Opinion: The CRTC’s UBB decision and a solution to retail Internet metering

In a decision announced on Jan. 25, 2011, the CRTC permitted the telcos and cablecos that own communications networks to apply usage-based billing (UBB) to the contracts they make with independent (reseller) ISPs. The CRTC, however, did a spectacularly bad job of providing an analytical justification for its decision. If UBB is to become the norm—and we think the economic case for it to become the norm in the pricing of Internet usage is strong—then there has to be a convenient, cheap and reliable way of metering usage at the household level. That is the issue under discussion here. Key aspects of metering usage for the Internet Currently it is only possible to measure usage at the ISP points. Trying to meter usage (usually measured in gigabytes) is problematic because there is no way for a strand/network to “know” either the sender or the receiver of any given packet (into which digital files are broken up) while it is en route from sender to recipient. Throughput in gigabytes can be measured only when the packets come together from their possibly far-flung journeys and are “reassembled” at the ISP used by the recipient of the digital file. Speed (called bandwidth) and throughput or usage—measured in gigabytes—are intertwined. Internet throughput speed is measured in amount (of bytes) per unit of time (usually seconds). Think of bandwidth as a hose that carries water. A larger hose will carry more water at the same speed—or pressure. But a smaller hose will carry more than that if you increase the pressure. Same thing with the Internet backbone. A copper wire, even if multiplexed, has physical limits on how much data it can carry, but an optical fibre can carry much more because it’s a bigger hose. Internet use always has two sides. Both the sender and the recipient use the Internet to move digital files. This is true whether the sender is in effect “passive” as in the case when the recipient accesses a website or downloads material from a website or from his ISP. The peak load problem seldom depends on the capacity of a single network. Over time, as more people went online at home and abroad, and the number of interconnected networks increased (the Internet is a network of networks), the time-of-day congestion problem also changed. Everywhere, the Internet became less a diurnal phenomenon than it was because it didn’t matter where the infrastructure was physically located. An optical fibre cable near Toronto was (and remains) every bit as accessible from Sri Lanka as a similar fibre in Delhi or Beijing, or Portland, Ore., or anywhere else. This fundamental characteristic means that the Internet can be accessed and data packets switched through it anywhere at any time, day, or night. Thus, peak load limitations will more likely derive from “server farms” that are owned by or leased to specific telcos or cablecos. They can handle peak load questions by installing more, larger, and faster hard drives. UBB plans of telcos and cablecos The actual proposed UBB pricing formula is far from a simple linear relationship between gigabyte usage and total price per month. One needs to look at the appendix to CRTC Telecom Decision 2010-255, issued on May 6, 2010, and the websites of the telcos and cablecos to begin to understand the pricing plans. The plans have several variables: (i) bandwidth or speed, (ii) the capped usage quota in gigabytes per month, (ii) the overage charge for gigabytes above the capped quota per month, and (iv) the duration of contract. Selecting a plan: Problems facing the household user First, there is a serious asymmetric information problem. How many buyers (households) know what their current level of usage (in gigabytes) is to guide in selecting a plan if they currently have an unlimited use plan with an independent ISP? How can they find out? We visited the website of two of the largest networks (Bell Canada and Cogeco Cable). Bell, for example, has a usage “Estimator” and a questionnaire with three broad categories, each referring to several types of usage: low usage (websites, email, instant messaging, photos); medium (online games, streaming radio, music tracks); and high (video/TV streaming, standard definition movies, high definition movies). The stand-alone ex ante usage calculators on the Internet that we checked-out, SimplySwitch and another on www.webuser.co.uk, are seriously flawed (even if one can provide reasonably accurate answers to the questions): Several major uses are omitted; the focus is on uploading; many types of downloading (e.g., photos, Skype, Twitter, online chat audio files, video files (e.g., sent to YouTube, a big user) were omitted. Responders would likely tend to significantly underestimate their usage. One site left out the mother of web usage, namely movies (whether downloaded or forwarded). One HD movie uses about two gigabytes, which is equivalent to sending 105,000 short text emails. Further, under Bell’s “Essentials Plus” 2 GB plan, the following information is provided: “emails you can send, no attachment, 200,000; web pages you can visit, 50,000; digital pictures you can download, 8,192; songs you can download, 409; movies you can download, 2.” Bell did not explain that each item is independent and not additive. A customer cannot do them all within the plan’s 2 GB limit. Second, if it is true that only about 10 per cent of the households on the networks’ UBB plans violate their cap in any given month, then the telcos and cablecos are benefiting from what might be called the Coleman Mustard effect. The ennobled owner remarked that he got rich from the amount left on the plate. This would seem to be attributable to fear of the deadly overage charge which as several critics have noted is 10 to 50 times the estimated costs incurred by the networks. However, The Globe and Mail reported on Feb.10, 2011 that, “ Shaw’s customers have monthly download limits, but are not charged so-called ‘overage’ charges for exceeding their set amounts—relying instead on a tactic used in the United States, where heavy users who go over are contacted about their contract terms. In the U.S., after being notified, customers who exceed the limit again in less than six months can be booted off.” Given the market power of the networks, Shaw’s policy could well change when UBB becomes ubiquitous. The Bell and Cogeco websites provide a “tracker” to measure ongoing usage by their household customers. It includes usage-to-date (for the current month), and average usage for past three months. There is a serious moral hazard problem of relying on suppliers such as Bell or Cogeco to measure usage. Contrast the situation at the gasoline pump. The pump is owned by the service station, but government checks the meter to ensure that it accurately measures the amount pumped. There is no government regulator checking the usage meters of Bell or Cogeco. Recall that Bell has had serious problems with its meters. Third, it is hard for households to find out, ex ante, the size of digital files they may wish to download. For example, Netflix is said to account for 20 per cent of all net usage in the U.S. during peak periods. We checked the Netflix website and could not find any information on the size of a movie file for any of its movies or other video programs. (It may be available to persons who use the service.) We found a site that lists over a score of software programs said to measure net usage in some form. We downloaded the free ShaPlus Bandwidth Meter 1.3 and found it to be simple and useful. It provides in real time the usage in gigabytes for each online session, the usage for the day, and for the month. The start of the billing period can be varied. We do have a concern, however. This meter is connected to the Internet interface inside the user’s software. It would be easy to insert a line or two of code instructing it to note and report which websites were being accessed, or even to divulge e-mail content, etc. And since the household user requests and installs the download, the program can then get around the usual virus/malware software firewalls. Proposal for ISPs The biggest problem facing households shifting to UBB is estimating their usage ex ante so as to select the “best” plan. Here is better way: Both reseller and network ISPs could provide new clients with a virus-free metering device as a means of assisting them to determine which plan would work best for them. The steps are simple: First, the customer contacts the ISP to discuss service. Second, the ISP makes available to the customer 30 days of service at a “reasonable cost” (obviously subject to discussion). The ISP requires a valid credit card as one condition of the offer. Then the household downloads and installs something like the metering device ShaPlus, discussed above. Third, the customer is instructed to use the service for the 30 days in their normal use pattern. Fourth, at the end of the period, with that use pattern information in hand, the customer is then directed to the most advantageous of the plans that would fit that use pattern. It would be something like the bandwidth used during the trial month plus about 20 per cent, to cover increased usage over time. The customer would then sign up. Fifth, if the customer did not choose to use that ISP’s service, he or she would be charged whatever appropriate rate for one month’s service to the credit card, and would then be free to seek service from a different ISP. Sixth, if the customer signed up, the monitor would remain in place to provide real time information as to future uses. Customers would more likely attend to that information in the future seeing that they used it at the start of the service. Seventh, it would be simple enough to have the ISP notify the user when they approach the upper limit of the plan prior to the end of the month, and to provide specific information as to what the overage charge will be if the plan limit is exceeded—or the terms of adopting another monthly plan. The advantage to ISPs of adopting this approach would be to show the world that they are going the extra mile to assist customers in choosing a service plan based on actual information derived from that user. A further step would be for the ISP to have an independent testing service monitor and verify the accuracy of its metering device. Perhaps it would be better to have weights and measures branch of government add Internet meters to their list of meters to be checked for accuracy. Don Homuth is a newly repatriated native-born Canadian currently living in the U.S., and an “early adopter” of Internet technology. W.T. Stanbury is professor emeritus, University of British Columbia. news@thewirereport.ca

Opinion: The CRTC’s UBB decision and a solution to retail Internet metering

News | March 28, 2011

Unlock all the Canadian telecom, broadcasting and digital media news you need.

FREE TRIAL

SUBSCRIBE

Inside The Wire Report

Information

Hill Times Publishing

Get in Touch With Us