Mobile musings and wireless wanderings

I have been talking with the press about the 700 MHz auction that is underway and thought I’d take the time to provide some context and expand as to the impact that the auction may have on mobile services in the next 3-5 years. I will start with a review of Google’s definitions of openness, then discuss the where, what, who and when of the auction. I will then try to put this in context of what this means for the future of open access and the underlying enablers for Mobile 2.0 services.

Four Types of Openness

For purposes of the following post related to the 700 MHz auction, there are four relevant types of openness as defined by Google: applications, devices, services and networks. Chris Sacca, the Head of Special Initiatives outlined these four types of openness on Google’s policy blog. They include:

• Open devices represent the ability to attach any device to the network. This means that the company operating the network (e.g., the network service provider) will allow any device to attach and utilize the network.
• Open applications include the ability to create and download applications and content to any device. As such, the company that owns the devices or operates the network (and likely even controls the distribution of the devices) will allow applications and content to run on the device.
• Open services provide the ability for other companies to utilize the networks on a wholesale basis. For example, this concept is akin to how MVNOs such as Virgin Mobile and Kajeet utilize Sprint with the further stipulation that the business terms are on a nondiscriminatory and reasonable commercial basis.
• Open networks allow third parties such as Internet service providers to interconnect at “any technical feasible point.”

This terminology proposed by Google is slightly confusing, e.g., it is not the applications that need to be open in order for users to be able to download and utilize applications on any devices - this is a function of the devices needing to be open. Similarly, what Google calls “open devices” really should be termed “open access” which connotes the idea of any device being able to utilize any network.

There are other aspects of openness that these four categories do not distinguish. I have discussed openness in more detail in a past post entitled Two by Six Degrees of Openness – Apple and Google’s impact on the mobile ecosystem.

The auction itself - who, what, where and when

Where
The auction includes a set of 181 licenses corresponding to various blocks of frequencies at the lower end of the spectrum (700 MHz) that has been used for the last 50+ years to transmit analog TV (UHF). The 700 MHz spectrum has been compared to prime beachfront property. The spectrum is likened to prime real estate based on the fact that it is based in the lower frequency range, which imparts special properties resulting from both basic physics and the way the licenses have been packaged by the FCC. Some special characteristics include (i) extended reach (and hence lower costs due to towers being spaced further apart than networks operating at higher frequencies), (ii) better coverage inside of buildings, and (iii) the possibility of an entire nationwide block from which a truly national wireless broadband network could be deployed.

What
The licenses include a packaging of the spectrum into blocks that are specified both in terms of their frequency range but also in terms of how the blocks are to be used and under what conditions. The conditions include both a deployment time frame (i.e. how rapidly the licensor will need to build out a service offering based on the respective blocks they have won) as well as stipulations related to the relative openness of the resulting services.

The stipulations related to the relative degree of openness can be thought of as weak provisions. Of the four types of openness defined above, they only include two: open access and open services. These weak provisions specify that whoever controls the spectrum (i.e., the winners of the licenses in the auction) cannot control how the network will be utilized and cannot control what kind of devices can be attached to the network.

In short, given these stipulations, service offerings resulting from the ensuing auction should be some type of generic wireless broadband networks (e.g., in contrast to an application-specific wireless broadcast TV network akin to Qualcomm’s MediaFlo) from which a variety of devices can connect to the network. The FCC could have gone farther and required that the service providers enable wholesale access to the network as we witnessed during the telecommunications deregulation. This would have likely given rise to a wider range of service providers akin to the various mobile virtual network operators (MVNOs) that we see today.

The available licenses can be grouped into three parts:

1. The C block (i.e., the potential nationwide block) includes a total of 22 MHz of spectrum (2 x 11 MHz), which must be built out under the open access service provisions as well as a relatively aggressive deployment timeline (40% of the build out must occur in the first four years, 75% within ten years).
2. The D block includes a total of 10 MHz of spectrum reserved for public safety to be built out in partnership with private entities.
3. The remaining A, B, and E blocks are economic area licenses (each 12 MHz, 2 x 6 MHz) and cellular market area (CMA) licenses (6 MHz, unpaired).

Who
The bidders include carriers (e.g., AT&T, Verizon, Alltel), cable companies (Cox, Cablevision - but does not include Time Warner and Comcast), Qualcomm and a rather electric array of regional providers and various bidding consortiums. FierceWireless maintains an up to date list here.

When
The auction started yesterday. The winners will be announced in the next month or so. The analog TV broadcasts will be shut down in a year from now. Deployments of services in these blocks will range from starting in the next 1-2 years (e.g., regional WiMax offerings) to 3–5 years (e.g., LTE deployments).

Fast forward - auction results

Fortunately or unfortunately - depending on you outlook, the general consensus is that the auctions will unravel as follows:

1. AT&T and Verizon get the prime real estate and it will be business as usual as they continue the progression of upgrading their networks as we witnessed as they moved from 2.5 to 3G and now 4G.
2. AT&T and Verizon will become “open” but it is still assumed that their general attitude will be more in terms of the “letter of the law” than in spirit.
3. Some action will happen in a number of local, residential areas where local providers will win some residual licenses that AT&T and Verizon do not focus upon. This could result in some interesting and rapid rollout of WiMax networks.

Innovation at the edge - the case for modularity and openness

All the action has always been and will always be at the edge of networks. As economist Odlyzko has noted, successful services from an end-user perspective have come from the edge of the network: the Web itself, email, IM, search, social networks (Facebook) and user-generated content (YouTube). In contrast, services that have originated from the network provider have been underwhelming, albeit often robust and scaleable, in comparison: ATM, multicasting, WAP, 3G.

There are a number of interesting historical examples of how markets and innovation resulted from various forms of openness between the network, the carrier and the customer premise equipment providers. There is always a natural tension between fully integrated service offerings and modular marketplaces in which the network providers are distinct from the carriers and the equipment providers. Originally, the rail networks themselves were separate from the carriers and the equipment providers. Consumers and companies acquired their own carriage and operated them on top of the railroads. See Odlyzko for more background on railroads.

Clayton Christensen has done a good job at articulating the importance of modularity with respect to disintegration, and hence openness. Vertically integrated service providers have integrated, interdependent architectures. These types of architectures enable service providers to control the customer premises equipment (CPE), the network and the service offerings (e.g., walled gardens). There are benefits of having full vertical integration (e.g., as exemplified today by Softbank in Japan or in the past when IBM created the mainframe hardware, operating system and software). Christensen has made a strong case for modularity and the resulting innovation. For example, in 1968s, the FCC ruled to decouple (i.e., modularize) the network provider (AT&T) from the CPE provider (Western Electric), hence enabling any device to be attached to the edge of the network. This gave rise to innovation and created a number of new markets driven by new types of CPEs: the PBX, the modem and the Fax.

Jason Devitt founder of Vindigo and Skydeck, referenced this 1968 FCC ruling during his FCC testimonial, claiming that “mere Carterphone-style open access wouldn’t be good enough.”. Professor Tim Wu re-introduced the relevance of the Carterphone ruling to wireless in his paper “Wireless Carterfone” published in the International Journal of Communication, Vol. 1, 2007.

Fast forward - what impact will the auctions have?

Although, the likely build out by AT&T and Verizon will only include 2 of the 4 types of openness outlined above, we will see the impact of openness start to emerge this year and, unfortunately the nationwide 4G wireless networks will take 5+ years to deploy.

In the short term, we have already seen Verizon accelerate their open access agenda. Verizon with the announcement of their Open Development Conference for March of this year. Arguably, Google’s pressure on the FCC and the resulting stipulations accelerated Verizon’s agenda for opening up their network. I believe that various aspects of openness have been on their agenda for quite some time. In terms of applications, they have been constrained by BREW in terms of being able to offer “off deck” capabilities. Operators will increasingly look to off-deck digital content retailers to help drive their data ARPU.

Some constituents are still attaching their hopes that Google will auto-magically deliver a nationwide wireless broadband network that offers the unfettered, open access of the Internet. The model for such an ideal open network includes connecting end-users via their devices using digital bit pipes (i.e., ISPs) connected to backbone network providers and the world of content and services (see Dr. Stagg Newman’s chapter Broadband Deployment in The Broadband Explosion).

We are seeing some glimpses of the types of innovation you get when you have open access on open networks. For example, the innovation arising from the open spectrum (2.5 GHz and 5 GHz “WiFi”) used as an on ramp to the Internet can be witnessed with the various innovation happening in terms of both devices and service offerings, e.g., Apple iPod touch and Starbucks network access.

Of course, there is a “cost” to openness. We may complain about the various managed networks that have infiltrated our lives: TV networks, plain old telephone service (POTS), radio, and wireless phone networks. But, I would imagine that as individuals, we spend much less time getting these services (radio, TV) to work than we do with managing our applications, devices and access to the Internet. Odlyzko, in Telecom dogmas and spectrum allocations has noted that at first glance the Internet seems like a low-cost network but if you take into account the total cost of ownership (e.g., end-user administration of devices, IT and services), it is a much more costly proposition. As such, there are prices to pay for open networks: issues surrounding quality, security, cost of ownership.

Nevertheless, having a wireless broadband network (i.e., ubiquitous mobile broadband) is one of three key enablers for Mobile 2.0 services, along with having open access with frictionless distribution and affordable, unrestricted access to enabling software platforms, tools and technology.

The network as platform - an after thought

Having read The Big Switch by Nicholas Carr, it is clear that we are further along than I would have expected in computing becoming an actual utility. Here are some brief excerpts from this book to provide context:

“By the early 20th century, manufacturers didn’t have to be in the power-generation business anymore. They could run their machines with electric current generated in distant power plants by big utilities and fed to their factories over a network of wires.”

“By supplying electricity to many buyers from central generating stations, the utilities achieved economies of scale in power production that no individual factor could match.”

Electrification, just like computerization, led to complex, far-reaching, and often for individual companies and entire industries — and, as households began to connect to the grid, for all of society.”

“But electricity and computing share a special trait that makes them unique even among the relatively small set of general purpose technologies: they can both be delivered efficiently from a great distance over a network. Because they don’t have to be produced locally, they can achieve the scale economies of central supply.”

Combining the innovation we are seeing at the edge of the network with utility-based computing will transform the concept of network as platform well beyond what we have seen with Web 2.0 services.

From the vantage point of mobile services, what emerges when computing becomes a centralized utility will be a powerful cocktail combining the social web, mobility and a full do-it-yourself computing utility network. This is either the ultimate in openness or the beginning of the dark ages where we are held hostage by another utility network with prices driven on new forms of scarcity (silicon) or some form of monopoly on the suppliers of chips.

Special thanks to my friend and colleague Epiphany Vera for bouncing ideas back and forth and suggesting some enhancements to this post. In fact, he noted that the cost of fabrication has led many big companies to depend on very few providers like Taiwan Semiconductor which could result in cartel pricing or massive shortages due to events like the memory shortage in the 1990s … but I digress.