed when you show the ombudsmen three speed tests that vary by way to much to be distance/server quality related.

[leadeater]

Just now, mr moose said:

that was my fear too.   I know telstra were experimenting with shaping for specific services, but I don't know whether they actually went ahead with it or not.  More than likely they did given the reports from a few of the members here.  

 

What would be good is if google, stan, netflix or anyone with the tech, could link a speed test up through their URL's.  That way ISP's would be caught red handed when you show the ombudsmen three speed tests that vary by way to much to be distance/server quality related.

I'm not actually against traffic shaping, it's something I know well and have implemented. The problem with ISP traffic shaping is it's very poorly thought out and not optimized to achieve best throughput for all.

 

I've done it on large networks and have been able to vastly improve the experience and download speed for everyone by doing it compared to not having it.

 

In a perfect world with infinite bandwidth QoS/Traffic Shaping would never be required.

[mr moose]

18 minutes ago, leadeater said:

I'm not actually against traffic shaping, it's something I know well and have implemented. The problem with ISP traffic shaping is it's very poorly thought out and not optimized to achieve best throughput for all.

 

I've done it on large networks and have been able to vastly improve the experience and download speed for everyone by doing it compared to not having it.

 

In a perfect world with infinite bandwidth QoS/Traffic Shaping would never be required.

Might be very good to shape a private companies or schools network access, it's their workplace and their right, but when I pay an ISP for internet access, I pay for full unhindered Internet access not for limited or restricted in any way.

[Dabombinable]

1 minute ago, mr moose said:

Might be very good to shape a private companies or schools network access, it's their workplace and their right, but when I pay an ISP for internet access, I pay for full unhindered Internet access not for limited or restricted in any way.

Eg. Telstra stating that out of the "up to 20Mbps" we can get 9 Mbps (they know full well that they are lying to consumers). We get that 9 Mbps for exactly 3 months, then it gets cut to 6-7 Mbps for everything except bandwidth testing websites-which also report double the actual upload speeds.

[leadeater]

1 minute ago, mr moose said:

Might be very good to shape a private companies or schools network access, it's their workplace and their right, but when I pay an ISP for internet access, I pay for full unhindered Internet access not for limited or restricted in any way.

Technically that's not actually what you are paying for but who reads the fine print 

 

Just remember shaping doesn't mean slow or unusable, often reducing per connection bandwidth by as little as 10% can completely eliminate congestion and can be done only when required. If you had a choice between 10% slower during that problem period versus an unreliable service I'm sure you know which is the logical choice.

 

Shaping isn't bad and it isn't what is commonly thought of by the public, it just has a bad name due to bad implementations and misunderstanding of internet problems which may have nothing to do with it at all.

[mr moose]

3 minutes ago, leadeater said:

Technically that's not actually what you are paying for but who reads the fine print 

 

Just remember shaping doesn't mean slow or unusable, often reducing per connection bandwidth by as little as 10% can completely eliminate congestion and can be done only when required. If you had a choice between 10% slower during that problem period versus an unreliable service I'm sure you know which is the logical choice.

 

Shaping isn't bad and it isn't what is commonly thought of by the public, it just has a bad name due to bad implementations and misunderstanding of internet problems which may have nothing to do with it at all.

 

Except when I can get a 2MB/s download but a website literally takes 30secs to load.   Something fishy going on here.

[leadeater]

9 minutes ago, mr moose said:

Except when I can get a 2MB/s download but a website literally takes 30secs to load.   Something fishy going on here.

That's why I'm only saying I'm not against traffic shaping the technology. I don't know what's going on over in Aus since I'm not there but I have heard enough complaints to know you aren't getting the quality of service you are paying for or equivalent to what we get for the same money.

 

Good to see that it's getting looked in to I just hope something actually comes of it.

[Jito463]

I'm glad that's not an issue I've had to deal with (the bandwidth "shaping").  While I may overpay for my 12 Mbps DSL, I almost consistently get 1.4 - 1.5 MBps downloads (note: that's MB, not Mb, which translates to 12 Mbps).

[Eroda]

On 2017-6-19 at 11:21 AM, adams said:

Wireless technologies (like 5G) will soon be way faster than the NBN anyway so it is useless.

thats great unless you want to do anything that is ping sensitive i get 80ms more ping when using wireless

[mr moose]

4 hours ago, Eroda said:

thats great unless you want to do anything that is ping sensitive i get 80ms more ping when using wireless

Ping is improving on wireless technologies,  It may not be there yet as a replacement for fixed line.  But I won't be surprised if you can't tell the difference in a few years.