Geekzone: technology news, blogs, forums
Guest
Welcome Guest.
You haven't logged in yet. If you don't have an account you can register now.

View this topic in a long page with up to 500 replies per page Create new topic
1 | 2 | 3 | 4
1332 posts

Uber Geek
+1 received by user: 152
Inactive user


  Reply # 473321 24-May-2011 09:44 Send private message

michaeln:
1080p:
I cannot see a reason that unlimited or at the very least extremely capacious plans are unfeasible.


The two are not at all the same thing.  I would very much like the cost to come down/caps to increase, but 'unlimited' has some problems, to wit:


I didn't say they were the same. I think for all but a few users, however, 1TB per month and unlimited would be close to one and the same.

michaeln: ~ mathematical distribution discussion ~


The problem is much simpler than you make it seem. There is no normal distribution of internet usage as you have stated and this actually works in favour of the ISPs. According to Cisco the 'average' amount of data consumed per month across the globe is close to 15GB per month: javascript:mctmp(0);

This means that the majority of customers are using a very small amount of data per month and in most cases paying much more than they should for the service consumed.

What happens is that most ISPs simply have to provision enough bandwidth to service their peak usage period and make that profitable from their customer base. This leaves huge amounts of unused bandwidth during the aptly named off peak periods.

So the situation is as you have stated but in reverse. Most customers are subsidising the ISP already by underusing their internet connections. Consider that the '1TB per month pirate' only uses the 'average' monthly connection of 66 customers and when you have an average base of customers using much less than 15GB per month the math will average out nicely if you ensure a good ratio of customers using small amounts of bandwidth to those using more.

You only run into issues when the heavy bandwidth users make use of the available bandwidth during peak times. Otherwise the ISP could care less if they were leeching all night as most of their customer base is asleep or at work.

michaeln: Of course, the usage pattern would change. At the moment, even the heaviest users are still constrained by cost. That constraint would vanish and they would use even more. At the moment the average users are also constrained by cost. They would certainly feel justified in using more, given that their usage was now unconstrained and that their charges had gone up. That would inevitably lead to higher usage, leading to higher costs to the ISP, with no concomitant revenue. The result would of course be further price rises.


The first part of your claim is true but the second does not follow and has been shown to not be so elsewhere. The solution for ISPs is to charge accordingly. Most users will not make full usage of their connections, a small minority will always want to push the limits. Allow for both and charge for the service.

A simple answer to me would be to offer a set of base plans like this:

Regular Usage (the majority of customers, usage during peak times mainly so bandwisth priorotised for this pool during that time otherwise equal to other traffic during off peak)

- 25GB
- 50GB

Gaming (require a low jitter connection, value a lack of congestion over data consumption)

- 80GB (guaranteed bandwidth pool 24/7 for better ping etc..., at a premium)

Heavy Usage (leeches, torrenters, other heavy usage, regular priorotisation, makes use of bandwidth when available)

- 250GB
- 500GB
- Unlimited

The heavy usage plans will all be priced differently but the key here is to maintain a ratio. You need a system to ensure that your percentage of 'unlimited' users do not saturate your base of regular customers. If you reach the maximum number of sustainable 'unlimited' users then the option needs to be removed from your sales page. The same would apply for the 250 & 500GB plans as well as, to a lesser extent, the gaming plan.

Making this idea known to to your customers will make you very popular I would think. This is the kind of transparent business that lands you with a loyal customer base.

michaeln: This contention for a common resource pool is exactly the classic “Tragedy of the Commons”, where game theory predicts that pursuing the most rational individual strategy inevitably leads to destruction of the resource and a loss for all concerned.


With a policy of transparency to your customer base and sticking to a ratio of customers this will never happen.

michaeln: Usage charging. Usage charging/data caps has the following desirable characteristics:





    • It matches costs







    • It is easy to do





Simplicity here results in low cost, compared to complex methods which result in the customer paying largely for the privilege of being sent a bill.


You forgot to add that it creates a commodity where one does not exist (data); instead customers should know that what they are paying for is contention. I also think that using a ratio system would be much easier than calculating data usage per customer.

michaeln: The fundamental driver behind the success of the Internet is that all power and innovation comes from the edge. The Internet is not driven by the carriers deciding what services to sell, but by end-users creating innovative new services to be carried on the Internet.

Volume charging doesn’t care what is carried, so there is zero incentive for the carrier to prefer one sort of traffic to another and hence to try to control what its customers use the Internet for.


I would argue that volume billing does care what is carried, specifically, large volumes of data. This can be seen in the trend to zero rate certain websites over others. This does not look like an incentive not to care what data is carried, does it?

My point here is that an unlimited volume of traffic is not something that cannot be done, it simply needs someone with the ability to think outside the box. 

178 posts

Master Geek
+1 received by user: 3

Trusted

  Reply # 473536 24-May-2011 17:14 Send private message

1080p:
I would argue that volume billing does care what is carried, specifically, large volumes of data. This can be seen in the trend to zero rate certain websites over others. This does not look like an incentive not to care what data is carried, does it?
 

Volume billing does not care what kind of data is carried. It doesn't care if it's http, ftp, VoIP, video, pictures of kittens or plain text.

25 posts

Geek


  Reply # 473706 24-May-2011 22:17 Send private message



KAREN doesn't work like that. (It doesn't make that distinction, as below, it's your router)
KAREN only advertises routes (v4 and v6) to other research networks such as AARNET and Internet2 and the EU research networks. (They also advertise routes to most google sites)

AFAIK, you only pay for a port and physical connection to KAREN. You still need to buy a commodity internet connection. Your router (a proper Cisco/Juniper/etc) can figure out which to route packets through. (Google/stanford.edu/canterbury.ac.nz will go via KAREN, the rest via your normal ISP)

I think you misunderstood my wording. I understand exactly how KAREN works as I install it for customers. My point was, when customers are connected to KAREN, they live with this apparently evil phenomenon of not being able to track where their data comes from, and where if it's sourced from one place, it's free, but not if sourced from the other.

Re using Traceroute to determine where from, I thought the argument was that 'normal' (read 'users with minimal technical knowledge') users couldn't tell where their data was coming from?

2329 posts

Uber Geek
+1 received by user: 78


  Reply # 473722 24-May-2011 22:30 Send private message

greaneyr:

I think you misunderstood my wording. I understand exactly how KAREN works as I install it for customers. My point was, when customers are connected to KAREN, they live with this apparently evil phenomenon of not being able to track where their data comes from, and where if it's sourced from one place, it's free, but not if sourced from the other.

Re using Traceroute to determine where from, I thought the argument was that 'normal' (read 'users with minimal technical knowledge') users couldn't tell where their data was coming from?


Oh, I see. Well it's up to customers to understand that. *shrug* There are ways to track where it goes. (Including which interface it goes out/in/etc)

I'm sure there's some analogies you could use for other (in)tangible things, but I can't think of any right now.


5184 posts

Uber Geek
+1 received by user: 747


  Reply # 473749 24-May-2011 23:20 Send private message

greaneyr:

KAREN doesn't work like that. (It doesn't make that distinction, as below, it's your router)
KAREN only advertises routes (v4 and v6) to other research networks such as AARNET and Internet2 and the EU research networks. (They also advertise routes to most google sites)

AFAIK, you only pay for a port and physical connection to KAREN. You still need to buy a commodity internet connection. Your router (a proper Cisco/Juniper/etc) can figure out which to route packets through. (Google/stanford.edu/canterbury.ac.nz will go via KAREN, the rest via your normal ISP)

I think you misunderstood my wording. I understand exactly how KAREN works as I install it for customers. My point was, when customers are connected to KAREN, they live with this apparently evil phenomenon of not being able to track where their data comes from, and where if it's sourced from one place, it's free, but not if sourced from the other.

Re using Traceroute to determine where from, I thought the argument was that 'normal' (read 'users with minimal technical knowledge') users couldn't tell where their data was coming from?


are KAREN users typical consumer broadband users?

1833 posts

Uber Geek
+1 received by user: 362

Trusted
Spark NZ

  Reply # 473757 25-May-2011 00:23 Send private message

1080p: [snip proposal for unlimited with constraints on number of users etc snipped]
Making this idea known to to your customers will make you very popular I would think. This is the kind of transparent business that lands you with a loyal customer base.


If you think ultra heavy Internet users in NZ have _ANY_ concept of loyalty, you're sadly mistaken. I've seen the real numbers, and even those not in the industry have been able to observe the massive users switching between the unlimited ISP de jure and killing those plans, one by one.

Several ISPs have tried unlimited. Some have even tried it more than once. It seems corporate memory for how badly unlimited plans have worked in NZ is about 24 months.

If we take the assertion of many people - ISPs are all money hungry evil companies that try desperately to separate users from their money while making as little investment as possible... then given that, if there WAS a profitable way to offer unlimited in the current regulatory environment, don't you think one of the ISPs would have done it?

Some of the economic and logical principles you put forward work in theory, and some work if you assume rational user behaviour. Others would work without current regulatory constraints. But the fact is that so far, not one single ISP in NZ has managed to make an unlimited consumer plan work out for both the customers and themselves.

There's only two possible logical reasons for this I can think of.

1) ISPs know how to offer a profitable unlimited plan, but are choosing not to do so to deliberately not make more money.

2) They can't figure out how to make a profitable unlimited plan.

Personally, I'm picking option 2. Call it negativity or realism, I don't care. But if you disagree with option 2, please be prepared to defend why option 1 is true. (or of course, offer an option 3 - but I'm afraid that I would reject the assertion of one internet user that they have the magic formula when all ISPs in NZ since Efficient Software in Dunedin in 199mumble haven't been able to figure it out)

Cheers - N



25 posts

Geek


  Reply # 473850 25-May-2011 11:39 Send private message



are KAREN users typical consumer broadband users?


At lot of them are schools, so they're even less technical than consumers!

7679 posts

Uber Geek
+1 received by user: 265

Trusted
Subscriber

  Reply # 473908 25-May-2011 13:38 Send private message

For a NZ ISP to offer un-metered they need to have:

1: Awesome traffic management equipment
2: Dedicated skilled engineers to manage/oversee the system
3: Extensive local caching
4: Lots and lots of domestic transit
5: Lots and lots of international transit

It's just not possible yet... for most NZ ISP's - to offer an acceptable level of service and a price consumers would find attractive.

I think Telecom have the sacle to do it, but they have failed twice.

Personally from the outside both times the planning and execution seemed poor eg: BigTime was $100 at release too cheap/attractive to leechers, http caching wasn't in place at launch, google/youtube caching wasn't in place at launch, traffic management rules were poor with some bone dead white listing people quickly exploited.

It could have been done a lot better.

I'm not even going to talk about Slingshot AYCE other than to say I think it fails on #5 and #2 at least possible some of the others too.

Also all the customers who currently have ADSL delivered via BUBA or older Telecom wholesale services over the old ATM backhaul and handover are still subject to the nasty handover link dimensioning which is a real handbrake on large caps for ISP's.

1332 posts

Uber Geek
+1 received by user: 152
Inactive user


  Reply # 473918 25-May-2011 13:45 Send private message

Talkiet:
1080p: [snip proposal for unlimited with constraints on number of users etc snipped]
Making this idea known to to your customers will make you very popular I would think. This is the kind of transparent business that lands you with a loyal customer base.


If you think ultra heavy Internet users in NZ have _ANY_ concept of loyalty, you're sadly mistaken. I've seen the real numbers, and even those not in the industry have been able to observe the massive users switching between the unlimited ISP de jure and killing those plans, one by one.

Several ISPs have tried unlimited. Some have even tried it more than once. It seems corporate memory for how badly unlimited plans have worked in NZ is about 24 months.

If we take the assertion of many people - ISPs are all money hungry evil companies that try desperately to separate users from their money while making as little investment as possible... then given that, if there WAS a profitable way to offer unlimited in the current regulatory environment, don't you think one of the ISPs would have done it?


You sorta missed my point. I realise that no one has yet been able to pull off a sustainable 'unlimited' plan. This is generally due to the lack of any kind of transparency about contention for bandwidth between ISPs and consumers as well as a lack of customer ratio maintenance. Being transparent will encourage loyalty; rather than oversubscribing and failing to communicate.

What I mean by ratio maintenance is that the ISP needs to actively ensure that its number of 'unlimited' customers do not saturate its base of regular use-roughly-15-20GB-per-month-mainly-during-peak customers - which is the vast majority of broadband subscribers. This way, when the ISP provisions bandwidth for its peak period they will have enough spare bandwidth during off peak to satisfy their unlimited users.

If that means setting a maximum number of slots for an 'unlimited' plan (based on a ratio to your customer base) then so be it. Being brave enough to front up and let consumers know that it isn't economic to sign up 3500 leeches when you only have 5000 regular customers will garner respect and those able to sign up will respect you as an honest provider even more.

When a provider signs up hundreds or thousands of customers to an 'unlimited' broadband plan knowing full well that they will not be able to come close to providing a reasonable service to all of its customers (on or off peak) is where you see the Slingshot AYCE and the Telecom Big Time etc... They completely disregard any kind of a customer ratio ensuring a reasonable experience for everyone. This is where you see reactions like the stupidity of putting all the high data users in a single bandwidth pool (remember the 125.* pool anyone?) or shaping all those on the unlimited plan resulting in poor speeds and angry customers no matter what.

I see this as a 'best efforts' service without the effort. This is where many consumers begin to get angry and see ISPs as greedy and because of this they show no loyalty.

A lot of the time ISPs don't price the service accordingly. Just think what could have been offered by Telecom if they charged $150 p/m for Big Time instead of ~$60 (iirc). They would have weeded out a lot of the stingy leeches (and those who would be quite happy with 40GB p/m but willing to jump for something that looked better) but pulled in so much more cash per subscriber making the service much more commercially viable.

I agree that the majority of users will not want to pay $150-160 p/m for internet access but if you are targeting the majority of internet users with an unlimited plan you aren't thinking logically.

Talkiet: Some of the economic and logical principles you put forward work in theory, and some work if you assume rational user behaviour. Others would work without current regulatory constraints. But the fact is that so far, not one single ISP in NZ has managed to make an unlimited consumer plan work out for both the customers and themselves.


I am not an economics major, all of this is simple opinion from a student of a very different discipline. From my perspective is seems glaringly logical but I realise that in many cases reality does not match logic. I feel like much of the regulation would not have been necessary had these logical principles been applied from the get go but making an argument in hindsight isn't fair.

Talkiet: There's only two possible logical reasons for this I can think of.

1) ISPs know how to offer a profitable unlimited plan, but are choosing not to do so to deliberately not make more money.

2) They can't figure out how to make a profitable unlimited plan.

Personally, I'm picking option 2. Call it negativity or realism, I don't care. But if you disagree with option 2, please be prepared to defend why option 1 is true. (or of course, offer an option 3 - but I'm afraid that I would reject the assertion of one internet user that they have the magic formula when all ISPs in NZ since Efficient Software in Dunedin in 199mumble haven't been able to figure it out)

Cheers - N


I agree that number one is not true. There is no one in the ISP business trying to give you a bad service in order to take more of your money, that I know of at least.

Number two is almost there; I would assume my thoughts aren't original but I feel like they are sound. I believe they could be converted into a business case (given enough real data input and not just guessing) so the problem is more in the execution than an abstract inability to find a solution.

Imagine that a single ISP offers an unlimited plan; I think Slingshot is the major one doing so at the moment. What will the majority of internet users who desire unlimited data logically do? Sign up with Slingshot (or stick where they are once hearing the terrible stories).

Now imagine all ISPs offered an 'unlimited' plan. Each plan was priced accordingly (regular broadband plans are cheapest, unlimited roughly 2x-2.5x the cost of regular plans), suddenly you have a market in which only those willing to pay for unlimited data access will go for it and a large number of providers able to absorb that (relatively small, once dispersed among ten or so providers) section of the market.

Now if all these ISPs were making sure to maintain their customer ratios they would be getting 2x-2.5x the money from their large data users for effectively allowing them to use the pipes when they are idle anyway. If necessary, shaping or priorotisation will be applied during peak times to ensure everyone can make use of the internet during the ISPs most stressed period. Perhaps a clause that unlimited users have a soft cap of 80GB p/m during the hours of 4-11pm would ease up on that peak period. This is all detail that would vary per ISP. 

1833 posts

Uber Geek
+1 received by user: 362

Trusted
Spark NZ

  Reply # 473929 25-May-2011 14:01 Send private message

1080p:  [snipped everything]


Sounds like we're actually very close to agreement. Your points are all valid in isolation, and are good ideas but do assume an awful of of behaviours that simply don't happen.

Heavy users aren't loyal. They won't moderate usage. Technical solutions will be circumvented.

ISPs won't all co-operate to launch similar unlimited plans to make sure it's shared about - and unless all are identically priced and have identical performance, users won't balance across them either.

There are also regulatory issues at the moment that should disappear with UFB - so that will help a bit.

The basic problem remains though, the cost structure for ISPs is not reflected in their retail pricing structure. This means they need to make assumptions about their userbase and traffic usage.

Economics does funny things to peoples' behaviour when price hits zero. People (and entire populations) that exhibit rational behaviour at $5/unit, or $3/unit, or even $0.10 /unit suddenly go batsh1t insane at $0.00/unit.

$0.00/unit is what unlimited plans effectively offer for the very heavy users, for all traffic above what they could get elsewhere in the included monthly cost.

They are not currently profitable customers, ISPs could absorb a few of them, IF they can move their traffic use out of peak times, and if they can somehow (through magic, because technology doesn't work) ensure that heavy users still get decent interactive performance during peak times. Unfortunately free is a crazy value, and it destroys performance.

I do know how I would do it, but it only works in theory, and I hold no illusion that it would be palatable to the marketing departments of any ISP.

I'd price data so that price per unit changes with ISP free resources. IF (at ANY time of day) there is 0% packet loss across the whole customer base, lower the price per unit. When packet loss is observed anywhere, raise the price for everyone, in REALTIME. This, over a relatively short period of time would reward those that use traffic when there is capacity available, and punish those that try to use capacity when it's congested.

However, this is a complex billing system, and there's no sensible way to communicate to customers the price in realtime - I don't care if you understand it, or if 10% of customers understand it - it would drive so much confusion that it just wouldn't work.

What I'm illustrating is that theories are nice, but special values (like free, or unlimited) screw things up, and theoretical models that work in theory sometimes simply don't fit the real world.

Cheers - N


wjw

157 posts

Master Geek


  Reply # 473956 25-May-2011 14:46 Send private message

Talkiet:
1080p:  [snipped everything]


I'd price data so that price per unit changes with ISP free resources. IF (at ANY time of day) there is 0% packet loss across the whole customer base, lower the price per unit. When packet loss is observed anywhere, raise the price for everyone, in REALTIME. This, over a relatively short period of time would reward those that use traffic when there is capacity available, and punish those that try to use capacity when it's congested.



Electricity spot market perhaps? 

1833 posts

Uber Geek
+1 received by user: 362

Trusted
Spark NZ

  Reply # 473966 25-May-2011 14:57 Send private message

wjw:
Talkiet:
1080p:  [snipped everything]


I'd price data so that price per unit changes with ISP free resources. IF (at ANY time of day) there is 0% packet loss across the whole customer base, lower the price per unit. When packet loss is observed anywhere, raise the price for everyone, in REALTIME. This, over a relatively short period of time would reward those that use traffic when there is capacity available, and punish those that try to use capacity when it's congested.



Electricity spot market perhaps? 


Great analogy... Very very similar. The difference we have in NZ at the moment is that the electricity spot market is mostly (completely?) at the provider level... Retail customers are insulated from the often wildly varying prices.

For a small number of very well educated customers that understand and REALLY accept that at times they might pay 100x what they pay at another time, it works great - for a retail consumer population... Well, let's just say there isn't a marketing person on earth that could explain to a broadband user why 1GB at 3am costs $0.30 but at 9pm on the busiest night of the year with a royal wedding on costs $3000.

Cheers - N :-)



333 posts

Ultimate Geek
+1 received by user: 88


  Reply # 474160 25-May-2011 21:10 Send private message

WOW, what a response thank you to everyone who has contributed, alot of good thoughts and points have been brought up. Will go through all the poat later on when I have time. This has promted a very good discussion.




1615 posts

Uber Geek
+1 received by user: 66

Trusted

  Reply # 474206 26-May-2011 00:37 Send private message

CoolAs101: WOW, what a response thank you to everyone who has contributed, alot of good thoughts and points have been brought up. Will go through all the poat later on when I have time. This has promted a very good discussion.


A very poor discussion i would have thought. Its the same arguments as always that can never be resolved because we haven't got extra international bandwidth to go around, with the exception of the change from BUBA to EUBA that does affect some opportunities for local traffic that weren't possible before. The fact remains that the only way to ensure an unlimited plan is successful is to guarantee that some other equivalent plan stays faster and cheaper, thus attracting the worst of the leechers. I've been involved with these unlimited plans too and the best way is to seriously throttle heavy users at peak times to fairly share traffic among all users on the plan.




Qualified in business, certified in fibre, stuck in copper, have to keep going  ^_^



333 posts

Ultimate Geek
+1 received by user: 88


  Reply # 474527 26-May-2011 18:43 Send private message

webwat:
CoolAs101: WOW, what a response thank you to everyone who has contributed, alot of good thoughts and points have been brought up. Will go through all the poat later on when I have time. This has promted a very good discussion.


A very poor discussion i would have thought. Its the same arguments as always that can never be resolved because we haven't got extra international bandwidth to go around, with the exception of the change from BUBA to EUBA that does affect some opportunities for local traffic that weren't possible before. The fact remains that the only way to ensure an unlimited plan is successful is to guarantee that some other equivalent plan stays faster and cheaper, thus attracting the worst of the leechers. I've been involved with these unlimited plans too and the best way is to seriously throttle heavy users at peak times to fairly share traffic among all users on the plan.


I thought it was quite a good discusion actually as for someone like me who is still learning about fibre broadband and new zealand's internet infrastructure it tought me alot. 




1 | 2 | 3 | 4
View this topic in a long page with up to 500 replies per page Create new topic




Twitter »
Follow us to receive Twitter updates when new discussions are posted in our forums:



Follow us to receive Twitter updates when news items and blogs are posted in our frontpage:



Follow us to receive Twitter updates when tech item prices are listed in our price comparison site:




News »

Trending now »
Hot discussions in our forums right now:

Orcon Global Mode launched
Created by freitasm, last reply by gehenna on 30-Aug-2014 23:11 (121 replies)
Pages... 7 8 9


Judith Collins: I am resigning
Created by Presso, last reply by KiwiNZ on 31-Aug-2014 07:53 (45 replies)
Pages... 2 3


Suddenly limited to 1mbps download speed on spark VDSL
Created by Jase2985, last reply by hio77 on 30-Aug-2014 17:39 (45 replies)
Pages... 2 3


Vodafone TV multicast settings on pfSense?
Created by kenkeniff, last reply by kenkeniff on 27-Aug-2014 10:32 (182 replies)
Pages... 11 12 13


Sluggish Macbook
Created by SATTV, last reply by SATTV on 30-Aug-2014 22:21 (19 replies)
Pages... 2


2Degrees Fraud Prevention "Text TALK to 233 now to keep talking"
Created by ArcticSilver, last reply by sonyxperiageek on 30-Aug-2014 22:03 (19 replies)
Pages... 2


Lightbox press event release
Created by freitasm, last reply by IcI on 30-Aug-2014 17:54 (562 replies)
Pages... 36 37 38


Recommendations for soundbar or similar for <$1k?
Created by jonathan18, last reply by michaelmurfy on 29-Aug-2014 21:11 (17 replies)
Pages... 2



Geekzone Live »
Try automatic live updates from Geekzone directly in your browser, without refreshing the page, with Geekzone Live now.

Are you subscribed to our RSS feed? You can download the latest headlines and summaries from our stories directly to your computer or smartphone by using a feed reader.

Alternatively, you can receive a daily email with Geekzone updates.