Must Reads

Nowak on 10 Myths from UBB Supporters

Peter Nowak has a terrific post that responds to ten frequently heard myths about usage based billing.

10 Comments


  1. My question is: since all major ISPs already have caps in place, why do they also throttle users that are well within their monthly cap?

    One may expect that this would only apply to “heavy users” that are well beyond their cap? Or no caps accounts?

    If I have Bell DSL with 30 GB/month and I respect this limit, why do these 30 GB have to be throttled too?

    Nap.

  2. I do have a bit of a comment with respect to reason #9… mostly his math and some of the assumptions. A dialup connection, running flat out at 56 kbps for a month, will use up about 14 GBytes according to my calculations. So, for about twice the price you are getting about 4 times the data.

    In addition, if we are talking about spending on an internet connection, you may need to add a portion of the monthly phone bill, since that is how the dialup is carried. The $170 quote from statscan implies that is what the households paid to the ISPs… without a phone line dialup is not much use.

    With respect to item 10, he is partly correct. However, the bandwidth of the links is finite; barring compression you can’t stuff bits down the link faster than the link supports. Thus if I am using 1Mbps then that is not available for others to use.

    @Napalm: I agree that throttling should only kick in once you approach your cap… however this needs to be considered in combination with QoS requirements. For instance, a VOIP connection needs a consistent throughput (let’s call this a real time link). Where too many non-real-time links are running the assure the real time links requirements, then one of two things needs to occur.

    1) The available bandwidth is split up; the real-time link requirements is allocated first and what is left over goes to the non-real-time links; or
    2) The non-real-time protocols are throttled.

    Where the link does not support QoS requirements then option 2 is all that is left. 1) is an option where QoS is available. However, this does not handle apps which advertise themselves as real-time but aren’t.

  3. Anon-K said: “With respect to item 10, he is partly correct. However, the bandwidth of the links is finite; barring compression you can’t stuff bits down the link faster than the link supports. Thus if I am using 1Mbps then that is not available for others to use. ”

    You obviously didn’t read the entire item. I quote from the article: “The “pipes” and other equipment over which these electrons flow are, of course, finite and therefore need to be continually expanded as the amount of traffic grows.”

    Now whether he’s referring to bandwidth (ie: available speed) by that statement, which I believe he is, or the actual physical equipment being subject to “wear and tear”, could be debated. However, the point he’s making isn’t about bandwidth (ie: speed). In fact in point 10 itself he never uses the word “bandwidth”, but rather “data”. His point is about the concept that the “stuff” being transferred over the lines is somehow a limited resource which you should pay for the bits of. That’s not the case. He’s saying there isn’t a big magic box of “gigs” at the ISP of which they may have a limited supply, as there is with hydro or gas.

  4. @Anonymoose
    Actually. I did read his entire post. His comment about the pipes needing to grow is beside the point… exactly the same applies to normal “utilities”. As the population grows and draw becomes greater, the pipes delivering these utilities also need to be expanded to handle the load.

    I’ll give you a for instance. There is a lot of talk about “intensifying” the urban areas of Ottawa. There is a few reasons, but part of it is so that they don’t have to expand the serviced land (water, septic sewer, natural gas, electricity, etc). Which is all well and fine, except that the services are built to handle a particular population density. Once you exceed that density then you need to expand the utilities or risk not being able to supply water, natural gas or electricity to the population, nor handle the septic load at peak times of the day. Many of the people that are advocating for intensification are ignoring this potential consequence.

    In terms of data, he is correct in that the consumer is not consuming data. However he has not allowed for the consumption of the capacity of the link. I admit that there are at least two schools of thought on this.

    1) The “dedicated connection” model, where the consumer has the right to use the link at the advertised speed constantly; for instance on the 6 Mbps link they continuously push 6Mbps of data. This is closer to what was described in an earlier entry as the connection negotiated by the ISPs with Bell.

    2) The shared link model. Picture this as an ethernet model. The link provides 10, 100 or 1000 Mbps of capacity however an individual computer shares this with the other computers in the collision domain. This means that, if you have 15 computers on the network you probably can get away with a single 1000 Mbps network. In this model the individual computer doesn’t consume data on the network; they consume network capacity.

    The utility model is more like the shared link model. The main difference is that the consumable price charged by the utility contains not only the price for the consumable but (normally) also the cost of delivery of the consumable.

    The Ontario Hydro pricing model is a good example of this. A home pays a fixed price for the connection to the grid. You pay a price (legislated) per kW/hr of electricity consumed, and a price per kW/hr for delivery charges. To put the shared link model into the data world, since there is no consumed data, the first variable cost to the consumer is 0.

  5. @Anon-K
    So what you’re really saying is you actually agree with the author’s point. If the “utility model” is to be billed for the access and delivery in addition to the cost of an actual consumable, then comparing the internet to a utility is ultimately false, since there’s no consumable.

    I’ll give you this, you’re right that the need to expand pipes is completely besides the point, that’s just the cost of doing business. And that’s exactly what he’s saying. Those things are “finite” and that’s the only way the internet is comparable to a utility, in that they have maintenance and expansion costs.

    That alone doesn’t mean you should be billed for your internet usage in the same way you are for the utilities BECAUSE there’s no consumable being transfer away from them to you. “Data” is not something they have that they no longer have once you have it. THAT is what he’s saying, and that is what would be required to realistically equate internet usage to that of a utility.

    The only way I would accept comparing a utility to internet usage would be if electricity, gas and water were actually INfinite, which they are not. They only have value because A. they’re tangible, B. they are finite resources, C. there’s a cost to creating and storing them and D. once they’re used by the consumer, they’re lost forever until more is created/cultivated. That’s what make them consumables.

    Different IS different.


  6. How about discussing some other utilities like water? Where they bill you for “sewage” for all the water you consume, even when, during summer, more than half of it goes to watering lawn, flowers and trees not into the sewer system?

    What gives?

    Nap.

  7. @Anonymoose
    Perhaps I didn’t make myself clear (see how I am not putting words into your mouth ;-). My point is that comparing an ISP to a utility has some merit. There are three costs associated with a utility:

    1) Fixed connection costs to the delivery mechanism. This covers the cost of maintaining the connection for the individual consumer. In the case of an ISP, this is the physical access at the customer premises plus always on services, such as email accounts.
    2) Variable costs related to the cost to the utility of what is delivered.
    3) Variable costs related to the cost of delivery, which includes the cost of maintenance of the required amount of capacity. For an ISP, this is the maintenance of the backbone.

    For data (2), the price to the ISP is $0 (agreed). However, an ISP still needs to recover costs for 1) and 3).

    @Napalm: The comparison to a water utility is the same. With respect to your point about sewage, I think you are partly correct, however please consider the following points:

    1) Metering sewage would require another meter… since we are talking about sewage this would likely be fairly expensive to deal with the solid stuff that goes down the pipes such as #2, used condoms, etc.
    2) At least some of the stuff that does not go down the septic sewer system will go into the storm sewer system. This includes water used to over-water lawns, trees, etc (which may contain fertilizer) as well as water used to wash vehicles (which may contain detergents, oils, road salt, anti-freeze, etc). This means that, even where the septic and storm sewer systems are separate, the storm system will need some treatment before discharge.

    Perhaps, you incorporated the storm sewer into your argument; if you did my apologies, that is not what I took from it. However, if it did, then your point becomes more meaningful with respect to ISPs, as the people that are putting more sewage down the sewers are paying less for the capacity that they are using on a per-unit basis.

  8. Anonymoose says:

    @Anon-K
    Sorry, I only mean to point out that your argument, at the very least partially, actually proves the author’s point in that item, whether you want to admit it or not.

    The point is that it’s disingenuous to say that “A and B are the same and therefore B should be billed like A”, when in fact, and as you pointed out, “A” and “B” in this case are not “the same”. That’s the argument I believe he’s making.

    Nobody is denying that ISPs have costs no doubt exactly as you’ve outlined in 1 and 3. I also don’t doubt that they have costs which you’re unaware of or have omitted. But simply because some of those costs are similar to a utility is not a valid reason, in the author’s opinion, to use “because they’re the same they should be billed the same” as an argument.

  9. Anonymoose says:

    @Anon-K
    Let me add this addressing specifically the argument in the context of UBB. We agree it costs ISPs $0 for the data, and I believe we also agree that the data on the ISP side of things is what they equate to the consumable on the utility side. So in the context of UBB, ie: billing for “the bits and bytes” so to speak, that passed over the lines, which we agree is basically a $0 cost, what they’re really saying when they say that “ISP is like Utility” is that those bits and bytes are like the consumable, and somehow cost something, which we agree they don’t in the case of the ISP. Therefore the argument of “ISP = Utility so lets bill for bits and bytes” falls apart.

    If they’re going to use that argument, IMO, THEY, the ones using this comparison as the argument, must specifically outline in PRECISELY which ways they’re alike, and why those specific ways mean one should be paying per byte of data received. I’m not aware of any such explanation, are you?

    Otherwise it’s like saying “well this green apple is about the same size and shape as this tennis ball, so we should sell it for the same price as the tennis ball”. I’d be like, but um… the apples grow on trees right? there’s a difference isn’t there?

  10. Anonymoose says:

    … And then you’d be all:
    1) There’s delivery costs to ship apples to grocery stores, and same for tennis balls to sporting stores.
    2) Granted the apples do grow on trees, so to a certain extent (NOT literally of course), they’re “infinite” ie: as long as there’s apple trees there should be apples
    3) It takes equipment to get those apples, and to create those balls. So there’s maintenance, repairs, and upgrades in both cases.

    So there’s an argument for apples and tennis balls to cost the same because of 1 and 3?

    Yes I’m fully aware that this argument is full of holes and makes absolutely no sense…. oh wait.