The Hole ISPs Dig When They Choose Throttling Over Capacity

If you complain so much about the service jump ship!

Thats all well and good, however Rogers and Shaw have a Monopoly of their coverage areas. Lots of places which aren't able to be served by ADSL technology or extremely poor ADSL coverage.

Whats funny is all of the smaller ISPs do no such throttling. You would think these huge corporations would have pleanty of money for upgrades and bandwidth.

I own smaller ISPs. I can assure you, many of us do this.

Users at some point have to realize that prices are low and margins are lower. Many of us don't want to play the transfer cap game, so throttling high demand users is a solution.

When business has bent over backwards for the last 12 years to drop the price of consumer bandwidth it has cost us in the realm of profit margins. Many broadband ISPs are taking a $200 to $400 loss on the front end hoping to see profit in 12 to 24 months.

DS-3's are not $19.95 per month. Sorry. Low prices are due to the high over-subscription rates we have been able to maintain as users went from occasional web surfing on dial-up to occasional video watching on broadband. Now the trends have changed to many people running dedicated servers (bittorrent or otherwise). This low price Internet ideal that people are holding on to is almost impossible to maintain when I go from a 20:1 subscription rate down to 5:1. Essentially the overhead in this case has increased nearly four times. Those aren't hard numbers, but you get the idea.

You are not paying for a dedicated and unlimited connection. If you are being told otherwise, then they should change their marketing.

We have a policy of letting people know exactly what is and is not allowed on our network. If they don't like that, they can go deal with someone who will lie to them. It won't be me.

I should have said ALOT of the smaller ISPs HERE do not throttle. There is a big difference between capping certain users or having upfront caps to providing a service and blanket throttling protocols and services simply because your infrastructure or pricing model doesn't cover it. Don't be shady and say up front you have an alloted amount per month etc. There is no REAL reason to throttle for a company like rogers except that their infrastructure can't handle it. They ALREADY have KNOWN monthly caps so using the excuse that bandwidth is expensive isn't really a valid one when everyone is capped the same. They simply don't want to pay for needed upgrades on their crowded network. There are alot of differences between having an unlimited connection and being able to use your connection to its ability anytime you want. Why should I pay for 6mbit cable if I can only ever use half or so of that at any given time because their network is so poor. This has nothing to do with a need for unlimited/dedicated bandwidth. I should be able to get ~95% of my throughput whenever I need it for a download or service.

There's also a difference between throttling some to manage capacity, and throttling in order to avoid spending money on necessary upgrades in the face of video demand.

Deffinately.

You guys have missed my point and here in lies the problem with the modern consumer mentality and businesses desire to market to it. Bit Torrent is a dedicated server, kazaa, limewire, bearshare, these are server apps. The current system they have was never meant for this.

They should do the smart thing and create a new system to accomodate these new consumer desires and charge accordingly.

Some people live to complain... even after leaving!

The only real alternatives for high speed service, in a given area in Canada, are DSL through Bell - provided you live in a supported area and a single Cable provider...that's it.

Thats really a bold statement considering there is no real choice in most places.

There is -always- a choice. You can choose to have what you've got, or nothing at all. Nothing is always an option. But one most people won't take and the companies of the world know this.

Dude, you have reviews for two different ISP's and two different VOIP providers.

You don't even have the room to say "nothing at all" is a choice. There are people out there that have uses/obligations that can't be satisfied with "nothing at all" as an option. If this ISP is all that they have to work with, and they use their market stance to offer shoddy service, then people are at a disadvantage and don't really have any other option, do they?

Sure I do. Nothing at all is a choice for me as well. However, I'm realtivly content with my service providers, so I continue to pay for my choices. If I weren't and I felt that dialup or changing my lifestyle as to not rely or have use for the internet meant doing away with it, then that is a choice (one I'm willing to make, if it came to it).

People, anymore, refuse to go to the extreme of voting with their wallets by not having any service at all. And, as I said before, the companies know this. So that gives them all the room in the world they need to screw you in the current climate in the US.

Some of us don't have the luxury of being able to say "no."

I, for one, use my connection for work. If I disconnect, I don't work. I don't work, I don't make money. I don't make money, I have no place to live. I think you get the idea.

If an internet connection is nothing but a luxury, I would agree... having none is an option. However, for those of us who work and make money with our connection, we don't have a choice if there's only one company.

YEP.

And you are exactly the type of person I was thinking of.

Not intending to be snitty, but if you are dependent upon broadband for your business, those costs better be in your business model or you are in trouble. If you are depending upon consumer grade connections for something that puts bread on the table, perhaps you should be a fan of any internet access provider.

As a business person, you do realize there is a plethora of choices for connectivity that existed long before the consumer "broadband internet experience", right?

In either case, it still appears that you can work without a connection to the internet. So many possibilities of doing something and transmitting the data from point a to point b without the internet come to mind.

In principle I agree with you.

Depends how you "work" from home. Company I worked for kicked all the workers out of the office to save on office leasing costs.

One day, they said "now you're teleworking. No we don't pay for your internet access at home. Since you're saving on your gas mileage, fork up the internet connection cost. No we don't care that working on a customer grade connection is violating a TOS."

And I know of at least 1 other company that did the same thing. And I'm not talking 10 user employees. Our office had over 500 people, and we had over 100000 employees worldwide.

In my case, I had the choice to work on a basic $40 cable, or cough up an additional $150 a month to get a business grade cable or DSL connection. Hard to swallow if you ask me.

So they jump through all these labyrinthine hoops and spend all kinds of $$$$ to thwart users from using what the cable co. considers excessive bandwidth, but by the same token, they refuse to invest in infrastructure to support the existing consumer base? Sounds like the Canucks have been taking notes from the American telcos/cable cos. If they're not careful, they may just end up in the broadband penetration toilet where the U.S. currently resides. Meanwhile the poorest 'hood in 3rd world cesspools has FIOS.

...you build them for the average user and you throttle the hogs or boot them off. No matter how fast you make your network or how much bandwidth you provide, the abusers will always find a way to eat up all available bandwidth. Unless, of course, you move to the pay-by-byte model. Then everyone, except the non-rich hogs, will be happy with speeds.

Yes, you build to support all your users. My provider, and most others do not seem to have a problem with this. Mis-management?

TCH is just pissed at the pummeling his ILEC stocks took when people stopped getting gouged by insane per-minute long distance rates. He wants it to return, only in the form of per-byte charges.

Actually he is partially right here. They should expand their capacity, but they shouldn't do it for those who routinely download gigabytes of material a month and strain the system. I'm not talking about 50 or even 100gigs. I'm talking about those users who download 300-400gigs a month every month.

I agree. However, usage of the percentage model is flawed. What if average usage was 500MB a month and you happened to download 3 GB a month because you receive and send lots of videos to your family? You are now the unprofitable customer - or the low percentile bandwidth hog - is it time for you to go?

While I definitely don't believe in the pay-per-byte method, I like to tease TCH. He is completely entitled to his opinions, as am I, but I am also entitled to tease him about them.

Educate yourself.

»www.ipsphereforum.org/index.html

Far from limited to ILECs...pay per use is coming as soon as the logistics are figured out. (S)/he is correct.

Pretty tough to build a business where customers pay less than the costs to serve them (even on average). Incorrect signals produce long term problematic results (fiscally as well as engineering wise, which are fiscally as well).

I think you are on to something...and it probably addresses the "network neutrality" issue as well. All ISPs should build a model that satisfies their direct customers at a price point that their customers are satisfied. Obviously there are customers that are willing to pay for more bandwidth, such as yourself, and should do so if your demands exceed the "average person". Business rates in telecom have always been driven under similar ideals...the network is engineered for peak usage at x connectivity/quality, which was/is in turn driven by the business user. The internet pricing model, as far as consumer dsl/cable is still far from that model from a pricing perspective, but from an engineering perspective,
the network is designed to meet the needs of the people paying for premium connectivity (dedicated lines / higher priority), not the average consumer.

Even the smallest of ISPs have followed this model since dial up days in my recollection.

not mis-management, it's probably cheaper and easier to throttle than it is to increase capacity.

Why build capacity to serve new customers when you can throttle existing customers and squeeze the new ones into the same bandwidth? I'll answer my own question: NO COMPETITION.

Well, if the pay per byte model is so good, why don't any of the major ISP's do it? I remember back in the 80's, I used to pay about 5.00 an hour for compuserve access, but I could use as much bandwidth as I wanted.

The reason that pay per byte will never happen is twofold.
First: There is no 'minute' cost for using your connection. Unlike a phone call (switched network), where a phone call ties up a circuit for x amount of time, a packet based network doesn't use any 'resources' if it's not being used. So the concept of charging by 'time' is meaningless.

Now, charging per byte. Granted, the pipes are a finite resource. The limiting factor is how much data the equipment can handle, AT ANY INSTANT. Unlike oil, or other 'physical' resources, the limit on an internet connection is not 'how much' (i.e. if you're not using 100% of your bandwidth, it's lost) but how much per user. If I travel 100 miles, my car uses about 3 gallons of gas. So, I'm out about $7.50. However, if I download 100GB, my 'pipe' didn't consume any 'resources', but I DID use a percentage of the total pipe.

To put it in numbers, if a particular node can handle 100mb/sec, in a perfect world, that node would move about 30TB of data every month. In a typical 500 user node, that means every person could download about 60GB per month. And, if every user DID download 60GB/month, the incremental cost to the ISP would be, umm, ZERO.

We don't live in a perfect world, so some users do 500GB/month, and others do 1GB/month. If the ISP's were to charge 'by the byte', then 500 people at 65.00/month, would mean that the ISP would charge you $1.00 per GB/month. Now, I have NO PROBLEM paying $5.00 for a 5GB/month plan, but the ISP would never do that.

Of course there is no minute cost if its pay per byte or pay per bit, that would make it a "byte" or "bit" cost.

As much bandwidth as you wanted in the 80's was more than likely over dial up and not even close to as much bandwidth one may want over broadband.

Because the marketing departments fear the customer attrition that would result if the rest of the market doesn't follow suit in charging for actual data transferred.That's why you don't bill by time Huh???? If you download 100 GB, you are most definitely consuming resources.Actually, it's 62 TB/mth...don't forget duplex That 100 Mbps connection needs to be paid for somehow. And in reality, to pay for that connection, the ISP most likely requires a lot more than 500 paying customers. Hypothetically, let's use 5000 customers to cover the ISP's expenses, so now you're talking about 6 GB/mth (12 GB full duplex) of transfer. Now, a large majority of customers won't use anywhere near 6 GB/mth, so the ISP is able to over-subscribe their network to lower the costs to the customers across the board. But what happens if only 10% of the customers transfer 100 GB/mth? That right, the network reaches a choke point. So now the question becomes, how do you pay for network upgrades without charging an exorbitant amount of money for your high margin customers (simple web surfing and e-mail)? Simple, charge the customers that drove the need for network upgrades more money.You're right, an ISP can't afford to do that. But ISPs could charge for a connection (e.g. $19.95/mth) which includes a specified amount of data transfer per month (e.g. 5 GB), and then bill per MB or GB (e.g. $2.95/GB) above and beyond the initial service offering. Very similar to cell phone and long distances plans

You're right, an ISP can't afford to do that. But ISPs could charge for a connection (e.g. $19.95/mth) which includes a specified amount of data transfer per month (e.g. 5 GB), and then bill per MB or GB (e.g. $2.95/GB) above and beyond the initial service offering. Very similar to cell phone and long distances plans

Because if you are going to bill bytes per month, then bill bytes per month. The ISP's costs aren't reduced during minimal times of use why should yours be? If you want that model, then expect your "connection" charge to be higher than my guesstimate of $19.95 and closer to the $40 or $50 that a lot of people pay now.

First: Time is not an issue with a per bi/yte charging. You first point is meaningless.

Second: The model would look more like the current cellular model with a minimum plan for up to XXXX bytes then then YY for each additional bi/yte. Along with all of the other pricing variations that are out there and probably some new ones.

The reason it probably wont happen is the operational logistics of counting the total "used" bytes because of issues such as overhead, computer generated traffic, viruses, spam, etc. It would just be a complete mess at the operational and customer support level.

Imagine calling in and arguing about how many bites and bytes you actually consume with tech support....ah!

It won't make the provider happy, because the majority of customers don't use many bytes. It's better for them to get $40 from 1000 people, than $10 from 900 and $100 from 100.

I am more than willing to pay by the gig (not the bit or byte) but they better include a fair amount in my normal flat fee, and they also better make my pipe one hell of a lot faster. If I am going to pay by the gig, I don't expect to have to wait a month to get what I want.

The funny thing is, in most data centers, they don't even count your downstream amounts, only outbound. For $50 a month, I bet I could get a web hosting account that includes a ton more of bandwidth than I get now.

Boy do I wish my isp had as many competitors, as web hosts do. Just think of the broadband utopia we'd live in.

Bit torrent is an application that simply does not work well with today's internet infrastructure. The internet, is still ultimately a client-server environment, and the cost of trying to properly decentralize it would be huge.
I personally really like Comcast's burst speed idea, it provides quick bursts of upstream data when you need it, and keeps bit-torrent from destroying the network. Personally, I wouldn't mind if my ISP (Shaw cable) gave me 2mbit burst upload speed (up to 45 seconds), and reduced my sustained upload to 128k (down from 1mbit).

Remember, bit torrent is still largely pirated material, so pointing out the legitimate uses for the program is a moot point.

If you lease a dedicated server box from a data center, you're usually given so many gigs of bandwidth per month, above which point, you pay overage fees.
Your server's network connectivity speed is a symmetrical amount, usually 10 or 100mbits/sec, and if you stay within your allotted bandwidth limit, which is usually in a good facility plenty for even the highest end average usage such as streaming audio/video, that kind of plan seems to work, and you get the full capability of the network.
Why can't or won't cable/DSL provider offer similar tiers of service, where users could enjoy the full upstream/downstream possibilities of the network they're on? I'd think doing something like that would be a neat thing to try.