Yesterday Google officially announced Project Fi, its much anticipated wireless service, which I’ve previously blogged and tweeted about during its pre-announcement rumor/leak phase. Now that more details, including pricing, are available directly from Google, an updated post seems in order, especially following recent posts about newly launched municipal Wi-Fi services in NYC and Boston (later on this post I’ll consider how these two developments may be related and synergistic).
As expected, Google’s wireless service will route user traffic over a mix of Wi-Fi connections and, via MVNO arrangements with Sprint and T-Mobile, the two carriers’ cellular networks. This “three network” approach alone makes the service pretty unique. In a blog post yesterday, VP of Communications Products Nick Fox explains:
As you go about your day, Project Fi automatically connects you to more than a million free, open Wi-Fi hotspots we’ve verified as fast and reliable. Once you’re connected, we help secure your data through encryption. When you’re not on Wi-Fi, we move you between whichever of our partner networks is delivering the fastest speed, so you get 4G LTE in more places…If you leave an area of Wi-Fi coverage, your call will seamlessly transition from Wi-Fi to cell networks so your conversation doesn’t skip a beat.
The Project Fi FAQ page explains further that its “software is optimized to not put extra strain on your battery by only moving you between networks when absolutely necessary.”
Also as expected, Project Fi’s multiple-network functionality will initially only be available on Nexus 6 smartphones, via a special SIM card. The device costs $649-$699, depending on its storage capacity, with a no-interest, no-fee option to pay for it over 24 months (an approach similar to what most cellular carriers now offer to their customers). Under this plan, the monthly cost for the device is $27.04-$29.12.
Though the multi-network functionality will initially only be available on the Nexus 6, Google aims to make it easier to switch not only between networks, but also between devices. As its Project Fi announcement explains:
Talk, text, and check voicemail with the screen nearest you. Your phone number now works with more than just your phone. Connect any device that supports Google Hangouts (Android, iOS, Windows, Mac, or Chromebook) to your number. Then, talk and text with anyone—it doesn’t matter what device they’re using.
While much of the above was previously confirmed via rumors and leaks, I hadn’t seen anything about Project Fi pricing until Google officially announced the service, which has a pretty simple fee structure.
For a base price of $20 per month, Project Fi customers receive the following:
The above graphic from arstechnica illustrates that, although Google’s pricing becomes less competitive the higher the monthly data allowance, it is the least expensive service for a 1 GB tier of service.
So far, this series of blog posts has focused on what private companies are doing in the unlicensed spectrum space—including startups, cable operators, Google and mobile carriers. In the next few posts I’ll consider what cities and local neighborhoods have done, are doing, and are planning to do with unlicensed spectrum.
As a first step in considering current and future “community WiFi” projects, it’s worth taking a look back at an earlier wave of municipal WiFi networks.
These date back roughly a decade, with one of the most ambitious early projects, in Philadelphia, being announced in April 2005, just a few months before the first iPhone was shipped. The Philly project was one of multiple urban deployments involving Earthlink which, at that time, was grappling with the transition from dial-up to broadband as the dominant form of Internet connectivity. Frustrated with its limited and not-very-profitable access to cable and DSL networks (which weren’t subject to the same network-sharing requirements that applied to dial-up service), Earthlink viewed these muni-WiFi projects as a way to offer Internet access services independent of wholesale arrangements with cable and telco network operators that were not only its main competitors, but increasingly refused to offer Earthlink and other dial-up ISPs access to their networks.
The problem with this effort to use WiFi as a competitive “bypass” technology was that the public-private partnership model embraced by Earthlink, its city partners and similar ventures, was flawed in multiple respects. Though some projects were relatively successful (the largest one probably being the Minneapolis network operated by USI Wireless) Earthlink eventually abandoned its WiFi ambitions after launching several high-profile projects in major U.S. cities. Other players, including MetroFi, which launched several networks in the Bay Area and in Portland, also exited the business around the same time.
Among the problems faced by these early network deployments was that they attempted to offer a service that could compete with wireline broadband services, at least at the low end of the market. But the inability of WiFi signals to reliably penetrate walls made in-home service a serious challenge, especially for the earlier generations of equipment used in these networks.
Another issue was that, ten years ago (as the June 2005 iPhone launch date suggests), there was nothing comparable to today’s nearly insatiable demand for nomadic (but not necessarily ubiquitous) outdoor Internet connectivity. Yes, there were plenty of laptops being lugged around and used in coffee houses, etc., but today’s increasingly universal presence and intensive use of high-performance WiFi-capable smartphones and tablets was, at that time, nothing more than a twinkle in Steve Jobs’ visionary eye.
As discussed in prior posts, today’s dramatically increased demand for nomadic Internet connectivity is spurring a range of efforts by private service providers, including startups, cable operators, Google and cellular providers (as well as restaurants, cafes and other venues providing WiFi hot spots as a customer amenity) to satisfy that demand.
At the same time, some municipalities and neighborhood groups are discovering unmet needs and exploring ways to address them. These efforts will be the focus of the next few posts in this series.
In a post yesterday I discussed the disruptive potential of Google’s Project Nova. Having just discovered an article by Christopher Williams published last weekend in the UK’s Telegraph, I thought I should add an update on international aspects of Nova’s ambitions and potential impacts.
Williams reports that, according to industry sources, “Google is in talks towards a deal with Hutchison Whampoa, the owner of the mobile operator Three.” He also notes that “Google and Three declined to comment.”
The two giants are discussing a wholesale access agreement that would become an important part of Google’s planned attempt to shake-up the US mobile market with its own network. It is understood that Google aims to create a global network that will cost the same to use for calls, texts and data no matter where a customer is located. By linking up with Hutchison, it could gain wholesale access to mobile service in the UK, Ireland, Italy and several more countries where the Hong Kong conglomerate owns mobile networks. Sources said Hutchison was a natural partner for Google in the plan, because it has also sought to eliminate roaming charges for Three customers.
According to CNET:
Hutchison Whampoa would be a potentially powerful global partner to help Google cut roaming fees. It operates the UK’s Three network and is trying to acquire the UK’s O2 network from Telefonica. It also operates networks in Hong Kong, Macau, Indonesia, Vietnam, Sri Lanka, Italy, Sweden, Denmark, Austria and Ireland.
Though Google has not revealed much in the way of details, the Internet search giant is expected to launch a WiFi/MVNO wireless service sometime in the near future. Based on limited comments from company executives and reports in the Wall Street Journal (see here and here; subscription may be required) and elsewhere, it seems that the service will:
Speaking at the Mobile World Congress held March 2-5 in Barcelona Spain, Sundar Pichai, Google senior VP of products, appeared to downplay the scope and disruptive impact of the MVNO service, known internally as Project Nova.
As reported by TechCunch, Pichai said “We don’t intend to be a network operator at scale. We are actually working with carriers.” And according to Wired, Pichai also pointed out that “[c]arriers in the US are what powers most of our Android phones [and]…[t]hat model works really well for us.”
Google may, in fact, have limited ambitions for Project Nova. But I suspect the cautious nature of Pichai’s comments reflects a desire to avoid prematurely upsetting the industry’s dominant carriers—whose customers purchase huge numbers of Android devices—more than it does a lack of Google-scale ambition for Project Nova. And the fact that Google won’t be operating its own network and will be “working with carrier partners” doesn’t mean Nova doesn’t have potential to seriously disrupt the mobile industry’s status quo.
I expect Google to approach Nova the way it approaches most new product introductions: start small in “beta” mode, then adapt to market developments. Sometimes this leads to products being killed, revised or merged with others (see a partial list here), while at other times it leads to aggressive expansion, as was the case with the Android operating system.
While there are parallels between Project Nova and Google Fiber, the company’s investment in local fiber optic networks, there are also important differences that could translate into much faster growth potential for Nova. The key difference is that Project Nova doesn’t require Google to undertake the time-, labor- and dollar-intensive task of building fiber networks city by city and block by block. As a result, while Google Fiber is intended to be a profitable business and is gradually expanding to more cities, Project Nova could allow Google to move very quickly and relatively inexpensively to deploy a nationwide service using other companies’ physical networks.
Having read a mix of pre-launch speculation available online (see excerpts below), I’m inclined to believe that:
Below are excerpts from online commentary that have helped inform this point of view. As always, comments are welcome, especially from those who see things differently.
After writing two posts on potential carrier use of LTE technology in unlicensed spectrum (see here and here), I came across some information that helps clarify the functionality of and relationship between LTE Unlicensed (LTE-U) and License-Assisted Access (LAA). In those posts I referred to these as if they were different names for the same technology. A more accurate statement would be that:
1) LTE-U is an earlier iteration of “LTE in unlicensed spectrum” technology that conforms to the less stringent spectrum sharing requirements of countries like the U.S., South Korea, China and India;
2) LAA will provide a standardized technology that goes further than LTE-U by satisfying the more demanding spectrum-sharing requirements in other markets, including Europe and Japan.
The clarification comes courtesy of the Qualcomm web site, which provides a summary of LTE-U here, and of LAA here. As noted in an earlier post, Qualcomm is a leading advocate of carrier deployment of LTE in unlicensed spectrum, having introduced the idea in late 2013. Selected excerpts from both descriptions are below.
In the last post in this series I reviewed several different points of view regarding the pros and cons of cellular carriers using “LTE Unlicensed” (LTE-U) to expand their network capacity. In this post I’ll take a closer look at movement in this direction among U.S. carriers.
[Note: The deployment of LTE in unlicensed bands is referred to by multiple names, including “LTE Advanced in unlicensed spectrum,” “LTE Unlicensed” (LTE-U) and, most recently “Licensed Assisted Access” LTE (LAA). In this post I’ll refer to it as LTE-U, though other names may appear in some excerpts included in the post.]
The two U.S. cellular providers that have so far expressed most enthusiasm for LTE-U are Verizon and T-Mobile.
[B]ased on the discussions I’ve had this week, it appears that Verizon…, Vodafone and other carriers last year decided they wanted to make LTE-U a reality–and they decided they didn’t want to wait for the 3GPP to standardize the technology. So they teamed up with some network technology companies to design real-world tests of the technology…
Verizon clearly has high hopes for the tests and the technology–it has said that it plans to commercially deploy it in the 5 GHz and 3.5 GHz bands in 2016. Verizon is not the only carrier that supports LTE-U/LAA. T-Mobile announced this year that it too will deploy what it calls LAA in the 5 GHz band in 2016. T-Mobile CTO Neville Ray said he believes the carrier can get LAA-capable handsets this year.
As Dano notes, “[h]owever, not all carriers are on board.” Specifically, he points to comments from Tom Keathley, senior VP of wireless network architecture and design for AT&T. As one might expect from a carrier that has invested in a network of more than 30,000 WiFi hotspots, AT&T’s concerns include the risk that LTE-U deployments will not share unlicensed spectrum fairly and efficiently with WiFi.
Keathley said that current approaches to LTE-U are vague about how exactly to check for existing users in unlicensed bands, and how long LTE users can occupy unlicensed spectrum.
Dano also cites comments from Eric Parsons, an executive at Ericsson, a leading wireless network equipment vendor, regarding how these spectrum sharing issues might be dealt with in different regions of the world. As Parsons explains, “there are very specific guidelines in Europe and Japan that cover these areas, but countries like the United States don’t have specific guidelines.”
T-Mobile, which has less licensed spectrum to work with than its competitors (see here for T-Mobile CEO John Legere’s perspective on this issue), seems particularly interested in LTE-U. In anticipation of commercial deployments in 2016, it has announced plans for multiple tests of the technology, in cooperation with Alcatel-Lucent and Qualcomm, Ericsson and Nokia.
In a January 5, 2015 blog post T-Mobile chief technology officer Neville R. Ray shed some light on the company’s plans:
Currently, there is approximately 550 MHz of underutilized spectrum in the 5 GHz Unlicensed National Information Infrastructure (UNII) band, which is available for any use within the FCC’s rules for the UNII band. LAA is a new and innovative approach that allows for licensed and unlicensed spectrum to work seamlessly together. And, we’ve already begun work with our various chipset, radio infrastructure and device partners to bring LAA production trials to life this year and bring the technology to our customers in the near-future.
During T-Mobile’s February 19, 2015 yearend earnings call, Ray provided an update on the company’s LTE-U plans:
A key theme in this “unlicensed spectrum” series of blog posts has been the potential negative impacts on wireless carriers of lower-cost services built on WiFi connectivity, either in a “WiFi-first” or “WiFi-only” mode.
In this two-part post the focus will shift to potential LTE deployments in unlicensed spectrum by licensed carriers, as they seek to increase network capacity while retaining tighter integration with their existing LTE-based networks than they can achieve with WiFi technology.
The prospect of carriers deploying LTE in unlicensed bands marks a new phase in the history of unlicensed spectrum. In this new phase licensed carriers and their preferred technologies (e.g., LTE) could play a much bigger role in the unlicensed space, potentially disrupting the existing spectrum sharing model embodied in WiFi standards and familiar to users of WiFi technology.
The deployment of LTE in unlicensed bands is referred to by multiple names, including “LTE Advanced in unlicensed spectrum,” “LTE Unlicensed” (LTE-U) and, most recently “Licensed Assisted Access” LTE (LAA). In this post I’ll refer to it as LTE-U, though other names will appear in some excerpts included in the post.
Not surprisingly, there exists a fairly broad range of views on the balance of benefits and harms likely to occur from carrier deployment of LTE-U. As one might expect, Qualcomm, the wireless tech giant that first proposed the idea in late 2013, is enthusiastic. In a November 20, 2013 blog post, Prakash Sangam, Director, Technical Marketing summarized his company’s perspective:
Consider the length that operators are going to address increasing data traffic with small cells and utilizing all spectrum assets….Wouldn’t it be ideal for them to deploy small cells that support LTE not only in their regular licensed spectrum but also in unlicensed spectrum?…[I]nstead of managing two separate networks for licensed and unlicensed spectrum, and dealing with the complexities of interworking between them, they will have one unified network accomplishing the tightest possible interworking. How cool is that?
Okay, the operators are covered. What about the mere mortals like us, the users? Well, remember all the juggling between LTE and Wi-Fi networks; making sure you are connected, and connected to the right technology to get the best speed; worries about the media not seamlessly moving over between the networks, and tolerating video freezing, breaks, restarts etc.? All of that will be over with LTE Advanced in unlicensed spectrum…Because it’s one network, with an anchor in the highly reliable licensed band, you are always in safe hands. Add to that carrier aggregation, across licensed and unlicensed bands, and you, the user, get higher data rates and an enhanced broadband experience.
This is all good, but one natural question someone might ask (we asked it ourselves) is, “will it affect the Wi-Fi networks out there now?” Well, LTE Advanced in unlicensed spectrum has been carefully designed to protect Wi-Fi, so that both can co-exist harmoniously. So, when an operator switches from Wi-Fi (“carrier Wi-Fi” as it is called in the industry) [to] LTE in unlicensed, not only do LTE Advanced users in the unlicensed spectrum benefit but also, in many cases, the neighboring Wi-Fi users.
Moreover, LTE Advanced in unlicensed can be brought to fruition in countries such as the United States, Korea and China using the existing standards (Rel 10) and, of course, by leveraging the existing LTE core networks.
Given the cable industry’s growing enthusiasm for a WiFi-based wireless strategy (see here and here), it’s not surprising that they are less enthusiastic than Qualcomm about wireless carriers deploying LTE in unlicensed spectrum. In a May 21, 2014 post on the CableLabs blog, Ian MacMillan expressed some of their concerns: