A letter requesting Unmetered National Data Transfer

Forums › ICT Policies and Regulation › A letter requesting Unmetered National Data Transfer

1 | 2 | 3 | 4 | 5

8223 posts

Uber Geek

Trusted

#236766 18-Jul-2009 21:49

I can think of two legitimate and viable uses of heavy local traffic straight away, there's probably many more:

Remote backup services
On Demand Pay TV, Movies, Music (like IPTV)

Regarding free national peering, we already have APE and WIX but both Telecom and Telstra decided to de-peer a few years ago now.

SamF

1578 posts

Uber Geek

Trusted

#236783 18-Jul-2009 23:19

At the end of the day, yes, national traffic costs for ISPs, but it is a LOT cheaper than international traffic! Enabling national peering & then letting customers have it for a reduced rate will reduce encourage people to use national sources and reduce international traffic throughput. This in turn will save ISPs money.

nate

6473 posts

Uber Geek

Retired Mod

Trusted

Lifetime subscriber

#236833 19-Jul-2009 10:40

I like the idea of the peering exchanges, the uphill battle would be convincing the government that the usage would warrant the significant investment in such infrastructure.

Ragnor: I can think of two legitimate and viable uses of heavy local traffic straight away, there's probably many more:

Remote backup services
On Demand Pay TV, Movies, Music (like IPTV)

Good thinking, I think the IPTV would have to be widely adopted by someone such as Sky for everyone to take it up. Remote backup would only be used by the true geeks, like those here :)

Screeb

698 posts

Ultimate Geek

#236883 19-Jul-2009 14:40

nate: I don't think national traffic should be free, as with anything if it's free it will be abused.

Like I asked of coffeebaron, how is that different to other countries? How do THEY handle this "abuse"? Why is NZ any different? Other countries are doing just fine with unlimited national (and international for that matter) traffic.

insane

3240 posts

Uber Geek

ID Verified

Trusted

#236908 19-Jul-2009 17:09

nate: I don't think national traffic should be free, as with anything if it's free it will be abused.? The flipside of this is an ISP's billing engine may not be setup to count traffic differently so there will be costs involved in installing hardware/software to count traffic differently.

National traffic also does cost to backhaul around the country.? Kordia don't give ISPs free bandwidth, so if national was free, more capacity would be needed, so ISPs would require a larger pipe down the country.? With ADSL margins already very, very slim, this overhead has to be covered somewhere else.

Maybe a smarter solution would be that content providers such as TVNZ, Sky, et al peer directly at a peering exchange (such as the APE), then ISPs can pass data directly between them, and zero rate this.

Last query, for the normal ADSL user, who would use copious amounts of national bandwidth anyway?

I can see you you've got a good understanding of how things work, its refreshing :)

The original writer spoke of how their business would do better with free local traffic, however we can see that a similar company Datalock.co.nz managed to get a partnership with Maxnet whereby Maxnet internet users are not changed for local traffic to Datalocks backup servers.

I don't see why this Databack couldn't do the same with Iconz, or more their services to another provider who would offer such a service for them. Would be interesting to know whether they have even asked.

Without sounding too negative I don't see national traffic being free or much less until telecom/telstra peer again, or they are regulated.

I do wonder what would happen if the big content providers decided to host with providers who do peer openly, maybe then it would give telecom/telstra more reason to.

cokemaster

Exited
4929 posts

Uber Geek

Retired Mod

Trusted

Lifetime subscriber

#236911 19-Jul-2009 17:14

Screeb: Other countries are doing just fine with unlimited national (and international for that matter) traffic.

Yes, indeed.

webhosting

Loose lips may sink ships - Be smart - Don't post internal/commercially sensitive or confidential information!

SamF

1578 posts

Uber Geek

Trusted

#236912 19-Jul-2009 17:17

Actually, we already have an arrangement with ICONZ regarding national traffic, however this doesn't help our customers who are not with ICONZ...

Sharesies

Trade NZ and US shares and funds with Sharesies (affiliate link).

exportgoldman

1202 posts

Uber Geek

Trusted

#236915 19-Jul-2009 17:28

I can't for the life of me find the PDF of Telecom's proposal for free local peering for what I thought was 13 but perhaps 26 local peering exchanges, from this document about their NGN they talk about offering local peering, and the writer of the report notes that this is the notable different between them and TelstraClear

Page 18

Telecom Response to Commission questionnaire Commerce Commission NGN study

Tyler - Parnell Geek - iPhone 3G - Lenovo X301 - Kaseya - Great Western Steak House, these are some of my favourite things.

coffeebaron

6235 posts

Uber Geek

Trusted

Lifetime subscriber

#236919 19-Jul-2009 17:47

Screeb:
nate: I don't think national traffic should be free, as with anything if it's free it will be abused.

Like I asked of coffeebaron, how is that different to other countries? How do THEY handle this "abuse"? Why is NZ any different? Other countries are doing just fine with unlimited national (and international for that matter) traffic.

As cokemaster points out, they aren't managing, and more and more we are hearing of these overseas ISPs beginning to put restrictions on data usage.

Because we in NZ are on a much smaller scale, there just isn't the same economics of scale at play.
There are also more choices overseas, e.g. UK or USA have several fixed line players including a large cable TV network. We only have Telecom, plus a bit of Telstra cable in Wellington / Christchurch.

We don't have big apartment blocks: what do you think the economies of scale are for feeding a bit of fibre backhaul into a few massive apartment blocks, Vs a whole suburb with the same population?

I definitely think there is a lot more that can be done, but at the same time the grass is not as green as some people think on the other side.

Rural IT and Broadband support.

Broadband troubleshooting and master filter installs.
Starlink installer - one month free: https://www.starlink.com/?referral=RC-32845-88860-71
Wi-Fi and networking
Cel-Fi supply and installer - boost your mobile phone coverage legally

Need help in Auckland, Waikato or BoP? Click my email button, or email me direct: [my user name] at geekzonemail dot com

PenultimateHop

637 posts

Ultimate Geek

Trusted

#236930 19-Jul-2009 19:08

There is so much misinformation in this thread, I will try to clear some of it up over a few posts as I have time.

Screeb:
coffeebaron:
Unfortunately it is not absurd, and is exactly what happened in the previous days of unmetered NZ traffic that some ISP's had. One of the main culprits was DC Hubs, shifting up into the Terabytes of data per month.

coffeebaron:
Try signing up for a guaranteed 80GB of traffic per month = $165 on a 256Kbps UNS connection. OK, so this is a commercial hi grade service, but how much would it cost to do the same for residential? ISP's work on very slim margins, you would need a reasonable price hike to make anything near unmetered national data a reality.
This was all fine a few years a go when 5GB per month was a huge amount of data, but not so now.

But how is that different to other countries? How do THEY handle this "abuse"? Why is NZ any different?

I'll start with this one.

Three (well, two and a half) reasons:

1. Subscriber density. Many of the operators in other countries have subscriber bases which equal the population base of New Zealand. It's not unusual to see 'mid size' ISPs with 1-2 million subscribers. In New Zealand there's only one ISP which even comes close to that.

Even in Australia, a small DSL ISP might have 200K customers. A small DSL ISP in New Zealand is likely to have less than 20K.

When you have subscriber density in these numbers, economies of scale come into play. 10% of the subscribers still use 90% of the resources, but you have a huge amount more income to work with, allowing you to afford those 10% of subscribers. Even then it can still be tough for an ISP to actually make money. When your subscriber base does not give you massive income that is required to service those customers, you just don't have the flexibility in your business to allow unlimited downloads.

2. Competition. In all of: access loops to the home, wholesale DSL networks, and wholesale IP transit. In many countries with bigger populations, it is cheaper to buy transit than it is to peer. This is certainly the case in the USA and in lots of European countries. If you can buy 10Gbps of transit to anywhere for $4000 per month, then obviously it gives you a lot of options for your subscribers vs. paying $200/Mbit!

NZ does not have this competition, and does not (yet) have the bandwidth demand to provide it. Additional trans-Oceanic cables might help, I suspect an increase in bandwidth demand is more likely to provide the stimulus to reduce costs and introduce more IP wholesale competitors. There are various plateaus that are reached in order to bring costs down, and so far nobody is running circuits bigger than STM-16 into New Zealand. Breaking that plateau will help to reduce the wholesale costs to ISPs.

Of course, international transit bandwidth is only one component of an ISPs costs (although it is usually their second largest, after staff).

3. No distinguisher between peering, offshore, and onshore. Many countries just don't distinguish because they are either:

a. Sitting in a strategic location for interconnection (e.g. Singapore, USA)
b. Not on the very far end of a piece of wet string (e.g. Anywhere in Europe)
c. Are seen as the centre of the Internet (USA!)
d. Have a significant population with particular language requirements, which means the majority of their traffic is sourced locally (e.g. China, Korea, Japan, some European nations...)
e. Are in a highly competitive, developed, wholesale market where there is often a financial disincentive to peer vs. buy transit (USA, parts of Europe).

It is definitely also worth pointing out that there are significant infrastructure and wholesale costs in offering free domestic bandwidth, such as:

1. Increase in wholesale costs from your wholesale operator (e.g. TNZ). Their network is not dimensioned to sustain significant volumes of traffic over the subscriber loop, although this is definitely improving. However this is a significant stranglehold until alternatives become available for backhaul services, and TNZ deciding to alter their models.

2. Network investment at the ISP. This will include more or bigger BRAS - this is a significant investment in the millions of dollars. Additionally the ISP may now need to upgrade their core infrastructure and their switching infrastructure, as well as their accounting/metering infrastructure and all their support systems. This could be anywhere from $0.5M to $10M depending on the size of ISP, their support systems, and their vendor selection.

3. Assumes that there is actual content and business to drive this investment. At the moment there really isn't, but this is a usual circular chicken-and-egg debate.

As a final note, it's worth understanding that many of these "overseas operators who provide unlimited broadband services" oversubscribe and traffic manage their networks absolutely ferociously. I have unlimited 100Mbps HFC service at my home costing me NZD$130 per month, and I am lucky if I get beyond 2-4Mbps on anything - the highest I have ever seen was 40Mbps to a local speedtest server, and about 16Mbps from Akamai for AppleTV content. Additionally it is DPI'd (at every CMTS!), aggregate traffic shaped, per subscriber traffic shaped (punitively based on subscriber profile), and transparent proxied and content filtered (badly). This makes the overall performance frankly suck. I had better experiences in Australia and New Zealand - at least I had a much better chance of hitting my PIR.

The grass is not always greener.

Screeb

698 posts

Ultimate Geek

#236931 19-Jul-2009 19:11

cokemaster:
Screeb: Other countries are doing just fine with unlimited national (and international for that matter) traffic.

Yes, indeed.

Great understanding of the issues there! The reason they're doing that isn't that the network can't handle it, but because Time Warner and Comcast are protecting their cable TV business revenues by making it impossible for users to download movies and TV shows instead of watching TV. Non-cable companies like Verizon have promised that they will never do throttling or capping. In fact a bill has been introduced to the US government to ensure plans aren't capped/throttled.

The US isn't the only other country in the world.

coffeebaron:
As cokemaster points out, they aren't managing, and more and more we are hearing of these overseas ISPs beginning to put restrictions on data usage.

See above.

Because we in NZ are on a much smaller scale, there just isn't the same economics of scale at play.

On a much smaller scale? The total population of a country has nothing to do with the viability of uncapped national traffic, and neither does population density for that matter.

There are also more choices overseas, e.g. UK or USA have several fixed line players including a large cable TV network. We only have Telecom, plus a bit of Telstra cable in Wellington / Christchurch.

Well yes, of course, competition = better deals. If we had more competition we would already have unlimited national traffic. No one is politely asking Telecom to kindly give us free national.

We don't have big apartment blocks: what do you think the economies of scale are for feeding a bit of fibre backhaul into a few massive apartment blocks, Vs a whole suburb with the same population?

Neither do a lot of places overseas. And you don't need FTTH to have uncapped national traffic. NZ's local infrastructure is fine for unlimited national traffic. The issue is backhaul and peering, both due to lack of competition. And what about NZ ISPs that offer uncapped local traffic?

I definitely think there is a lot more that can be done, but at the same time the grass is not as green as some people think on the other side.

No, the grass is infact greener than some think. Don't look at the US's relatively (compared to Europe and Asia) poor broadband situation and conclude it's the best in the world.

PenultimateHop

637 posts

Ultimate Geek

Trusted

#236935 19-Jul-2009 19:20

Screeb:
cokemaster:
Screeb: Other countries are doing just fine with unlimited national (and international for that matter) traffic.

Yes, indeed.

Great understanding of the issues there! The reason they're doing that isn't that the network can't handle it, but because Time Warner and Comcast are protecting their cable TV business revenues by making it impossible for users to download movies and TV shows instead of watching TV. Non-cable companies like Verizon have promised that they will never do throttling or capping. In fact a bill has been introduced to the US government to ensure plans aren't capped/throttled.

The US isn't the only other country in the world.

Actually, a lot of operators in the US and even Europe are looking at metering as a way of controlling their network investment[1] - by reducing the number of expensive-to-serve users. This includes non-MSOs who don't have a TV offering, so it is not necessarily related to that.

Screeb:
Because we in NZ are on a much smaller scale, there just isn't the same economics of scale at play.

On a much smaller scale? The total population of a country has nothing to do with the viability of uncapped national traffic, and neither does population density for that matter.

I beg to differ. My job entails network architecture for carriers and ISPs, and I have worked in many countries around the world. The total population, subscriber population, and population density absolutely have a huge amount to do with the viability of the network for reasons I outlined previously. They are not the only components to the equation (retail competition is obviously a big one), but they do have a significant impact to it.

Screeb:
There are also more choices overseas, e.g. UK or USA have several fixed line players including a large cable TV network. We only have Telecom, plus a bit of Telstra cable in Wellington / Christchurch.

Well yes, of course, competition = better deals. If we had more competition we would already have unlimited national traffic. No one is politely asking Telecom to kindly give us free national.

Absolutely agreed here. However there are some interesting market forces and economics at play which continue to leave New Zealand stifled.

Screeb:
I definitely think there is a lot more that can be done, but at the same time the grass is not as green as some people think on the other side.

No, the grass is infact greener than some think. Don't look at the US's relatively (compared to Europe and Asia) poor broadband situation and conclude it's the best in the world.

The grass is not always greener. As I've mentioned, I've lived and worked in a lot of countries and can categorically tell you that there are the same discussions and debates going on everywhere.

I do agree that the US is a poor example for retail broadband. Frankly they are well behind the curve compared to many countries. On the other hand, they have a highly developed and competitive IP transit, peering, and transport ecosystem which makes them very powerful in the global stakes. Little of the world except parts of Europe can compare to the US in that regard.

[1] I have been involved in network strategy discussions with quite a few global operators about this, including one of the MSOs mentioned in the links. The background material is very interesting.

Screeb

698 posts

Ultimate Geek

#236937 19-Jul-2009 19:32

PenultimateHop: There is so much misinformation in this thread, I will try to clear some of it up over a few posts as I have time.

I'll start with this one.

Three (well, two and a half) reasons:

1. Subscriber density. Many of the operators in other countries have subscriber bases which equal the population base of New Zealand. It's not unusual to see 'mid size' ISPs with 1-2 million subscribers. In New Zealand there's only one ISP which even comes close to that.

Even in Australia, a small DSL ISP might have 200K customers. A small DSL ISP in New Zealand is likely to have less than 20K.

When you have subscriber density in these numbers, economies of scale come into play. 10% of the subscribers still use 90% of the resources, but you have a huge amount more income to work with, allowing you to afford those 10% of subscribers. Even then it can still be tough for an ISP to actually make money. When your subscriber base does not give you massive income that is required to service those customers, you just don't have the flexibility in your business to allow unlimited downloads.

The user base of ISPs has nothing to do with their ability to provide unlimited local traffic. I have heard numerous times about how the best internet access in the US for instance, is provided by small local ISPs.

Also see Romania, which has >1300 ISPs, most of which have 50 to 3000 customers. They provide ~100Mbps unlimited plans cheaply http://en.wikipedia.org/wiki/Internet_in_Romania

2. Competition. In all of: access loops to the home, wholesale DSL networks, and wholesale IP transit. In many countries with bigger populations, it is cheaper to buy transit than it is to peer. This is certainly the case in the USA and in lots of European countries. If you can buy 10Gbps of transit to anywhere for $4000 per month, then obviously it gives you a lot of options for your subscribers vs. paying $200/Mbit!

Already adressed in my last post.

NZ does not have this competition, and does not (yet) have the bandwidth demand to provide it. Additional trans-Oceanic cables might help, I suspect an increase in bandwidth demand is more likely to provide the stimulus to reduce costs and introduce more IP wholesale competitors. There are various plateaus that are reached in order to bring costs down, and so far nobody is running circuits bigger than STM-16 into New Zealand. Breaking that plateau will help to reduce the wholesale costs to ISPs.

Of course, international transit bandwidth is only one component of an ISPs costs (although it is usually their second largest, after staff).

What do international cables have to do with national traffic? If anything, unlimited national traffic would reduce international traffic.

3. No distinguisher between peering, offshore, and onshore. Many countries just don't distinguish because they are either:

a. Sitting in a strategic location for interconnection (e.g. Singapore, USA)
b. Not on the very far end of a piece of wet string (e.g. Anywhere in Europe)
c. Are seen as the centre of the Internet (USA!)
d. Have a significant population with particular language requirements, which means the majority of their traffic is sourced locally (e.g. China, Korea, Japan, some European nations...)
e. Are in a highly competitive, developed, wholesale market where there is often a financial disincentive to peer vs. buy transit (USA, parts of Europe).

Already adressed (lack of peering in NZ is a simply a "political"/competition issue).

It is definitely also worth pointing out that there are significant infrastructure and wholesale costs in offering free domestic bandwidth, such as:

1. Increase in wholesale costs from your wholesale operator (e.g. TNZ). Their network is not dimensioned to sustain significant volumes of traffic over the subscriber loop, although this is definitely improving. However this is a significant stranglehold until alternatives become available for backhaul services, and TNZ deciding to alter their models.

2. Network investment at the ISP. This will include more or bigger BRAS - this is a significant investment in the millions of dollars. Additionally the ISP may now need to upgrade their core infrastructure and their switching infrastructure, as well as their accounting/metering infrastructure and all their support systems. This could be anywhere from $0.5M to $10M depending on the size of ISP, their support systems, and their vendor selection.

Again, this is a competition issue. OF COURSE it's more expensive to provide unlimited national traffic than not.

3. Assumes that there is actual content and business to drive this investment. At the moment there really isn't, but this is a usual circular chicken-and-egg debate.

Like you say, chicken and egg. Hell, it's essentially the topic of this thread!

As a final note, it's worth understanding that many of these "overseas operators who provide unlimited broadband services" oversubscribe and traffic manage their networks absolutely ferociously. I have unlimited 100Mbps HFC service at my home costing me NZD$130 per month, and I am lucky if I get beyond 2-4Mbps on anything - the highest I have ever seen was 40Mbps to a local speedtest server, and about 16Mbps from Akamai for AppleTV content. Additionally it is DPI'd (at every CMTS!), aggregate traffic shaped, per subscriber traffic shaped (punitively based on subscriber profile), and transparent proxied and content filtered (badly). This makes the overall performance frankly suck. I had better experiences in Australia and New Zealand - at least I had a much better chance of hitting my PIR.

Congrats, you have a bad ISP. I know of a person in Sweden who hosts high traffic web content from their house, totaling ~3TB/month with their residential plan costing < $100 NZD.

Screeb

698 posts

Ultimate Geek

#236940 19-Jul-2009 19:38

PenultimateHop:
Actually, a lot of operators in the US and even Europe are looking at metering as a way of controlling their network investment[1] - by reducing the number of expensive-to-serve users. This includes non-MSOs who don't have a TV offering, so it is not necessarily related to that.

That's the result of poor network investment and lack of local competition. Like I say, companies like Verizon have no problems.

Screeb:
I beg to differ. My job entails network architecture for carriers and ISPs, and I have worked in many countries around the world. The total population, subscriber population, and population density absolutely have a huge amount to do with the viability of the network for reasons I outlined previously. They are not the only components to the equation (retail competition is obviously a big one), but they do have a significant impact to it.

The grass is not always greener. As I've mentioned, I've lived and worked in a lot of countries and can categorically tell you that there are the same discussions and debates going on everywhere.

So your personal experience trumps various studies and reports released by organisations such as OECD that show findings to the contrary?

insane

3240 posts

Uber Geek

ID Verified

Trusted

#236945 19-Jul-2009 19:50

Need to remember that not all the current network equipment ISPs are using today easily scale to speeds great enough coupe with free national traffic.

While national traffic probably only accounts for around 20-25% of all usage, with free local traffic this would balloon with file sharing and other traffic that isn't actually contributing to business / economic growth.

ISPs will be having to upgrade their switching equipment, core routers, DPI equipment, handover links. Depending on the current equipment used some may even need to completely upgrade entire platforms if their current gear does not simply have upgradable components.

All this investment with no return as traffic is free? I can't see many ISPs wanting to sign up to that.

1 | 2 | 3 | 4 | 5