Which would really help us all...... NOT!!! The problem is that is exactly what more and more are actually doing � and with unlimited quotas there is nothing to stop them. Providing unlimited NBN quotas is a really, really dumb thing.
Of course it won't but if it goes to crap, then more people will complain and bring awareness to the issue.
People are asleep because they don't know whats happening, they will only notice when they get dialup speeds.
It is what it is, I remember there was quotas and I had to watch what I got or scheduled things in off peak.
After I got unlimited about 8 years ago. I didn't even bother. To be honest, I downloaded less because I didn't feel I needed to get my moneys worth.
Overseas works on this model, because people just don't care and use it when it they want to and the bandwidth was always there.
Enough said, the whole thing is a sham and the faster people get educated on what MTM is and how bad it is the better.
so created this virtual circuit which costs $17.50 per megabit
Just so effing dumb! No RSP is going to be able to offer what people expect at that rate, it should be $1.75/mbit.
If I was forced onto NBN i'd max out my connection getting linux isos then deleting and repeat.
To add to Fast is Good's post, that's the current problem in Au....so many people looking to get as much as they can from 'the system' and generally courtesy of others :p
s I understand it the CVC bandwidth purchased by RSPs from nbn for TC-4 is a peak information rate (PIR) not a committed information rate (CIR), so the RSP won't necessarily get the stated bandwidth 24x7 � do I have that right
No. At the level of CVC, its effectively a CIR. The RSP connects at the POI with 1 Gbps or 10 Gbps ethernet links. NBNCo simply drops packets when the rate for a CVC exceeds the purchased amount. There is nothing in the NBNCo infrastructure at the POI which is likely to be congested. However, connections to individual premises might be congested at times and so only a PIR is provided for the overall premises to POI link(TC4).
The ACCC successfully ran a performance monitoring pilot last year. They put probes in some people's homes (like TV program monitoring companies) and measured the performance of the Internet link over time. They are now seeking funds to runs this as an on-going operation and presumably to publish the results. Amazingly the industry have found some reasons to object to this.
The metrics selected for testing included download/upload speeds, web browsing time, latency, packet loss, video streaming, jitter and DNS resolution. A range of results were observed against these metrics, with a noticeable trend in deteriorating performance during peak use periods, particularly in relation to download speeds.
�The ACCC is greatly encouraged by the results of the Pilot Program; however a decision to proceed with a future program has not been made and any finalised program would involve further stakeholder consultation,� Mr Sims said.
�Similar broadband monitoring programs have been established in the United Kingdom (2008), United States of America (2010), Singapore (2011), and more recently, Canada. Such programs have led to improved transparency of information and increased performance-based competition for broadband services.�
Just so effing dumb! No RSP is going to be able to offer what people expect at that rate, it should be $1.75/mbit.
Sure. Why not make it free even? It's not like there are billions of dollars worth of capital to be paid off with the current network without considering the network upgrades that would be required to deliver the capacity 2 dollar back haul would require.
Maybe people should adjust their expectations to what can reasonably be expected for the amount they are willing to pay?
None of this CVC shit exists overseas.
Really? So how do wholesalers do interconnects in other areas? BT for example have interconnect pricing for wholesale services in their pricelists.
Seriously the links between all POIs should have unlimited capacity, its an internal national network. It doesn't cost $17.5 per mbit of bandwidth within this network.
Why should they have 'unlimited' capacity? How is that even rational? There should be racks upon racks of fiber cable termination and infinite switching backplane connecting them all together. And this miracle you created shouldn't cost $17.50?
The network cost 50 to 80 billion or more, depending on who is talking. If CVC is paying it back it would take 30 to 40 years at 1000 Gig of CVC. And that's not including the upgrades that would be required by the time they get to that kind of CVC dimensioning, like upgrading Nodes to 10 gig, or upgrading POI switching and interface capacity to handle it.
Or were you just worried about the cost of a single fibre and assuming everything else was magically paid for by the money fairy?
The price is higher than what it should be thats for sure especially given its ongoing and the RSP also needs to pay a access charge for the last mile.
I think if it was annual it would be closer to the correct figure
Id have to double check but if memory serves me correct ISP peering with pipe is like 3k pm for 1g and 5k pm for 10g, and thats with partial internet routes as well, pretty sure you can get a 10g interconnect with megaport for less than 10k per month, flap i can do metro wavelength cheaper than 10k per month.
I know there is a payback required but the price is way too high in this market for a single component that is charged per mbit, i understand all the costs involved as well.
Sure. Why not make it free even? It's not like there are billions of dollars worth of capital to be paid off with the current network without considering the network upgrades that would be required to deliver the capacity 2 dollar back haul would require.
The costs should come down. $17.5 per mbps will render the whole network useless. The plans are not viable if RSPs only buy 1mbps per subscriber. In case you havn't realised data needs have moved on. We are not living in year 2000 any more.
Really? So how do wholesalers do interconnects in other areas? BT for example have interconnect pricing for wholesale services in their pricelists.
In Asia you can go for your life and download as much as you want. Within reason the bandwidth is mostly there. The 1gbps plans at times your going to be getting 200-400mbps depending on sources. This is not 1gbps, but its not 1mbps either. Its usable internet.
Why should they have 'unlimited' capacity? How is that even rational? There should be racks upon racks of fiber cable termination and infinite switching backplane connecting them all together. And this miracle you created shouldn't cost $17.50?
Why shouldn't it be unlimited? If NBN / MTM planned properly they would be pulling in many strands or provisioning enough bandwidth / backhaul so that it never becomes the issue. Fibre can do 255 Tbps in the labs on a single strand. Come down a bit and I'm sure there is heaps of bandwidth there. It should never be a real issue.
The most expensive links for the RSP should be their international links, this nations links should not be that because its should be cheaper to run fibre within this country then transcontinental. The whole thing stinks and it extends to this whole countries rip off nature. Literally rip everyone off and its no different with the internet.
Yes the network should pay itself, why not have RSPs commit to the same amount of average cost per user, but give them 10 times the CVC. Making it $1.75 mbps (even this is too much imo).
CVC is an artificial construct that has no place in a national network. It should cost nothing if you are sending data between 2 NBN users. This is what will affectively give us the competitive edge globally.
Enough said, the network is shit, enjoy your congestion. I'll make sure I do my part if I'm forced to go onto the NBN. I'm staying on ADSL2+, currently enjoying my 20/1 (nearly anytime I use it, I get 15mbps+). If this gives me a better experience than fighting for the last chip like seagulls, clearly the NBN has failed the nation with its artificial limits.
The price is higher than what it should be
Based on your feelings? You don't use any math or justification for it.
I think if it was annual it would be closer to the correct figure
So back to 2 dollars per meg with no idea of how to provide the infrastructure to handle those volumes.
Id have to double check but if memory serves me correct ISP peering with pipe is like 3k pm for 1g and 5k pm for 10g, and thats with partial internet routes as well, pretty sure you can get a 10g interconnect with megaport for less than 10k per month, flap i can do metro wavelength cheaper than 10k per month.
And NBN's NNI costs are $200 per month for 1 gig and 400 per month for 10 gig. With pricier tiers when you want extended range 1 gig.
A fairer comparison would be Telstra's AGVC charges which perform the same geographical aggregation function. Until NBN came along they were riding at over $33 per meg for the same type of service. And that didn't include things like satellite base stations in the mix.
CVC is half the price AGVC was 3 years ago and is providing a heck of a lot more for its cost.
I know there is a payback required but the price is way too high in this market for a single component that is charged per mbit, i understand all the costs involved as well.
Based on feelings again. The money needs to be paid back. The price of AGVC, CVC and NNI needs to pay it back. There isn't another chargeable element that will pick up 40 billion of slack for you. And while AGVC and NNI costs are fixed, CVC costs increase as you need to add more interfaces or upgrade interfaces and switching fabrics from 1Gig to 10Gig to 100 Gig capable.
with 93% fttp the cvc made some sense, its was a way to get back some of the initial spend, the pricing could have been adjusted if it was too high. But with cost blow outs and lower returns the cvc has become the only way to slow down the bleeding. it n o longer makes sense.
What I don't really like is this term in this thread 'all you can eat hogs' Whats with the attitude ? The nbn was for exactly that, giving us access to faster speeds and more content. I for one intend to have at least one connection that syncs bac,kups from our remote sites, so I expect that will be fairly well utilised, at 100/40 fttp unlimited, and that's exactly what this network was supposed to be built for, if it cannot handle todays requirements then it really is not fit for purpose
everyone should be an all you can eat hog, international quota charges are tiny, the cvc is just a way to make the nbnco profitable and make back some of the investment, if it was a proper fibre network it would have been fine CVC would have gone down quickly as people demanded the faster plans, hell I would be getting 1Gbps if anyone actually offered it. nbn does but no rsps are game.
A fairer comparison would be Telstra's AGVC charges which perform the same geographical aggregation function. Until NBN came along they were riding at over $33 per meg for the same type of service. And that didn't include things like satellite base stations in the mix.
CVC is half the price AGVC was 3 years ago and is providing a heck of a lot more for its cost.
And this is the BS cost we had to put up with since the internet was introduced in Australia. Telstra has ingrained these monopolistic charges for internet in this country. Why should this be the norm? when its not anywhere else in the world.
It does not cost $33 per mbps or even $17.5 mbps for local transit data.
Sorry I don't support these gouging practices. The whole system needs to collapse for any sane price to be introduced. What better way would that be to expose the problems and get your fair share when you sign up to a service.
It will only get worse as time goes on. I hope more applications like netflix comes along. More bandwidth hogging apps the better.
For the record what good is a national network, when it is so expensive to send data between 2 homes within the same city.
Seriously the links between all POIs should have unlimited capacity, its an internal national network. It doesn't cost $17.5 per mbit of bandwidth within this network.
100% agree. It really makes you think about what the primary objective of the NBN is doesn't it?
it's kind of like a tax to help pay for the NBN
Whoa! Hold up! Isn't the entire NBN tax payer funded?
Whoa! Hold up! Isn't the entire NBN tax payer funded?
No. As soon as the entity received its first cheque from an RSP it started on the road to sustain itself.
The whole thing stinks and it extends to this whole countries rip off nature.
This country does have a bit of a "rip you off culture", I agree.
Amazing how people try to screw each other for a percentage.
The costs should come down. $17.5 per mbps will render the whole network useless. The plans are not viable if RSPs only buy 1mbps per subscriber. In case you havn't realised data needs have moved on. We are not living in year 2000 any more.
I realise that. Except increased data needs also have increased costs. And RSPs are still trying to charge the same as they did for DSL services at half or less the speeds and run the same contention ratios.
The problem is consumers want the providers to magically absorb the costs because apparently businesses can run without actually recouping their costs.
In Asia you can go for your life and download as much as you want. Within reason the bandwidth is mostly there. The 1gbps plans at times your going to be getting 200-400mbps depending on sources. This is not 1gbps, but its not 1mbps either. Its usable internet.
Is this Asia places like Bali, or rural China, or Tibet? Or does your version of Asia only consist of Hong Kong, Seoul, Singapore and Tokyo? Because I've been outside the cities and those plans are magically not available.
Singapore has about a million more people than Sydney in a tenth the area. And it isn't subsiding aggregation for a rural Singapore region outside of metro Singapore and doesn't have a requirement for satellite services. Your cable runs just for metro are a tenth as long, your points of aggregation are 100 times larger in capacity. And that's just for metro, which is all most of those city plans care about.
Why shouldn't it be unlimited?
Because infinity is an unattainable concept.
If NBN / MTM planned properly they would be pulling in many strands or provisioning enough bandwidth / backhaul so that it never becomes the issue
Yes because fibre is the only cost here right? Things like interfaces and switching fabric don't have to be dimensioned and increased to increase capacity. The fibre cable is the cheap fixed price bit. And the only thing you look at apparently when doing your calculations.
Fibre can do 255 Tbps in the labs on a single strand. Come down a bit and I'm sure there is heaps of bandwidth there. It should never be a real issue.
And a 255 Tbps card will be free when they release them? Or will cost a few million? And the kind of switch fabric that can handle 40 Pbps will also be supplied free of charge. And there will be no power or aircon capacity required with this increased throughput.
Try buying 10Mbps back haul from any provider. Now tell them to deliver 20Mbps for the same price because the fibre can handle it. Watch as you get laughed at.
he most expensive links for the RSP should be their international links, this nations links should not be that because its should be cheaper to run fibre within this country then transcontinental.
They aren't real links, you get that right? The CVC will deliver connectivity from multiple exchanges to the POI, where international links are generally point to point links.
But then you are fixated on the fibre as a cable. How many optical transmission devices are involved in the link from here to the US? How many are involved in connecting up a single FTTN POI? There is far more active equipment in one POI region than in any international link.
The whole thing stinks and it extends to this whole countries rip off nature. Literally rip everyone off and its no different with the internet.
You have that rip off nature here. The network costs $X to build but you don't want to pay for it. You want it at your price and to heck with who actually foots the bill. You still can't talk about how you would repay the Capital to build the network, because you don't give a crap. Only about what you want to take for whatever you want to pay.
Yes the network should pay itself, why not have RSPs commit to the same amount of average cost per user, but give them 10 times the CVC. Making it $1.75 mbps (even this is too much imo).
How does that work? They need to increase network capacity tenfold in the same revenue. And please try to understand that network capacity involves switching fabric and optical transmitters and isn't just fibre in the ground. So you have increased costs drastically while retaining revenue.
Until you get to the point where it breaks because back haul from a DSLAM doesn't allow you to meet CVC capacity. You're running dual 10 gig links, (How you bought the interfaces with your reduced revenue model is a mystery...) but now you need to split the chassis in two and add more interfaces, but yet again with what money?
I know you will still say there is sufficient fibre in the ground, which is true, but totally irrelevant to actually addressing this issue.
CVC is an artificial construct that has no place in a national network. It should cost nothing if you are sending data between 2 NBN users. This is what will affectively give us the competitive edge globally.
So the network between the two should cost nothing to build and maintain. It should just have magically risen from the ground fully realised? Yeah not paying for network maintenance and build will give us that 'competitive edge'. I'll buy into your daydream when you explain how your model repays just the capital build over even 50 years. Try to at least include some OPEX and capacity upgrades in your figures.
Amazing how people try to screw each other for a percentage.
I think that's called "capitalism" and "free enterprise".... but comes down to human nature I'm afraid!
Whoa! Hold up! Isn't the entire NBN tax payer funded?
If it was it wouldn't have to be paid for under user pays..... it would be like the Sydney Harbour Bridge or the Snowy Hydro scheme.... but in this case we (the users) have to pay for it � not the taxes everyone pays.
Based on your feelings? You don't use any math or justification for it.
$17.50 per Mbs per month (on top of the port costs) to get data the 10kms from my door to my poi.
An ISP will get internet connectivity, so from their presence to anywhere in the world for $2 per Mbs.
A fairer comparison would be Telstra's AGVC charges which perform the same geographical aggregation function. Until NBN came along they were riding at over $33 per meg for the same type of service. And that didn't include things like satellite base stations in the mix.
No its not a fair comparison. Read the reports from Akamai. Because of Telstra, bandwidth in Australia is the most expensive in the world. The $33 per Mbs is a ridiculously expensive charge and only possible because Telstra is a monopoly.
NBN was designed to do it better. Under FTTP, CVC charges were to drop significantly. They aren't actually required � they are and artificial charge designed to constrain bandwidth.
CVC is half the price AGVC was 3 years ago and is providing a heck of a lot more for its cost.
Yet internet traffic in 2016 is 5 times higher than 2013. And it will increase another three fold in the next five years.
The CVC costs are rediculous, and do not reflect the true cost of carrying data. Its just being used to try and recoup the screwed up MTM financials.
How does that work? They need to increase network capacity tenfold in the same revenue. And please try to understand that network capacity involves switching fabric and optical transmitters and isn't just fibre in the ground. So you have increased costs drastically while retaining revenue.
Until you get to the point where it breaks because back haul from a DSLAM doesn't allow you to meet CVC capacity. You're running dual 10 gig links, (How you bought the interfaces with your reduced revenue model is a mystery...) but now you need to split the chassis in two and add more interfaces, but yet again with what money?
I know you will still say there is sufficient fibre in the ground, which is true, but totally irrelevant to actually addressing this issue.
384 people max on a Node. 20G is over 50mbs per user. Thanks to the way internet traffic works, you wont be maxing that out at all.
However, as Ive pointed out before, there are three unavoidable realities the MTM NBN will face:
1. It will remain a CVC constrained mess.
2. Internet costs will have to rise $50 or so per month.
3. NBN will take a massive financial hit.
If NBN continues to ignore it, NBN will manage to achieve all three.
FTTP has exactly the same problem as FTTN, it's just that there's fewer subscribers on the shared medium.
Yeah... should have said fairer!
You potentially could suffer the same issue but the bandwidth per end user is about 10 times higher so is far less likley as CVC congestion problems will arise well before the GPON bandwidth becomes congested (unless CVC prices miraculously fall by a lot!)
The problem is all modelling for internet provision assumes a fairly high contention ratio (which was fine before the online streaming became so popular).... they have to re think the supply and pricing to take this change into account but are yet to refocus on end user requirements...... this has to change or the internet is going to be pretty unusable at peak times.
Disagree; FTTP has exactly the same problem as FTTN, it's just that there's fewer subscribers on the shared medium.
Have 32 people on the same GPON all order 4x services each and you will get congestion too.
Yes, however the chance of that is 0. And if it does, you can simply switch the FSAM to 10GPON � which will coexist with GPON.
The chance of FTTN hitting its limits right now is near 0. However, this will need to last us 20 or so years, and the chance of it hitting its limits in even the next couple of years is much greater than 0. Cisco expects data usage to treble in the next 2 years. And their predictions have been the most credible, unlike the laughable Vertigen report that MTM was based off of.
All services have their limits. Even POTS � if everyone tried to make a call at the same time, it wouldnt work. It was designed to service a typical capacity. GPON will do this for the foreseeable future, and when it doesnt, its a very simple upgrade to 10GPON.
You are aiming at option 2!
Well, yes. It's the only viable option and it avoids #1 and #3.
One of the problems with this plan � you have significant number of people who want a phone and only a phone
Not really. You don't need the NBN to offer phone service. You have multiple options: NBN, copper, mobile phones, fixed wireless and satellite. Pick the cheapest option and price accordingly. If the cheapest option is deemed to cost too much, then the government can apply direct subsidies as needed and leave the rest of us alone.
NBN, copper, mobile phones, fixed wireless and satellite. Pick the cheapest option and price accordingly. If the cheapest option is deemed to cost too much, then the government can apply direct subsidies as needed and leave the rest of us alone.
Copper is gone after 18 months. FW is only in rural areas. For the majority of the population, we'll only have NBN.
Under FTTP, this is by design. From a whole country perspective, there is cost savings in only running the following networks: FTTP, FW, Sat, Mobile.
Under MTM, we cant ditch the legacy of copper. Not only are we running FTTN, we are still running a large number of copper based services that FTTN and HFC cant provide. This isnt nbn, but the people of the country have to pay it all the same. All this was to be migrated to fibre. So NBN saves on the CapEx (which is doesnt) and we all pay through the nose for the extra maintenance of all the extra networks.
Disagree; FTTP has exactly the same problem as FTTN, it's just that there's fewer subscribers on the shared medium.
Have 32 people on the same GPON all order 4x services each and you will get congestion too.
Except you are overlooking a few things, the original FTTP design does allow for this...
Firstly, the original design didn't fully utilise the 32way way split, this was to provide for future growth etc.
Secondly, NBN Co could "rebalance" the high demand users across different splitters in the FDH.
Thirdly, there was an additional 12f ribbon allocated per FDH which could be used if absolutely required.
Finally, if the demand at the FDH was so great NBN Co couldn't use the first three options they could upgrade to NG-PON2, 10GPON or whatever was most appropriate at the time.
Too bad it will never happen, as it would be political suicide just to suggest it.
Really? Have you looked at the ARPU figures over time? I don't think they look too bad, adjusting for inflation and making some assumptions for business revenues.
There is nothing in the NBNCo infrastructure at the POI which is likely to be congested. However, connections to individual premises might be congested at times and so only a PIR is provided for the overall premises to POI link(TC4).
What causes the local congestion then? So you're saying that a relatively small but unknown number of connections at a POI may be congested at times due to factors other than CVC bandwidth purchased by the RSP, and therefore beyond the RSP's control?
How can/do RSPs monitor this, if they are even aware when it occurs? Does nbn provide them with diagnostic tools / reports, or are they simply left to argue it out with customers? If the customer wins the argument (by jumping through enough of the RSP's hoops?) does the RSP simply log a fault with nbn who will address it, or is it deemed acceptable because of the PIR thing and nbn will just ignore until they are ready to fix/upgrade later?
Would be good to know what goes on behind the scenes at nbn and RSPs when things like this occur.
What causes the local congestion then?
One thing that is often overlooked because it is not in the "final" design is the temporary network elements (POIs and FANs etc).
These clearly have limited bandwidth, it is clear that the higher AVC tiers are not available in the temporary sites. You can't get the 250/100, 500/200 or 1000/400 everywhere yet (the CVC costs are a different issue).
I am also suspicious that NBN Co might be using 3rd party providers for the FTTN cabinet backhaul, but I have no hard evidence.
I am also suspicious that NBN Co might be using 3rd party providers for the FTTN cabinet backhaul, but I have no hard evidence.
What's there to be suspicious about? Go digging in the docs, they freely admitted they used Telstra Fibre to speed up the rollout. Same for HFC. That's why FTTN cabinets only have 2x 1Gbps, when it would've been a minuscule extra cost to use 10Gb interfaces. This is also allegedly the reason with HFC is only available at 100/40 at launch.
Really? Have you looked at the ARPU figures over time? I don't think they look too bad, adjusting for inflation and making some assumptions for business revenues.
Yes, really.
This isn't about ARPU or some other rational thing, it's about politics.
It's politically untenable to raise the base cost of a NBN connection. It would raise granma's costs to check her email and the NBN was supposed to be affordable, so we can't have that.
It would also require politicians to admit they were wrong and face the cold, hard facts of what a cluster**** the NBN has become. No true politician is capable of either.
What causes the local congestion then? So you're saying that a relatively small but unknown number of connections at a POI may be congested at times due to factors other than CVC bandwidth purchased by the RSP, and therefore beyond the RSP's control?
NBN has already experienced FTTN nodes in some areas hitting congestion in peak-periods, due to a disproportionate number of heavy users . Fixed Wireless has also experienced congestion due to a "higher-than-expected" uptake of higher speed services in specific regions.
The way we deal with it as RSPs in lodge speed faults to NBN, get told it's congestion and then told there's nothing we can do about it and should move customers to lower speed plans to minimize the impact.
What's there to be suspicious about? Go digging in the docs, they freely admitted they used Telstra Fibre to speed up the rollout.
The Network Design Rules say they should be allocating a single 12f ribbon per FTTN, that is not third party infrastructure.
I know they talked about it for the transit network, using dark fibre from other providers but I didn't think it was included in the DFN/LDN.
Copper is gone after 18 months.
Only if you chose to decommission it. If it's cheaper than other options to use the copper to provide plain phone service then it should be kept. If not, good riddance.
FW is only in rural areas.
Fixed wireless can be extended into urban areas to provide phone only service if need be. Obviously this is not needed if there are other cheaper options.
For the majority of the population, we'll only have NBN.
And mobile phone service. Which is probably the easiest and cheapest way to provide voice only service. The government already pays for some rural towers. It can pay for a few urban towers too to cover problem areas.
If we go for providing voice only NBN service then at least the end user price should cover all the direct costs of providing that fiber drop. People with no other voice options should qualify for subsidies, if needed.
From a whole country perspective, there is cost savings in only running the following networks: FTTP, FW, Sat, Mobile.
Agree.
With respect, I think this thread is going slightly 'off topic' � what I was trying to get at was some way of customers / subscribers having some basic but reliable info on which RSP's are having bandwidth / CVC issues.
Much like a 'league table' so customers and potential subscribers can have a benchmark of RSP's from which to choose from.
Cheers
Ernie
what I was trying to get at was some way of customers / subscribers having some basic but reliable info on which RSP's are having bandwidth / CVC issues.
I think the discussion just highlights how hard (and sadly impractical) it would be to achieve this. Also it wouldn't really be fair to the ISPs involved as they could be a victim of another ISP's overuse at a node etc.
It really is a can of worms and whilst not being able to find out who is the best provider maybe hard for the customers � it is a product of bad design and bad implementation by NBNCo (under Government direction). You can't punish the ISPs for not being able to deliver when the foundations of their network is a crumbling pile of deep, smelly doggy doo doo! It is Telstra off net all over again � only far worse. The ISPs have no control over the most important part of the network.
With respect, I think this thread is going slightly 'off topic' � what I was trying to get at was some way of customers / subscribers having some basic but reliable info on which RSP's are having bandwidth / CVC issues.
The only answer is to find an RSP that publishes network graphs.
Only if you chose to decommission it. If it's cheaper than other options to use the copper to provide plain phone service then it should be kept. If not, good riddance.
But its more expensive to run multiple networks.
So you could have many heavy users at a node (they don't even have to be on your ISP) and the backhaul from that node is full whilst there is stacks of CVC capacity at the POI � and is being totally unused. (The ISP could purchase enough for 1:1 contention and this would still be the case)
Not really. It's a theoretical model but in practice with ISPs only buying 1 meg per user of CVC, you need a lot of completely idle services elsewhere to let those users dominate the uplink. And not only must you have idle services, but your heavy users must all be concentrated in that node, because heavy users in other nodes would grab that spare capacity.
So while I can envisage what would need to happen, I can't picture it happening except in the early rollout, where ISPs get their free 150Mbps CVCs and only have a handful of customers in a POI when the first area gets turned on. But after that with a diverse user base the odds of having enough free capacity and having all the heavy users in one area is pretty remote.
And all traffic will attempt to use the backhaul and will only be dropped by NBNCo when it reaches the POI and is dropped at the CVC as it is handed to the ISP. The ISP has no say in what gets dropped � only the amount which gets passed (depending on how much they are prepared to pay).
Your traffic is travelling in the wrong direction. Downstream, the congested path, goes Internet > ISP > CVC. So the ISP definitely gets to police their traffic before NBN has a crack at it.
Some ISPs will have more higher usage customers than others, there will be more of these high end users at some nodes
Yes but those higher usage customers are still only buying a meg of CVC. So an ISP with more high usage customers will have more contention for their limited CVC and so each user will get less down the pipe. A mix of high usage customers is self defeating, because CVC blocks them before they saturate the back haul.
If you have 100 high usage customers with 100 meg of CVC to play with and they all suck as hard as they can they each pull down only a meg each. Which sort of makes them not be high usage customers. That composition could never saturate a back haul.
You need a mix with a select few high usage customers down one back haul only and the rest of the customers of that ISP for that POI being essentially idle to make it break. And you would need 40 or 50 idle customers for every high usage customer. And the odds of that particular deployment of customers is pretty slim. Fantastic if you are one of the high usage people but incredibly unlikely,
But its more expensive to run multiple networks.
That's why I said pick the cheapest option. Keep if it's the cheapest option, decommission if it's not.
A very simple solution which solves all problems is to set CVC to zero and raise the AVC price to cover the full line rate cost to the POI, operations and maintenance, upgrades and the prorated share of the NBN build costs.
No, it doesn't.
You've just designed a cost model where either a 100/40 service is unaffordable, or your back haul melts down because everyone is trying to go full speed. Which means you sort of end up back where you started with no way out.
Charging for CVC is just an ugly hack which is used to paper over the financial problems of the NBNco.
No. I'm not even sure what the 'hack' is. If NBN needs 80 bucks per subscriber per month then it doesn't matter to them whether it comes in in CVC or AVC form. The advantage that CVC gives is that as traffic grows the revenue grows to finance upgrades to back haul.
CVC was designed as a tool for ISPs, and there are white papers out there about it. An ISP could offer a low CVC low usage network to people like my folks who 'do email and Facebook'. They could offer a different CVC amount and contention to those who want SD streaming and another to High Def / download the Internet types. And each user type would be paying according to their needs. Where a flat rate means non users would be subsidising over users, and that kind of economic environment encourages more over users.
Since all costs are fixed, there is no need for a variable CVC cost.
They aren't fixed. Upgrading from 1 gig to 10 gig interfaces costs money. Adding a second interface costs money. Adding switching infrastructure, and the power and cooling to support that infrastructure costs money. And those costs are incurred as more bandwidth is used, not based on the AVC purchased or on a per user basis.
To have fixed costs you would need to move from PIR to CIR. Which would give you fixed costs. Massively high fixed costs, but still...
You want a service. If you don't get that service from a RSP you can do the following:
Call their support. Hopefully they buy more capacity.
If they don't you can try and get a break in contract (if you are) or move to another RSP.
Contact TIO.
All of which costs potentially thousands of dollars in time by the time you deal with outsourced customer support and obscure escalation processes. It would be far quicker to have the information available up front to consumers.
You've just designed a cost model where either a 100/40 service is unaffordable
Incorrect. You are stuck in the $17.50 CVC mentality. The correct reference price is NBNCo's current ARPU. The fixed monthly cost of a 100/40 service where CVC is zero is roughly equal to the amortized cost of a single NBN line to which you add the line's prorated share of NBNCo's operating costs.
Of course if your definition of unaffordable is more than what we pay now or anything less than what it actually costs to provide the service, then, yes, it will by definition be unaffordable.
or your back haul melts down because everyone is trying to go full speed.
That's a whole another problem, one which we are better equipped to deal with via competition in the backhaul market. Something we can't have with the NBN.
Which means you sort of end up back where you started with no way out.
Quite the opposite. My model solves all problems both technical and financial.
I'm not even sure what the 'hack' is.
Artificially inflating the cost of CVC. The true marginal cost of CVC capacity is zero and the true dollar amount cost can easily be covered by a fixed monthly fee.
If NBN needs 80 bucks per subscriber per month then it doesn't matter to them whether it comes in in CVC or AVC form.
It matters a great deal. Charging for CVC suppresses usage and causes congestion. This is obviously not something we want. Charging for AVC aligns costs and prices, and it encourages usage, something we all want.
The advantage that CVC gives is that as traffic grows the revenue grows to finance upgrades to back haul.
It's just a bad idea to nickel and dime ISPs with CVC charges. It is far better to charge ISPs the fully loaded cost of the AVC and be done with it. Revenue still grows as AVCs are upgraded and more AVCs are ordered.
CVC was designed as a tool for ISPs, and there are white papers out there about it. An ISP could offer a low CVC low usage network to people like my folks who 'do email and Facebook'. They could offer a different CVC amount and contention to those who want SD streaming and another to High Def / download the Internet types. And each user type would be paying according to their needs. Where a flat rate means non users would be subsidising over users, and that kind of economic environment encourages more over users.
The CVC charges are just a way for the NBNco to paper over the failings of their finances. The CVC is totally unneeded and an entirely artificial construct. The NBNco does need to throttle users bandwidth. The scenario you described can just as well be handled by grouping customers and providing different amounts of backhaul to each group.
They aren't fixed. Upgrading from 1 gig to 10 gig interfaces costs money. Adding a second interface costs money. Adding switching infrastructure, and the power and cooling to support that infrastructure costs money. And those costs are incurred as more bandwidth is used, not based on the AVC purchased or on a per user basis.
Define: fixed costs
noun
"business costs, such as rent, that are constant whatever the amount of goods produced."
All the things you list are by definition fixed costs. It does not cost more or less to run a 1GE or a 10GE interface depending on the amount of traffic.
Yes, it costs more to buy more line cards or upgrade to 10GE, but it is still a *fixed* cost to do so.
Fixed costs can be spread out amongst subscribers at a fixed monthly cost. A network infrastructure will support a fixed amount of AVCs at line rate. Once you go above that you buy more equipment, but it's still a *fixed* cost per AVC as you upgrade.
To have fixed costs you would need to move from PIR to CIR. Which would give you fixed costs. Massively high fixed costs, but still...
We know how to build linerate networks. It is neither hard nor expensive.
Define: fixed costs
noun
"business costs, such as rent, that are constant whatever the amount of goods produced."
All the things you list are by definition fixed costs. It does not cost more or less to run a 1GE or a 10GE interface depending on the amount of traffic.
You're really struggling here. You go to the dictionary and then fail to comprehend it.
Yes, it costs more to buy more line cards or upgrade to 10GE, but it is still a *fixed* cost to do so.
Nope, that makes it not a fixed cost.
Fixed means it is constant regardless of the output. You're admitting that the costs change depending on the output. The fact that they change by a *fixed* amount still means they changed which means they are not fixed costs. The cost must remain constant for it to be a fixed cost, which means it cannot change.
I believe the correct term is step fixed cost:
"A step fixed cost is a cost that does not change within certain high and low thresholds of activity, but which will change when these thresholds are breached."
However, I'm not going to quibble with you. I already annihilated your main argument point by point, which is enough for me.
what I was trying to get at was some way of customers / subscribers having some basic but reliable info on which RSP's are having bandwidth / CVC issues.
Are these ACCC Reports any help? I guess at a very high level it may give a vague national indication by access seeker group, though the information will always be in retrospect.
http://www.accc.gov.au/media
https://www.accc.gov.a
Unfortunately the data provided is obtained from nbn only and therefore can only show NBN access seekers rather than RSPs specifically: As a wholesale access service can be used by an NBN access seeker to itself supply a retail service, or alternatively to supply a wholesale service to another NBN access seeker or a retail service provider, the report does not provide a view of the structure of NBN retail markets.
Also, it does not show stats for individual POIs, though there is a break down by TC-4 AVCs for access seeker group and state. The count of TC-4 AVCs by data download rate and access seeker group is in a separate table, which again is unfortunate.
what I was trying to get at was some way of customers / subscribers having some basic but reliable info on which RSP's are having bandwidth / CVC issues.
that's not going to happen, realistically. NBN won't release the info and I doubt the ISPs will, so at best you're stuck with anecdotal reports from users.
It will happen if the Govt funds the ACCC's proposed monitoring regime. This has been tested successfully so its just a matter of getting the go ahead. People need to pressure MPs to get this to happen.
It will happen if the Govt funds the ACCC's proposed monitoring regime.
And if they did it would expose the NBN for what it is..... a poorly thought out implementation as a result of poor Government direction.
The Government aren't going to be real popular if the taxpayer has to fund a failing NBN(Co) due to lack of funds or if our monthly fees have to increase by say $50 so that ISPs can afford to throw enough money at NBN(Co) so they can pay down their dept..... (paying for expensive CVC or some other charge to raise the income).
They would appear to be the only two options � the third being leave it as it is and blame everyone other than the Government � and if you were the Government making that decision what would you do? Bit like the Government starting a blog with their election promises and ticking it off as they deliver on each......
And if they did it would expose the NBN for what it is..... a poorly thought out implementation as a result of poor Government direction.
how does ISPs failing to provision adequate CVC capacity in any way translate to a network implementation?
if anything, blame the ACCC for insisting on 121 POIs.
The Government aren't going to be real popular if the taxpayer has to fund a failing NBN(Co) due to lack of funds or if our monthly fees have to increase by say $50 so that ISPs can afford to throw enough money at NBN(Co) so they can pay down their dept..... (paying for expensive CVC or some other charge to raise the income).
that ship has sailed. 121 POIs it is, and the network costs what it costs. it doesn't matter how you structure, package or restate it, the end users will end up paying for it. the government is not running the NBN as a loss to improve people's internet connections.
I believe the correct term is step fixed cost:
Hey you're the idiot who quoted the dictionary for the wrong term. I mean you didn't just pick the wrong term, which happens, but you actually supplied the dictionary definition for it to prove your wrongness.
However, I'm not going to quibble with you. I already annihilated your main argument point by point, which is enough for me.
By quoting the wrong definitions and repeating the same opinions without any supporting evidence? That isn't how you annihilate any argument.
And now that you 'corrected' your term to 'fixed step cost' you cannot have a fixed AVC cost based on a fixed cost, so you kind of annihilated yourself. Again.
And now that you 'corrected' your term to 'fixed step cost' you cannot have a fixed AVC cost based on a fixed cost
Do you have any facts to support the above statement?
I think you are wrong, and I refer you to my previous post for the underlying reasoning.
Going all ad hominem is not going to help your argument, it just reflects poorly on you. You'd do much better trying to refute any of my points with facts and logic. If you can manage, that is.
Không có nhận xét nào:
Đăng nhận xét