Bandwidth is the message. Taxing ISPs in the era of music as data flow

A few thoughts on copyright law in the digital world. Skip to part C if you’re hungry. Background info here

Music is consumed as an immaterial service rather than single object-unit. Accordingly, when we ask ourselves how to make money from music in the 21st century, we should be looking at both massive downloads and individual albums or songs. Monetizing music requires us to focus on the meta-level of aggregate music consumption, rather than copyright infringement of a specific album or song. We propose taxing Internet service providers (ISPs) who profit from increased bandwidth demands from peer-to-peer (p2p) downloads. Provided privacy safeguards are in place, deep packet inspection allows a fair redistribution of income to content makers.

Regulating ISPs

ISPs are the major winners of digital music consumption. They provide the main conduit for data transmission, and 70% of Internet traffic consists of file downloading–including movies and softwares. Since ISP revenue is tied to the availability of digital music for download, it would be efficient to collect a levy at this level of the transmission mechanism.

-Levy and only be collected on fixed medium

The Federal Court has restricted the ability to collect a levy to fixed mediums. Its reading of the enabling statute for imposing a levy on ‘audio recording medium’ excludes mp3 players. Because those devices have recording and playback capabilities, they are conceptually different than fixed medium. In first instance, the Copyright board adopted a functional understanding of mp3 players to determine that it was “ordinarily used by individual consumers” to copy music and as such was within the purview of material subject to a levy. On appeal, the Federal Court adopted a formalist stance and contrasted the fixed form of blank cds to playback devices. It relied on Parliament’s explicit rejection of levying audio tapes as evidence of its intention to limit the levy to fixed medium. Given ISP’s rising and prominent role in digital music consumption, it is high time for a legislative update allowing a tax to be collected from mp3 players. Although only tangent to ISP levy, this would be a significant step forward because mp3 players are also a significant piece of the puzzle in digital music consumption.

– ISPs are not “communicating”: beyond the SOCAN case

The Supreme Court has decided against imposing a tariff on Internet communication in the SOCAN case. The author collective society sought and was denied permission to impose a tariff on ISPs, as it did with blank audio media. The Copyright Act carves out an exception for companies “providing the means of telecommunication” from copyright liability. The Court decided that ISPs were covered by that exception, and as neutral transmitters could not be subject to a tariff.

The Court failed to recognize that en masse digital music consumption drastically altered the quid pro quo between expanded copyright protection and protecting ‘wholesale’ communication providers. IPSs are not innocent disseminators, they derive 70% of their demand from p2p file sharing. Citing 1998 parliamentary committee concerns over unfairly burdening infrastructure providers, the Court portrays ISPs as mere “conduits” in the communication process. Failure to charge for individual downloads does not suffice to call ISPs mere intermediaries. The point we belabored above is that music is now consumed as a service. It should therefore be charged as a service. Consumers do not pay for individual digital objects, they pay for the ability to download music en masse. While the Court agreed that “the availability of ‘free music’ is a significant business incentive’ for ISPs”, it did not connect this finding to the overarching goal of “a balance between promoting the public interest in the encouragement and dissemination of works of the arts and intellect and obtaining a just reward for the creator”. Had it considered the equation from the perspective of music as service, it may have found that taxing music consumption at the ISP level properly rewards creators and encourages the dissemination of works. ISPs charge for bandwidth. Digital music creates a demand for such bandwidth offer. Given the clear economic linkage between ISP revenue and music downloading, ISPs are a legitimate point of royalty collection. Given ISPs financial stake in music consumption, the 1998 parliamentary committee concerns over unfairly burdening infrastructure providers are no longer legitimate

In the cases outlined above, the Court adopted an originalist stance which may be legitimate under the doctrine of judicial restraint. However, it invited lawmakers to adopt specific copyright legislation tailored to the digital era. We now turn to what those possibilities might be.

Implementing a levy on ISPs

The levy on the sale of blank audio media distorts royalties distribution to the advantage of mainstream music. Functionally, ISPs are better positioned to collect royalties. Analyzing internet traffic with deep packet inspection would allow accurate computation of digital music consumption.

– Levy on blank audio media

In 1997, the Liberal government implemented a levy on blank audio media to mitigate music industry losses due to private copying. A collective society of music makers apply for tariff certification at the Copyright Board. Once the tariff is approved, the society administers its collection from CD makers and redistributes it to music makers. The Canadian Private Copying Collective (CPCC) has distributed over 200 million dollars in royalties since 2000.

industry capture : metrics leveling playing field in favor of analog success The CPCC scheme favours the entrenched property interest of established analog music industry. It relies on commercial radio airplay and CD sales to allocate the perceived tariff. By CPCC’s own commissioned report, in 2006-7 21% of the music copied onto CDs is sourced in the Internet. That’s 27.5 million songs. Downloads also accounted for 24% of music on mp3 players. Accordingly, digital download is relevant to calculating levy redistribution. Moreover, trends in digital music consumption may differ radically from analog sales. Viral songs–think Das Racist’s Pizza Hut and Taco Bell–are invisible in the music channels surveyed to establish the redistribution scheme. The scheme is divorced from digital informal sources, and data from the analog world maps uneasily with online music consumption patterns. This mismatch levels the playing field to the advantage of established music interests. Mainstream music, which is having difficulty getting exposure in the digital world, extends its advantage by sticking to analog metrics and formal sales channels. Newer music genres, which have completely bypassed the analog sales step, may be doing extremely well in terms of p2p downloads and college radio streaming. Yet downloading is invisible in CPCC metrics. SOCAN only samples college radio, offering a partial portrait at best. Those songs transferred to CDs will not get their share of the levy. As a result, the competitiveness of emerging music makers able to grasp digital consumption patterns is flattened. Their advantage in terms of digital consumption (streaming, download) is ignored in the calculation of royalties. In this sense, the current calculation model distorts the market in favour of established music makers in the analog world.

Existing tracking systems can be tweaked to include reporting sources more reflective of digital downloads. But technology exists to inspect the content of Internet traffic in order to obtain accurate data of digital music market share. If we can rely on the raw data instead of an approximation, what is stopping us?

– Deep packet inspection of bandwidth use for accurate levy redistribution

DPI addresses the fundamental mismatch between the way music is made–per unit, as in the analog world–and the way it is digitally consumed–downloaded and streamed as a service. DPI allows ISPs to analyze not just the header, but the actual content of the information packets requests handled by its servers. It can accurately determine what songs are downloaded and reflect the market share of each copyright holder, leading to an exact distribution of a copyright levy. Individual song download could therefore be compensated. In the context of digital music consumption, calculating revenue entails translating massive data flow in segregate revenue. Doing so through DPI is a workable compromise because it allows unit producers to get a fair share of the en masse distribution enabled by large bandwidth availability. It is an accurate calculation tied to actual network traffic, not an approximation securing analog vested interests as with CPCC’s current model.

Privacy DPI opens the door to intrusion on user privacy. Monitoring the content of bandwidth traffic for levy allocation may facilitate unintended uses of the information. Alleged national security concerns, for example, have overridden commitments to information privacy. For instance, the American government used a packet sniffing software (Carnivore) to monitor and intercept private email communications, arguably violating civil rights. Another concern is targeted advertising: ISPs may derive advertising revenue from online habits without the consent of consumers. Like any technology, DPI can be used both for laudable goals and in contravention to morality, law and public perception. Arguably, deep packet inspection is now a necessary network administration tool. What we need is regulation to curb illegitimate use of personal information data.

Legislative framework ensuring privacy of digital personal information The Personal Information Protection and Electronic Documents Act (PIPEDA) is a law of general application covering the collection of personal data across industries. Generic privacy protection obligations apply to ISPs gathering personal customer information in the course of their service. ISPs should provide an accessible, plain word explanation of the purpose and scope of DPI to ensure meaningful consent. Other safeguards include limiting collection, use, disclosure and retention to the strict necessary, as well as appropriate safeguards of the information. For instance, automating DPI and limiting access to preauthorized personnel would protect citizens from the risk of discretionary manual intrusion and data manipulation.

The Canadian Radio-television and Telecommunications Commission (CRTC) has specific jurisdiction to ensure the privacy of Canadians in telecommunications. It can tailor a privacy framework to DPI for the purposes of copyright levy calculation.

Net neutrality DPI questions the principle of net neutrality–indiscriminate data transmission. It will likely focus on p2p transmission because this is the preferred method of download. This scrutiny will likely mean slower p2p traffic, but network administrators have the tools to rebalance this effect if they put more resources to optimizing p2p traffic. DPI also threatens the End-to-End principle, defined as “intelligence at the edges of the network, not within its core infrastructure”. Inserting analysis capabilities in the middle of the transmission process, DPI changes the mission of the Internet from neutral transmitter to involved monitor. Critics argue that DPI is a core part of network performance tools and many if its uses are legitimate and desirable for consumers. As with the privacy argument, finetuned regulation and increased public awareness are more realistic avenues than a blanket ban of DPI.

A crude DPI-based copyright levy trades off user privacy for fair remuneration of music makers. But effective legislative privacy safeguards can balance the stakes of different constituencies and serve as a model for other challenges raised in the electronic ecosystem, where unauthorized use is a given. In a conservative tax policy context, will the Canadian legislature reject this idea as a distortion of markets? One must be clear that ISPs are the collection point of the levy, but the cost is passed onto the consumer. Nevertheless, the cost of implementing DPI falls upon ISPs. Given their expertise in network administration tools, they are the most efficient actor to carry the task.


Is it feasible to update copyright laws to the digital world? Given that music is consumed as a service, charging every single object in individual transactions is unrealistic. The pricing model has to apply at the level of massive downloads. While imposing a levy on ISPs provides an interesting avenue for upholding copyright benefits, the dynamics of digital music consumption strain the very applicability of copyright’s conceptual foundations in the property regime. The difficulty to exclude others and the non-rival nature of Internet-sourced music remind us that copyright is a tool to create a market for the fixation of an otherwise free idea. Copyright is not an end in itself, but a means to achieve the promotion and dissemination of ideas. Given the increased difficulty of enforcing the regime –DPI has privacy tradeoffs, and encryption will render it irrelevant sooner or later–, might it not be time to consider other ways to promote the dissemination of art and ideas? One avenue is to provide artists stipends–beyond welfare–, private endowments, and state grants. This will feed artists, not copyright lawyers. Another avenue is crowdfuding. Since p2p networks allow decentralized music sharing, this dynamic can, and does, work as an upstream finance model.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s