CommLawBlog : Broadcast & Communications Lawyers & Attorneys : Fletcher Heald & Hildreth Law Firm : Telecommunications, FCC Regulation, Communications Act

Within a month of the FTC’s exit, the FCC had fired off a Letter of Inquiry (LOI) in an effort to figure out whether Google’s data collections had broken the law. The law in this case is Section 705(a) of the Communications Act (which, oddly enough, is codified as 47 U.S.C. §605(a)). In relevant part (that would be the second and third sentences of Subsection 605(a)(6)), it bars the unauthorized interception, followed by divulgence, publication or use, of certain radio communications. 

The LOI sought vast amounts of information and documents about Google’s Wi-Fi data collection activities.  Google reacted like any public-spirited organization with nothing to hide would – by cooperating fully, opening its files to the FCC and happily walking the agency through the complexities of its data-collection process . . . NOT. Au contraire, Google mounted an impressive effort – some might call it stonewalling – to keep the FCC in the dark.

The LOI was designed to bring in huge numbers of documents – including internal emails relating to the data collection process – so the Commission was doubtless disappointed with what Google produced: a very small handful of documents, a few apparently unhelpful interviews, and no emails at all. The paucity of materials presumably stemmed, at least in part, from Google’s somewhat circumscribed approach to the LOI. According to Google, it had “not undertaken a comprehensive review of email or other communications” because doing so “would be a time-consuming and burdensome task”. (Having responded to our share of LOI’s, we are sympathetic to Google’s concerns here; we only wish that we had thought to raise the “Gee, that’s a lot of work – we think we’ll pass” defense.) 

Google also chose not to identify any of the individuals responsible for authorizing its collection of Wi-Fi data or any employees who had reviewed or analyzed the Wi-Fi communications collected. Consistent with this insistence on anonymity, it also redacted the names of its engineers from the limited documents that were produced.  Google claimed that identifying its employees “at this stage serves no useful purpose with respect to whether the facts and circumstances give rise to a violation”.

The FCC sleuths were able to identify the engineer who developed the software code that Google used to collect and store payload data. In the NAL he is referred to as “Engineer Doe”. We’re guessing that’s not his real name . . . not that knowing his real name would help anything. According to the NAL, Engineer Doe lawyered up and took the Fifth, effectively slamming the door on that potentially useful source of information.

And the icing on the cake: Google declined for nearly the entire length of the investigation to provide a verification, under penalty of perjury, from any corporate official with either first-hand involvement in the data collection effort or personal knowledge of the information contained Google’s response. Such a verification of the accuracy of the response was specifically required by the LOI and is SOP in dealing with the Commission.

Undeterred (and perhaps nonplussed), the FCC issued a supplemental LOI, but that didn’t result in more useful intel from Google. That was followed by a demand letter making sure that Google knew that the FCC was really, really serious – it ordered Google to provide complete responses to earlier requests and requested additional information. Same non-result. And then a final supplemental LOI. Ditto. Throughout the process, the Bureau was also in touch with Google by phone and in person. All to no avail.

So despite the fact that the Commission dipped its regulatory arms deeply into what appeared to be a trough brimming with useful factual information, the Commission came up empty-handed.

But Google was not completely silent. While it kept its factual cards close to the vest, it laid out a legal argument to show that the data collection that occurred during its Street View process was not illegal. 

Google pointed out that the Wiretap Act – a section of the federal criminal code the relevant part of which you can find at 18 U.S.C. §2511 – effectively trumps Section 705(a) of the Communications Act, at least as far as the alleged misconduct is concerned. The Wiretap Act provides that it’s OK

to intercept or access an electronic communication made through an electronic communication system that is configured so that such electronic communication is readily accessible to the general public.

The statute defines “readily accessible to the general public” as, among other things, not being scrambled or encrypted. According to Google, the payload data it collected and accessed was only from unencrypted networks, not from any encrypted networks. QED: Nothing illegal happened here. With a casual wave of its cloaked hand, Google assured the Commission that these were not the droids it was looking for. 

And even if the Commission weren’t susceptible to Jedi mind tricks, what was it to do? Thanks to (a) Google’s refusal to provide virtually anything in the way of hard information, and (b) Engineer Doe’s Fifth Amendment embrace, the FCC had no evidence from which to dispute Google’s self-serving conclusion.

Presumably recognizing its predicament, the Commission beat a quick retreat, at least with respect to the claim that Google may have violated Section 705(a). Acknowledging that there is no FCC precedent on this particular question, and acknowledging as well its lack of evidence, the Commission declined to find any such violation here.

But the Commission wasn’t ready to let Google off the hook entirely. Obviously miffed that Google had, um, largely ignored the Commission’s LOI, and its supplemental LOI, and its demand letter, and its second supplemental LOI, and its various blandishments delivered by phone or in person, the Commission wanted at least something to show for its efforts. (Think Glenn Close in Fatal Attraction.)

So with a carefully honed prosecutorial knife, the Commission lunged boldly at Google’s capillaries, proposing a fine of $25,000 because Google’s “level of cooperation with this investigation fell well short of what we expect and require”. The base fine for such lack of cooperation would normally have been $4,000, but the FCC wanted to give Google what for, in part “to deter future misconduct”. For sure the $21,000 bump over the base fine should scare the bejeebers out of Google, whose 2011 gross revenues were a paltry $38 billion or so. With a forfeiture amounting to, what, not even one ten-thousandth of one percent of its annual revenue, Google obviously has much to fear.

Actually, the NAL subliminally suggests that Google has very little to fear, at least from the FCC. That’s apparent from extraordinary redactions that make many pages of the NAL look like Rorschach tests. (See the graphic, above, which depicts one page from the NAL.) Some sample redactions:

• “In response to the Supplemental LOI, Google expanded upon [REDACTED]. Google explained that [REDACTED]." "Google further stated that [REDACTED]."
• "In interviews and declarations, managers of the Street View project and other Google employees who worked on the project told the Bureau they [REDACTED]."
• "One engineer remembered [REDACTED]."
• "During interviews with Bureau staff, Google employees stated that [REDACTED]."
• "In both his written declaration and his interview with Bureau staff, the engineer characterized [REDACTED]."

Such redactions are rarely seen around these parts. Our guess is that these were done at Google’s request. The Commission, after all, would appear to have no incentive to withhold the redacted information. Google, on the other hand, is facing potential liability in a number of other venues where its data collection activities are still under investigation. It’s reasonable to assume that Google plans to stick to its stonewall approach as long as possible in response to as many investigations as possible. Inclusion of lots of informational tidbits in an FCC NAL available to the whole world would not be consistent with that strategy. So we’re guessing that Google asked the FCC to be nice guys and black out vast swatches of the NAL.

What’s something of a puzzle is why the FCC would go along with that request. After all, the Commission seems, rightfully, to be cheesed off at the way Google cavalierly thumbed its nose at the FCC’s investigation. Why should the Commission turn around and do Google any favors?

In addition to that practical question, though, the NAL leaves open other far more important legal questions.

Is Google correct that the Wiretap Act effectively permits interception of communications carried on unencrypted Wi-Fi networks? The Act’s language can be read to support that position, and there is nothing contrary to that reading in FCC precedent, as the Commission concedes. But the Wiretap Act was drafted in the 1980s, a decade or more before the advent of even the earliest Wi-Fi networks. It’s not at all clear that, in referring to “electronic communication . . . readily accessible to the general public”, the authors of the Act had in mind a home wireless network, as opposed to the technologies of a generation ago.

And even if the “readily accessible” exception turns out to be applicable where individuals fail to encrypt personal home wireless networks, should data collection of the scale and scope as the Street View project nevertheless be subject to some legal constraint? And at what point does the massive aggregation of data provide sufficient information to violate our expectations of privacy? 

We, of course, have no official opinion on this matter. If we did, though, we would have to say [REDACTED].

On reconsideration, though, the FCC determined that some MVPD services and translator/LPTV/Class A stations relying on over-the-air reception to obtain and redistribute TV signals are located more than 80 kilometers from the originating TV station’s protected service contour. In order to avoid disruption in those circumstances, the Commission opted to expand the notion of “within reasonable bounds” temporarily: it provided a 90-day opportunity (commencing with the effective date of the rules) for MVPD’s, TV translator, LPTV and Class A TV stations to request a rule waiver to allow them to register their receive locations in the TV bands devices database. This opportunity is available only for locations at which the TV programming is received over-the-air more than 80 kilometers from the originating station’s protected contour.

The initial 90-day waiver request filing period will expire on April 5, 2011. (Facilities that meet the geographic standards but don’t get licensed until later will have 90 days, starting with commencement of operation, to file for a waiver.)

Waiver requests should demonstrate how the operation of an unlicensed device near the relevant receive site would act to disrupt current patterns of television viewing. After a waiver request is received, the FCC will put it out for public comment and then will make a determination as to whether it will be granted.

The Commission has not yet provided any special instructions for the filing of such a waiver request.  Check back here for updates on that score. But absent any such instructions, it would appear that filing through the Secretary’s office with a reference to ET Docket Nos. 02-380 and 04-186 should do the trick. Electronic filing in the dockets might also be a possibility – but, again, the FCC hasn’t given any guidance yet. We’ll post a follow-up on this as developments warrant.

This inclusive non-decision may reflect the FCC’s often-expressed distaste for “picking winners and losers.” Or it might follow from the FCC’s having neglected to state, at the outset, the criteria it would use for selection, an omission that leaves it vulnerable to challenge from the losers. This problem does not arise, of course, if there are no losers.

One applicant and a wireless microphone coalition challenged the impartiality of some other applicants. The FCC responded with a stern injunction against the administrators engaging in anti-competitive practices, and a promise of careful oversight.

Here at CommLawBlog, we have two concerns.

The FCC could have decided to manage the database itself. It certainly knows how; it keeps close track of millions of licenses. The FCC opted instead to farm out the work. With one or two administrators, that might have been a labor-saving move. But riding herd on nine of them, some inexperienced, each working with a database built to a different design, might turn out to be more work for the FCC than just doing the job on its own.

The other problem relates to data quality. Each administrator will keep its own database, but all nine must reflect the same underlying reality. Some of the data are slow-moving and should be easy to maintain – TV station contours, for example, and locations of protected TV receive sites, such as cable TV headends and TV translators. Potentially more troublesome, though, will be wireless microphone users’ frequent and changing registrations as they sign up for short-term interference protection at sporting events, political events, concerts, etc. These data will be volatile.

Suppose NBC, say, as part of its planning to cover an event, logs on to its preferred database administrator and registers a few dozen wireless microphones by date, time, place, and TV channel number. That information must be made available to every TVBD in the vicinity of the event, through every database administrator. Accordingly, the administrator receiving the registration must quickly and accurately disseminate it to the other eight, in a form that allows easy incorporation into their own, differently-designed databases. This kind of coordination is hard enough among two or three parties. We wonder whether nine can bring it off reliably.

And those nine will be competitors after the same business. It may become tempting for some to try making the others look bad by feeding them bad (or late) information. Even greater will be the temptation to cut costs by using ill-trained and badly supervised staff. Just as the hygiene of a shared kitchen quickly sinks to the level of the sloppiest person using it, so will the quality of the shared data reflect the least careful administrator.  (Users may appreciate the lower cost . . . at least until they realize that you do, in fact, get what you pay for.) 

To say, “You’re all winners!” is fine for T-ball. But maintaining a large and critical database takes real skill and a large measure of dedication. We may all come to wish the FCC had exercised greater adult authority in making its choices.

The technical problems with white space devices center on avoiding interference to TV stations and the wireless microphones that have long used vacant TV channels. The original plan called for each device both to use geolocation – ascertaining its own position using GPS and consulting a database to find locally vacant channels – and also to “sniff” for TV stations and wireless microphones, a process called spectrum sensing. (The FCC exempted from geolocation certain devices under the control of other devices and, separately, allowed for the possibility of some sensing-only devices.)

The new decision confirms the geolocation requirement, with many critical details still to be fleshed out by the Office of Engineering and Technology. But the FCC has pulled back on sensing. When it tested spectrum sensing technologies several months ago, none of them worked well. This result surprised us, as white space proponents had touted sensing as the ultimate safeguard against interference. In some other universe, the agency might insist the promised technology function properly before it allowed deployment. This universe, though, works differently: the FCC’s spectrum-exploding train will not be de-railed, so they simply dropped the sensing requirement for devices that use geolocation.

Sensing-only devices are still allowed, but only under very rigid technical constraints that will be hard for manufacturers to satisfy. Because database checking will usually be the sole feature for avoiding interference, the FCC promised a rigorous certification procedure to make sure devices handle this function properly. Again, no details.

The FCC struggled, with only limited success, to accommodate users of wireless microphones in broadcasting, theater, movie-making, sporting events, and public gathering places like churches and auditoriums. The FCC will reserve two TV channels in each geographic area for wireless microphones, which it thinks will accommodate 12-16 microphone voice channels. Some parts of the country will also have other channels closed to white space devices and available for wireless microphones. Large productions, though, often use 100 or more. Microphone operators may request to have specific events entered into the white space database, which should (if all goes according to plan, that is) automatically keep white space devices away. Requests to protect unlicensed microphones must show that the channels free of white space devices cannot do the job. These requests will be subject to public comment, which requires 30 days advance notice. Without a database entry, and in the absence of spectrum sensing, the microphones will have no protection against white space devices on the same channel. 

In the end, the FCC believes wireless microphones should move to more efficient digital technology. But it did not address the difficult engineering problems that so far have barred this option.

The question of using vacant TV channels for backhaul links in rural areas is deferred.

Read the FCC’s news release on the decision and the full text of the White Spaces order.

[Blogmeister's Note:  This post has been updated as of 9/24/10 to provide additional information culled from the full text of the Commission's white spaces order.]

The Commission’s latest inquiry seems to respond in large measure to two aspects of the Verizon-Google Proposal. According to Verizon and Google: (1) as long as they comply with four general Open Internet principles (relating to consumer protection, transparency, non-discrimination, network management), Internet service providers should be allowed to provide other broadband services that would not be subject to the Open Internet rules; and (2) wireless Internet access service providers should be subject only to the transparency principle at this time. 

Well, isn’t that special?

With respect to the question of “specialized” services, the Commission is concerned that carving out exceptions for such services could undermine the ultimate effectiveness of the Open Internet principles. “Specialized” services include things like subscription video, telemedicine, or business services to enterprise customers. They’re services that are provided over the same last-mile facilities as “broadband Internet access service” (BIAS). They can in many instances look just like the kind of services normally available to the public through standing Internet access. But they are available only to those who sign up, and they typically incur costs beyond ordinary Internet access.

And that’s the problem.

Such “specialized” services can attract important private investment and can provide those willing to pay with new and valued services. There is therefore good reason to foster them by, for example, exempting them from Open Internet principles. In the NPRM the Commission appeared to recognize this and, accordingly, sought comment on how to define such services.

But the Commission is now focusing more on the possible risk that, if providers avail themselves of such an exemption, the whole point of those principles might be defeated. Providers might use the exemption to avoid Open Internet principles with respect to delivery of services that are substantially similar to standard broadband service. Or providers might devote so much of their capacity to such “specialized” services that the incentive and resources to expand standard broadband service would “wither”. The potential for anti-competitive conduct exists as well. And the risk of any of these undesirable consequences would be exacerbated if the public’s choices of Internet broadband service providers are unduly limited.

With these concerns in mind, the Commission suggests six possible approaches:

“Definitional Clarity” – This would involve defining BIAS “clearly and perhaps broadly”, with the Open Internet principles applicable to such service. “Specialized” services not subject to the Open Internet principles would be those services with a “different scope or purpose than broadband Internet access service (i.e., which do not meet the definition of broadband Internet access service)”,. This is somewhat similar to the approach suggested in the Verizon-Google Proposal, which characterized the exempt services as “additional or differentiated services . . . distinguishable in scope and purpose from broadband Internet access service”. The main difference, it would appear, is that the FCC is contemplating a more inclusive definition of BIAS that would, presumably, narrow the range of services entitled to the exemption.

“Truth in Advertising” – This heading – quoted directly from the FCC’s inquiry – is curious. The Commission’s brief summary under this heading refers to prohibiting providers from marketing “specialized services” as a substitute for BIAS. The Commission also suggests requiring providers to offer BIAS as stand-alone service. It is not clear that either of those suggestions necessarily involves “truth in advertising”.

“Disclosure” – This approach would entail the required disclosure, by providers, of information about specialized services, including their effect on capacity and the BIAS market.

“Non-exclusivity in Specialized Services” – Commercial arrangements for the offering of specialized services would have to be offered to all qualified parties on the same terms.

“Limit Specialized Service Offerings” – Broadband providers would be allowed to offer “only a limited set of new specialized services, with functionality that cannot be provided via broadband Internet access service”. The Commission offers telemedicine as a possible example.

“Guaranteed Capacity” for BIAS – Broadband providers would have to keep “providing or expanding” capacity allocated for BIAS regardless of any specialized services offered. Moreover, the provision of specialized services would be prohibited from “inhibiting the performance of broadband Internet access services at any given time, including during periods of peak usage”. Some of these suggestions are strong medicine, although for now, merely a starting point for discussion.

Going mobile

With respect to mobile wireless platforms, the FCC has asked in the NPRM how, to what extent, and when openness principles should be applied. Again, the Commission is concerned about furthering innovation, private investment and competition in the industry. In the most recent inquiry, the Commission seeks to update the record on these questions in light of intervening developments.

The two intervening developments that appear most significant to the Commission are: (1) the Verizon-Google Proposal suggestion that wireless broadband be exempt from all Open Internet principles other than transparency; and (2) the recent rise of wireless pricing plans based on the amount of data the customer uses. The latter, in particular, raises a serious question.

In essence the issue boils down to this. The need for “network management” – i.e,, blocking or slowing traffic – generally increases to the degree that network traffic approaches or exceeds network capacity. If usage-based pricing reduces congestion on wireless networks, will wireless operators have less need to use traffic management techniques that trigger open Internet issues?  

The latest inquiry raises far-reaching questions, and poses potential solutions, that are likely to generate considerable debate. Look for a further influx of commentary, for and against, as the deadline for comments approaches. (As of this writing the deadlines for comments and replies have not been established. Check back to www.commlawblog.com for updates.)

There is two additional intriguing aspects of the latest inquiry (and Chairman Genachowski’s separate statement in support of it). First, according to the notice, the “discussion” triggered by the Open Internet proceeding “appears to have narrowed disagreement on many of the key elements of the framework proposed in the NPRM”. Genachowski’s statement strikes a similar note. It is, of course, impossible to say for sure whether that gloss on the on-going deliberations is accurate. Certainly the Chairman would prefer it to be so. The response to the most recent may or may not tell a different story. 

Second, whenever the comment and reply deadlines happen to be set, the window for replies will not close before November 2. . . which happens to be Election Day. That means that the conclusion of the Open Internet proceeding, once expected by some to be set for September, will not happen before the upcoming election. In view of the high profile the issue of Network Neutrality has had on Capitol Hill – it’s probably no accident that Verizon and Google titled their magnum opus a “Legislative” proposal – an intervening election could have a significant impact on the fate of the Open Internet proceeding.   We shall see.

Wasn’t it the Obama administration in general – and FCC Chairman Genachowski’s administration in particular – that arrived on the scene with a much ballyhooed promise of “transparency” in government decision-making?   Gone would be the bad old days of a few corporate giants sitting around with policy-makers in smoke-filled rooms deciding the fate of the rest of the industry and, for that matter, of the general public, whose representatives were offered no seat at the table.   Some tentative steps in that direction were taken, including announcing the expected agenda items at upcoming meetings far enough in advance to permit everyone to have a last word with the Commission before the cone of silence imposed by the Sunshine Act descends. A new spirit of openness seemed to be stirring over the waters of FCC policy-making.

Perhaps that is why the recent backdoor maneuvering seems like such a betrayal.   To be sure, the GSA’s “no smoking” policy ensured that industry titans would have to leave their Macanudos and Cohibas smoldering outside in their idling limos while they met with “senior FCC officials”. And these days mineral water and acai juice are more likely to be on the beverage bar than rye and sour mash. So a lot of the fun, not to mention the smoke, has been drained from smoke-filled rooms. 

But the essence of a smoke-filled room – the private, closed door, invitation-only, giant corporation-only session with high ranking policy-makers – certainly remains. The conception that something as important as Net Neutrality (with huge implications for the privacy of the American people), the development and growth of the Internet, and the expansion of broadband access could be hashed out by a few corporations over corned beef sandwiches with no involvement whatsoever from the rest of the world is appalling. It is everything that the Obama administration claimed to reject about politics-as-usual. 

The FCC needs to do some damage control – and fast – to reassure people that in its quest to deliver universal broadband to strengthen our economy and democracy, it is not trampling on the very principles that it seeks to further.   An open decision-making process cannot take short cuts and can thrive only in the full light of day.

The Digital Revolution

Flash forward to the future.  All phone calls are digital. The telephone handset converts the voice signal to a series of 0s and 1s, compresses the bit stream, separates it into “packets,” and gives each packet an address corresponding to the phone at the other end of the call. Traveling through the Internet, re-routed on the fly, the phone-call packets share the wires with packets carrying other people’s phone calls, video, web pages, baby photos, and everything else we send to each other. Among the different “protocols” for setting up and addressing packets, the ones used on the Internet are collectively called “Internet protocol” or IP. The term VoIP refers to the combined technologies for carrying voice over IP – that is, over the Internet.

Most home and office phones are still tied to circuit-switched local telephone systems. But companies like Skype and Vonage offer VoIP service to end users. People who receive their telephone service through a cable company or FIOS also use VoIP, although that fact is often buried deep in the service agreement.

VoIP has important advantages over the circuit-switched phone system. Because voice packets make up only a tiny fraction of the data carried on the Internet, transport is cheap. Calls within the U.S. are usually free; calls to other countries cost just a few cents per minute. Phone numbers need not be tied to location, so a subscriber in Sydney, Australia can have a local New York City phone number, which further lowers costs for calling friends in New York (although a careless pizza order can be expensive).

In contrast to its historically tight regulation of the circuit-switched phone system, the FCC takes a mostly hands-off approach to data services. Its Computer II decision of 1976, well before the public Internet emerged, kept full common carrier regulation over phone lines, even if they also carry data, but forbore from regulating the data. The Internet is squarely in the data category; and true to its word, the FCC left the Internet alone. As broadband emerged, the FCC largely deregulated the facilities that carry it, which left only voice lines under traditional regulation.

The voice vs. data distinction worked well for thirty years. But then came VoIP, which straddles both categories. Early VoIP drew little attention from regulators, being just another novel Internet application that needed special equipment and skills. But the technology soon moved beyond the hobbyists and into the mainstream. A subscriber could place and receive calls using what looked like an ordinary phone, yet bypass most of the phone system. Passing the ultimate test for ease of use, VoIP became popular with many elderly people as an inexpensive way of keeping in touch with far-flung family members.

The FCC, which dislikes ontological questions, had to make a decision: Is VoIP a voice service or non-regulated data service? It depends, was the answer. VoIP that meets the “quacks-like-a-duck” test – that functions as a full substitute for traditional voice service, phone number and all – became subject to many of the same rules as traditional voice: 911 call location, access by disabled users, law enforcement wiretaps, privacy rules, and universal service payments. Yet even quacks-like-a-duck VoIP remains exempt from most common carrier regulation.

The result is two phone systems operating side by side. The old one is ubiquitous and regulated. The new one is used only by some people, though more every day, and is mostly outside FCC regulation. The two interconnect to the extent that VoIP users can place and receive calls to and from ordinary phones. But not to and from everywhere – a point that has now caught the attention of the FCC, and could portend major changes in phone service.

Google Voice Muddies the Waters

VoIP can do a lot more than just emulate a circuit-switched phone. A service called Google Voice shows some of the possibilities. We described GV in an earlier post: A subscriber receives a new phone number, local in a region of the subscriber’s choosing.  Calling that number rings all the customer’s phones, wherever they are:  office, home, cell, etc.  Different callers can be automatically routed to different phones, or forwarded selectively to still other phones, or fed different voice mail greetings, or given different rings, or blocked altogether.  All the voice mails from all the phones end up in one place, where they can be read in printed form, like emails, or listened to online from anywhere. There are provisions for setting up conference calls, and for recording phone conversations for online storage. And all of this is free.

Except for the final connection to a dialed or forwarded number, the whole thing runs over the Internet.

But although the Internet goes everywhere, Google Voice does not. Complaints from many quarters allege that GV refuses to place or forward calls to certain rural areas. Google has said why: it wants to avoid paying high access charges. Of course, a GV subscriber whose Aunt Mildred lives in South Succotash still has the option of reaching her with an ordinary call over the old system. But GV’s form of discrimination, if true, raises one of those awkward definitional problems that so trouble the FCC. A common carrier is not allowed to pick and choose among call destinations. It must connect wherever the customer dials. A non-common-carrier is not subject to that obligation.

In response to the complaints, the FCC sent a letter to Google that asks two kinds of questions. First is whether and how Google Voice restricts calls to certain phone numbers or groups of numbers. Translation: do you discriminate? Second and third are whether GV charges end users for its services (no), and whether use of GV is by invitation only, as Google claims (yes, although anyone can request an invitation.)

Why do those last two questions matter? The law defines a common carrier service as one that is (1) offered for a fee (2) directly to the public, or to enough people that they effectively constitute the public. If the FCC finds the GV service is “for a fee” and offered to the public (or most of it), then GV is a common carrier service and must connect to anywhere. But that is not a likely outcome. Google no doubt will make a case that the service is truly free of charge, and truly limited in availability. The FCC would then have to agree it is not a common carrier service. The non-discrimination rules would not apply, and GV could decline those South Succotash calls with impunity.

Rewriting the Rules

The problem is, if we care about the future of the phone system, then none of Google’s possible answers yields a good result. The incongruity of shoehorning an IP service into circuit-switched regulation suggests that technology may have outpaced the rules.

Suppose the FCC, against all odds, decides that GV is a common carrier service. It must then require GV to provide rural connections and pay rural access charges. Google’s business model collapses and, with it, any incentive to continue offering the service. The public loses out. 

On the other hand, if GV is held not to be common carriage, then companies like Google can skim off the cheap, easy-to-provide services, and leave the more expensive ones, like rural voice calls, to the traditional regulated phone companies. As urban and suburban customers abandon their old phone service for cheap VoIP, the regulated companies’ revenues will fall. Some of their costs, like system maintenance and universal service fees, will become an increasing percentage of revenues. Local phone rates will climb, driving more users to VoIP alternatives, in a self-reinforcing spiral. That is also a poor outcome.

Unfortunately, the only lasting solution entails an overhaul of the Communications Act. When we last did that, in 1996, it was not fun for anyone, and we’re sure not looking forward to it now. But with the engineers having gotten out ahead of the lawyers, the lawyers have little choice but to catch up.

The rough outlines of a reform are easy to foresee. In the past, Congress and the FCC have eased regulation of services that became subject to competition. In the long-distance market, for example, the AT&T divesture of 1984 and a pro-competitive FCC created dozens of players. (We know exactly how many, because they all used to phone us at dinnertime.) Cell phone service has always been at least minimally competitive, and has never been subject to traditional telephone regulation.

Sooner or later, Congress will have to find that IP-based services like cable-provided VoIP, Skype, and Google Voice are giving local phone companies enough competition to justify some easing of traditional regulation. Rates, including access charges, will become more responsive to market forces. That will make traffic pumping less profitable, but will also give companies like Google Voice less expensive access to more areas of the country. Traditional phone companies will also benefit, in some respects, as they will find it easier to offer innovative services of their own.

But across-the-board deregulation will not do the trick. We still need a “carrier of last resort” to serve the always-expensive rural customers. We still need ways to subsidize service to high-cost areas and low-income subscribers, and to assist hearing- and speech-impaired users. The states, too, have a large role in regulating local telephone service, and may rank various outcomes differently. A largely rural state with a widely-dispersed population may have priorities at odds with those of a more urban state having dense concentrations of sophisticated users.

In short, VoIP and its spin-offs promise better, more flexible, and less expensive phone service. For the time being, though, these benefits will not reach everyone. Circuit-switching, although technologically obsolete, is everywhere, and will remain part of the telephone system for decades to come. The goal is a regulatory system that lets each deliver what it does best, keeping the enormous efficiencies and flexibilities of VoIP without sacrificing the ubiquity of the circuit-switched system.

Getting all the moving parts to mesh correctly will be an enormous undertaking. But we think the tone of the FCC’s letter to Google signals an awareness that the present regulatory scheme will not hold up. Inequities and discrepancies of the kind now presented by Google Voice will only get worse. Occasional patches might keep the current regime working a little longer, but an overhaul is coming due.

The FCC’s letter may be the first important regulatory step in that process.

The Wall Street Journal on December 15 accused Google of changing sides on network neutrality.

Other companies, though, have long collocated servers with the broadband providers. Akamai, Limelight, and others provide a service called “edge caching,” which locally replicates the content of major websites to allow faster access by subscribers. The practice is not widely seen as a violation of network neutrality. For one thing, edge caching is necessary to prevent logjams at popular sites. Today’s Internet could not function without it. For another, the companies that provide edge caching compete with each other in signing up as many websites as possible. Smaller websites work fine without edge caching; larger ones have a choice of providers. Everyone benefits.

Google now says it only wants to do what Akamai and Limelight and the others have been doing for years.

But there is a big difference. Akamai et al. do not manufacture their own content. They merely help with distributing the content of others. Google, with its search engine, YouTube, and dozens of other operations is probably the biggest single content provider on the planet. It has no incentive to share its edge servers with competitors. Despite the company’s protestations, Google’s plans seem to exemplify a fundamental violation of network neutrality.

So far, though, Google faces no legal barriers. Nothing in the FCC rulebook prohibits discrimination in broadband Internet services. The FCC did announce four Internet “principles” back in 2005, but never adopted them as rules, so they may not be enforceable. (We will find out, now that Comcast has appealed the FCC’s use of the principles in ordering it to alter network management practices.) But even then, Google’s plan may not amount to a violation. The most relevant principle just entitles consumers to “competition among network providers, application and service providers, and content providers.” So long as Google is not the only provider on the block, it may be able to discriminate with impunity.

But it can’t do that and still keep saying it supports network neutrality.