News

What to do About Retail Usage Based Billing: A Modest Proposal

OpenMedia.ca, which spearheaded the public uproar over usage based billing earlier this year, launched a Vote Internet campaign that quickly attracted political support. The campaign asks candidates to be pro-Internet, which includes standing up for an open and accessible Internet and stopping the “pay meter on the Internet.” While this predictably raises claims of retail price regulation, addressing concerns about retail UBB need not involve a return to regulatory approvals over retail pricing of Internet services.

I’ve argued that UBB is fundamentally a competition problem and that addressing the competition concerns (which OpenMedia also supports) will address many of the concerns.  Increased competition takes time, however, and in the meantime there are legitimate concerns about the use of UBB in Canada at the retail level given the approaches in other countries and the pricing far above costs.  In addition to discussing those issues, my UBB paper makes a modest proposal for addressing retail UBB that includes greater transparency and a reasonableness standard.  The proposal – which I’ve called the creation of Internet Billing Usage Management Practices or IBUMPs – is explained below.

Given the CRTC’s longstanding forbearance of retail Internet services, addressing retail UBB represents a more challenging regulatory and policy issue. While some have advocated for the elimination of UBB at the retail level, there is little political support for directly intervening in the pricing choices of retail ISP services. The optimal method for addressing retail UBB is a robust competitive environment in which it is consumer ability to switch providers in the event they are dissatisfied with UBB that serves as the primary safeguard against unreasonable UBB practices.

In many respects, the current debate resembles the early battle over net neutrality in Canada.  Indeed, substitute the words video streaming, Netflix, and heavy users for peer-to-peer, BitTorrent, and file sharers and the discussion points are virtually identical.
Net neutrality advocates maintained that fears about traffic shaping practices and preferential treatment of incumbent ISP content were the product of an uncompetitive marketplace that required regulatory intervention. Incumbent ISPs argued that there was no need for CRTC involvement since a market-oriented approach was sufficient to address any public concerns.

The CRTC ultimately adopted guidelines that ensured that Internet traffic management is not a free-for-all. It rejected arguments that the market would ensure ISPs provide adequate disclosure on how they manage their networks.  Instead, it mandated full disclosure of traffic management practices, including information on when they occur, which applications are affected, and their impact on Internet speeds. Moreover, it adopted a new test to determine reasonable traffic management practices. Where a consumer complains, ISPs are now required to describe their practices, demonstrate their necessity, and establish that they discriminate as little as possible.

a.    IBUMPs – Transparency and Disclosure

Given the current competition concerns and the fears that the UBB could be used for uncompetitive purposes, the CRTC should consider new guidelines for retail UBB – Internet Billing Usage Management Practices (IBUMPs) – expanding on the CRTC’s Internet Traffic Management Practice guidelines.

Much like the ITMPs, the IBUMPs would include a comprehensive transparency requirement so that consumers can better understand the limits of their service and reliably track their monthly Internet usage. While ISPs already provide some tracking tools, they have proven unreliable with admissions that there have been errors.  The CRTC’s ITMP decision appears to adopt the position that monthly usage bills provide sufficient transparency to Canadian consumers. Yet the incumbent ISPs have also sowed some marketplace confusion with marketing that breeds consumer uncertainty about what their data plans actually provide in practice. 

Marketing materials often overstate the benefits of data plans by citing unusually compressed and shorter MP3 music or video files to suggest that consumers can download far more than is likely to be the case.  Similarly, discussions often conflate streaming video with downloaded video, leaving consumers with the mistaken impression that they can access hundreds of hours of video each month.  This claim may only be true for low-definition streaming, not for downloaded video from various commercial services.

If UBB is to remain part of the retail Internet access landscape, the transparency and public disclosures must improve. The CRTC should adopt similar requirements as those found with ITMPs to ensure that consumers are better informed about the benefits and limits of their capped services.

b.    IBUMPs – Reasonableness

Much like the ITMPs, greater transparency and disclosures are a necessary but not sufficient step to address current concerns. The ITMP process also led to guidelines designed to ensure that ISP traffic management practices are reasonable. The CRTC’s basic test for acceptable ITMPs was as follows:

When an ISP is responding to a complaint regarding an ITMP it has implemented, it will use the ITMP framework. In doing so, the ISP shall:
    
Describe the ITMP being employed, as well as the need for it and its purpose and effect, and identify whether or not the ITMP results in discrimination or preference.
    
In the case of an ITMP that results in any degree of discrimination or preference:

  • demonstrate that the ITMP is designed to address the need and achieve the purpose and effect in question, and nothing else;    
  • establish that the ITMP results in discrimination or preference as little as reasonably possible;
  • demonstrate that any harm to a secondary ISP, end-user, or any other person is as little as reasonably possible; and
  • explain why, in the case of a technical ITMP, network investment or economic approaches alone would not reasonably address the need and effectively achieve the same purpose as the ITMP.

The ITMP second-level analysis is limited to instances of discrimination or preference. Although UBB may raise discrimination or preference issues, particularly for competing video products, many UBB plans do not. Applied to the issue of retail UBB, the establishment of an IBUMP policy would would ensure that retail UBB is reasonable in light of network congestion concerns and marketplace conditions. It could include the following four conditions.

First, CRTC approval for UBB has long been focused on addressing network congestion.  Indeed, incumbent ISPs now argue that the use of UBB is related to congestion concerns and the need for fairness among all subscribers. In the case of a complaint about UBB retail practices, ISPs should be required to demonstrate that the UBB approach is designed to address specific network congestion concerns. 

Second, the UBB model should be the least restrictive possible to achieve its intended goals. Much like ITMPs, this may mean giving consumers the option for accounts that are rate-limited after a certain monthly cap is reached instead of imposing overage charges (as noted above, many ISPs around the world have adopted this approach).

Third, if the ISP has implemented traffic shaping or other technical measures, it should be required to demonstrate why those approaches alone would not reasonably address the same network congestion concerns. This requirement is a mirror image of the ITMP requirement that envisions ISPs relying on network investment or economic ITMPs (ie. UBB) rather than technical ITMPs such as traffic shaping.  If traffic shaping is in place, ISPs should be required to explain why that has not adequately addressed the network congestion concerns.

Fourth, while the CRTC should not engage in reviews of the retail pricing or size of monthly data caps, it should, under certain circumstances, be entitled to examine overage charges should ISPs use this economic model (as noted above, many ISPs around the world rate limit rather than impose overage charges). The review of overage charges would only occur in local markets where both the cable and DSL provider employ UBB, thus leaving consumers with limited non-UBB alternatives.  In such instances, the CRTC should review overage charges to allow for a reasonable profit but establish safeguards against price gouging in light of actual ISP costs.

The IBUMP approach would provide additional guidance the CRTC neglected to provide in the ITMP decision.  That decision recognized that guidance on reasonable traffic management practices could co-exist with a largely unregulated retail Internet services market. The same decision left economic ITMPs largely unregulated, effectively giving the Commission’s blessing to UBB.  Given recent marketplace developments, it is increasingly clear that was a mistake. Much like the ITMP decision, the Commission should not dictate pricing models, but it should ensure that pricing models are transparent and that those premised on addressing network congestion do so in the least restrictive manner possible.

24 Comments

  1. Issue is not idealogical, it’s technical.
    As the title states, UBB is a false metric to measure network strain created by a user.

    This of itself is why it should be outlawed entirely and why people are so upset by UBB conceptually.

    I’ve illustrated dozes of times how you don’t measure network strain by data transmitted over a billing period.

    The first component of the measure is meaningless and the second is completely arbitrary. Would you pay for your produce at the grocery store based on what scent of underarm deodorant you use?

    That’s exactly how absurd UBB is!

    Technical experts (not marketers and greedy stakeholders) as well as Measurement Canada need to be involved.

    Network strain is measured in capacity during a given moment. Penalizing people for total data transmitted doesn’t reflect a real cost.

  2. Alex is completely right. Mbps IS a data transfer cap.
    I wholeheartedly agree with Alex: UBB is nothing more than a confusion of concepts, and is utterly invalid regardless of its strategy or so-called reasonableness.

    The very term “Mbps” (Megabits per second) is a measurement over time. Therefore, when I pay for a 15 Mbps connection from Shaw, a transfer cap is built into this service. The real issue is that the big ISP’s have oversold their residential nodes. Similar to reserve banking, now that everyone is trying to use what they actually pay for, at certain times of the day, the nodes cannot deliver to each customer their entitled bandwidth – i.e. congestion occurs.

    You know what is ironic, though? This congestion is completely artificial. It was created due to the nodes being recklessly oversold to increase profit margins. It’s not the customers that are now “using more and robbing from others” No. It’s ISP’s overselling for a quick buck – now that short-term thinking is coming home to roost and they want us to pay for their short-sightedness.

  3. Michael,

    Great proposal!

    Maybe emphasis on the fundament goals and allowance for modification and granularity for improve billing over time (like only data during congesting peak times would collected).

    Also, linking it to Net Neutrality is counter productive. Net neutrality is a just politics.

    Alex, I think UBB is absurd because people don’t trust the incumbent ISPs, not for a technical reason that can’t be fix and I don’t think that Measurement Canada could ever be dynamic or thorough enough to do this. There isn’t any standard measurement that everyone would be happy with anyhow. Every consumer has different needs, gamers may want low latency, on-line stock traders want reliability, etc.

    Overall, if UBB leads to complex bills made to hide extra unnecessary charges then that is opposite direction of where we should be headed.

    The fundamental goals should be low costs, high speeds, reliable service, good geographic coverage, understandable billing, innovation, adaptation.

  4. Not quite…
    Martin,
    “Low latency” and “more reliable” sound great on paper, but they’re really red herrings in the discussion. I can’t understand for the life of me how you can add value to an internet connection by making it “more reliable”. This claim is nebulous and gives the telecom companies fodder to spin our heads with.

    An internet connection couldn’t be any simpler than exactly what it’s described to be. The added features and economics behind it is really just a way to re/de-value and fracture service for baseless increases in fees.

    Yes, there is a standard unit of measurement and we’ve already had it since the dawn of computer networking. The service providers are already billing on this basis. It’s the potential rate of transfer. This is bandwidth correctly defined (bits per second). Most people know it as the various tiers of service: Nitro, extreme, lightning, lite, etc… You pay more as you go up because you get a faster connection. The total amount of data transmitted means nothing as per the strain you put on the network with the connection.

    Again, let’s not try to re-politicize the issue. It is technical in nature and has some very concrete science behind it. We’re not dealing with opinions here.

  5. Alex,
    If you look at any high end internet router they are built with reliability and low latency as one of their main features – redundant processors, dual power supplies etc. They have multiple connections to major providers, backup etc. So as a network this it may not matter to the average user but one a network basis it actually does matter if you don’t have it.

    I think you are only looking at your own usage. Statistically speaking if someone uses a lot of bits in a month then the probability is that you are contributing to congestion than someone who is not. Bernoulli and Poisson proved that. Sure its not granular, but it’s a starting point. Personally, I would like to have a faster bps rate so I could stream a HD movie but I can’t since I don’t have Nitro since I don’t do it often. If they gave everyone Nitro at a the price and of Lite and then just charged fairly for the overage then I would be very happy. Of course, the problem is can they do it ‘fairly’.

    If you skip past the opinion part to the bottom of Michael’s proposal, to his 4 conditions, his first condition says the ISP would have to demonstrate the specific network congestion issue that they UBB is combating. Along with other conditions, the proposal seems to allow for technical aspects.

  6. Do you really want to pay the price for a 10 Mbit connection?
    A lot of people here have pointed out that you should just be billed for the speed of line you want, and not how much you want to transfer over it. However, that doesn’t work, because in that model, most people wouldn’t be able to afford the 10 mbit connection that they currently have. If the ISPs weren’t overselling, we’d all be stuck with 384 kbit/s lines, unable to watch Netflix, watching as our web pages slowly download, and barely able to have a voice conversation online. By limiting the throughput, they have effectively put a cap on the amount of time you can use that whole 10 mbit. If you want to use 10 mbit 24/7, you are going to pay through the nose for it. However, most people are ok using their connection such that it’s not going full bore all the time, and for them, the caps are ok. The only problems I had with the initial UBB proposal was that you had to pay upfront for your bandwidth, and that going over what you paid upfront for resulted in huge overage fees. If you could pay after for what you ended up using, and only charge $5 for 40 GB, then I don’t think most people would have had as much of a problem.


  7. “In the case of an ITMP that results in any degree of discrimination or preference”

    Did they ask Bell if they’re throttling their IPTV together and in proportionate measure with the internet traffic?

    “IPTV is different” said Mirko, and Konrad was satisfied with the answer.

    Nap.

  8. Martin, I work with computer equipment daily. The things you’ve outlined as sophisticated features are pretty standard fare. If not at the very least expected.

    The whole idea that there is network congestion are not only misleading but have yet to be substantiated. So there’s zero reason at this point in the discussion to even consider traffic management of any type as necessary.

    If there is any situation, it’s at the idle hands of the telecom companies who decided to suck up profits instead of reinvesting in their business. That’s just common sense.

  9. If you look at Michael’s last posting, you will see that the ISP interconnect fees generally run between $1-5 per Mbps. A 1Mbps connection will allow about 324GB/month if used constantly all month. The ISP does not pay any more, or any less, if it is used to transfer 30MB of data or 300GB of data. They are paying for the constant availability of the connection at a certain speed, regardless of usage.

    There will be an overhead for the ISP to bring 1Mbps of connection through their network into your premises, which is currently an unknown number. Add on a markup over that.
    Bringing 10Mbps to your premises isn’t all that much more expensive for the ISP than bringing 1Mbps, as long as they have the available capacity. It certainly isn’t 10 times as much. If they don’t have the capacity, then they may need to upgrade certain portions of their infrastructure. If they can offer 100Mbps local connection capability, then obviously the local network segment must have the capacity.

    There is perhaps an element of truth in the possibility that a single subscriber may not be able to afford a 10Mbps “full usage” connection. This is where the ISP starts playing the statistics, and oversells connections at a rated “speed” for a lower price, based on the assumption that the subscriber won’t use it 100% of the time. When all of these “oversold” connections use it at the same time, we get congestion.

    At any rate, the “problem” isn’t usage per se, it’s all that usage during peak times on oversold connections. Blanket UBB isn’t the right answer. UBB during peak congestion times might be. I’ve mentioned it before, but a constant trickle feed of 40KB/sec uses more “monthly data” than a 3GB usage over a 2 hour peak period each night of the month. But the impact on congestion of that 40KB/sec trickle is unnoticeable compared to the congestion caused by that 3GB peak period usage. Some people have the option to schedule their usage outside of congestion periods. The ISP has to carry the overhead for the infrastructure connection “speed” even if it’s not used. Usage outside ISP congestion periods certainly isn’t contributing to the issue, so why should it be penalized?

    It’s not that hard to do the math and reverse the data caps into “Bps” values if you want to extrapolate the “bandwidth” the ISP is really allocating for your always available connection.

  10. Alex, your’e absolutely correct. The CRTC needs to not cave in to the telcos and focus on why the telcos aren’t taking care of the congestion by adding capacity. That’s the bottom line.

    UBB could have been an interesting billing option, but not as a congestion solution.

  11. Try not to misrepresent what I’m saying. I refute it both as a billing option and as a solution to congestion.

    How can it be either? It has no technical merit and is based on made up interpretations of what constitutes strain.

    If you’re going to bill someone for something, you have to bill them fairly. In this sense, UBB is unable to offer that as the metric it measures doesn’t accurately reflect the cost of the service.

    Basing anything on a measure of data transmitted over a billing period should be outlawed entirely. It’s completely made up.

  12. So when you get your electricity bill you don’t like that you pay by the killowatt hour and would rather just pay by the Amp?


  13. @Martin: “So when you get your electricity bill you don’t like that you pay by the killowatt hour and would rather just pay by the Amp? ”

    If electrical power would be free, like Internet data, then it would be a good option. Since Hydro’s costs would be strictly related to maintaining your line in working condition, without having to pay any power plants.

    Nap.


  14. Of course, if we had competition, we would have had the innovation and infrastructure expansion that would make all this moot. It still isn’t too late.

  15. @Nap: but once they put in a Hydro dam isn’t the electricity free? The water is doing all the work.


  16. @Martin: “but once they put in a Hydro dam isn’t the electricity free? The water is doing all the work.”

    Not really. You have to charge the price of the dam, turbines and their maintenance. So the dam works like any other power plant, charging Hydro which in turn charges you.

    Whereas on Internet it is Google & Youtube that are paying for pumping their data into the backbone. Imagine a power plant that pays Hydro to consume MW?

    Nap.


  17. …and not only that, but Hydro charging you too for those MW that you helped to consume?

    Nap.

  18. Scott Watkins says:

    Another Modest Proposal
    Why not nationalize the last mile?

    Seriously. I am pretty sure subsidies paid for most of it anyways, and given the fact that it is where most of the monopolistic tendencies are occurring, an independent, regulated governmental organization or crown corporation would level the playing field.

    Ownership of the last mile outside of a dwelling or business should belong to the crown, and be maintained by a local firm based on regular, competitive contracts. All “central office” facilities – increasingly that box down the street technicians are always fiddling with – would be required to provide space for competitors to place their fiber-to-DSL/fiber-to-cable equipment. From what I understand, that piece of equipment is usually a simple box not much bigger than the router in your home. a simple extender on top of the box would provide enough space. I doubt sincerely that power would become an issue.

    What we get from this? Decreased congestion due to multiple routers being in place (I’m told there is no appreciable congestion at the fiber level) and a much more competitive environment, with a level playing field for all providers.

    A great place to start and test this idea is in providing broadband to rural communities, as in many cases no fiber has been laid there yet. Government lays the fiber, Government sets the rules.

    If it works there, do it everywhere.

    K, I just fixed the internets. Order of Canada, please.


  19. @Scott: “Why not nationalize the last mile?”

    And what would Bell do then? When their IPTV would be considered “heavy usage”, throttled and charged by the GB?

    Muhahahaha.

    Nap.

  20. Scott Watkins says:

    @Napalm: “And what would Bell do then? When their IPTV would be considered “heavy usage”, throttled and charged by the GB?”

    Actually, it would be just another piece of traffic looking for bandwidth. If congestion is an issue on the lines they service, they’d have to beef up their fiber-to-DSL switches. No effect on other providers who have their own switches. Their bandwidth, their problem.


  21. @Scott:

    And what happens to Bell’s IPTV if the “Wired Carrier Crown Corp” decides to charge UBB? IPTV traffic will be taxed, right? (and those would be terrabytes per month no less)

    It all boils down to the fact the Bell carrier and TV divisions got a very sweet deal between themselves where the carrier guys apply UBB to everyone except the Bell TV guys.

    Nap.

  22. @Nap, that’s not the way peering works. Google does not pay ISPs to carry its traffic and YouTube is Google. It ruins your credibility if you make stuff up.

  23. Everyone is missing the simplest solution
    and it’s the most painful for these companies:

    DROP YOUR CABLE TV SUBSCRIPTION!

    This has nothing to do with congestion. This has nothing to do with “overuse”. Those are just red-herrings in this discussion. These companies don’t give a whit about “overuse” unless it results in people dropping them as an ISP.

    This has everything to do with content creation and delivery. These TV content delivery companies make money their money on package subscriptions. They do not make their money sending bits to computers.

    The fact is, Netflix et al will eat the cable companies’ lunch. They offer the content people want to watch at a price that is much lower than what the traditional delivery companies charge. For the $60/mo you pay just to watch TV, you can have a good internet plan _and_ pay for a subscription on Netflix. You can watch TV, send email and browse the web for much less than using traditional mechanisms.

    And it’s scaring the crap out of Bell, Rogers, Shaw, Telus…. They know the ride is over. These choices are coming up really fast and they don’t know what to do with themselves. Their business models are 30 years old and their infrastructure decisions are based on those models.

    And then this little thing called the Internet just messed it up for them. And now we have Bell and company pushing for UBB.

    Welcome to the new world.

  24. Scott Watkins is win. Good observations by napalm about how much congestion the actual TV programs being sent by Bell/Rogers cause, their internet problems are made up. I bet the net uses less data.