May 22, 2008
Technology Lessons from Municipal Wi-Fi 1.0
Jason Shannon is the Chief Technology Officer for Civitium, LLC. Prior to joining Civitium, he was the Lead Wireless Architect for EarthLink Municipal Networks, where he was responsible for the technical architecture of many of the largest broadband wireless initiatives to date, including Wireless Philadelphia, Anaheim and New Orleans.
In light of the recent announcements by EarthLink to discontinue operations of municipal Wi-Fi networks throughout the U.S., I thought it would be appropriate to share some of my insights from a technical perspective of Municipal Wi-Fi 1.0.
I've personally been involved in the design and optimization of more than a dozen active municipal Wi-Fi networks, and experience suggests that a dense, urban- scale municipal Wi-Fi network will likely fail to provide a universally available, technically viable, low-cost alternative to existing broadband services. This isn't to say that Wi-Fi doesn't have its place in the broadband ecosystem. In fact, it has become obvious that Wi-Fi will be a fundamental means of access for some time. With this in mind, it is important to know the limits of the technology, and what a city or service provider can expect from a large scale deployment.
So, what really were the technical issues with Municipal Wi-Fi 1.0? I've made a list of the six issues that are at the top of my list:
1. It is challenging to establish a reliable connection to an outdoor Wi-Fi access point using the low power client devices inherent in handheld, laptops and indoor customer premise equipment (CPEs). This challenge is due to the constrained power budgets associated with Wi-Fi radio equipment and the power output and emission mask limits established by the FCC. The typical uplink gain for municipal Wi-Fi systems is very low and does not withstand heavy loss factors due to exterior wall penetration or significant obstruction between transmitters.
2. An especially demanding use-case for municipal Wi-Fi where this problem is common is multi-dwelling units (MDUs), where the penetration through many densely constructed walls is required for communication between transmitters. Another example of where constrained link budgets limit the ability to establish reliable Wi-Fi connections is in high rise or mid rise buildings where: (1) construction is dense and highly absorbent of the 2.4GHz electromagnetic signal, (2) exterior metallic reflective windows reflect the electromagnetic signal, and (3) the higher elevation floors (often above the 4th or 5th floor) are outside of the vertical radiation pattern of the omni directional antenna.
3. We find that Wi-Fi is most reliable when client devices are within the main radiating lobe of the antenna and only minimal obstructions exist between the transmitters. A universally available service based on this operating assumption would require a massive node density in an urban environment, and would require a significant infrastructure of both outdoor and indoor mounted Wi-Fi access points. As we'll discuss in the following paragraphs, this massive node density presents other challenges outside of the obvious cost of infrastructure and operations.
4. The 2.4GHz spectrum used by Wi-Fi is an unlicensed frequency band, and is susceptible to interference from other radiators in the 2.4GHz band as defined by the FCC’s Part 15 regulations. An important factor in considering the availability of a Wi-Fi system is that transceivers are always susceptible from outside, non-cooperative interference. This interference will exhibit itself as complete loss of connection, or RF "jamming", or as significant degradation of data throughput. The intensity of the interference is dependent upon factors such as the power output of the interfering device, the proximity to the interfering device, and the baseline signal to noise ratio (SNR) of the Wi-Fi signal.
5. A dense, urban scale deployment of Wi-Fi transmitters is often susceptible to performance degradation due to the issue of media sharing. Wi-Fi systems operate across a total of 3 non-overlapping channels in the US and North America. In situations where a dense deployment of transmitters is needed, nodes operating on a given channel will have to share that channel with neighbor transmitters.
This issue is known as "mutual node visibility" and will cause system capacity to decrease by up to 1/n the number of mutually visible neighbors on the same channel, where n is the number of mutually visible nodes.
6. Municipal Wi-Fi systems require an infrastructure underlay to provide and distribute capacity throughout a network. Traditionally we find that this additional network tier leverages terrestrial fiber or some other unlicensed microwave frequency such as 5GHz. Both of these options present challenges. A terrestrial fiber infrastructure is often financially prohibitive to provide the number of physical connections required to support a dense, urban scale municipal Wi-Fi network. The use of 5GHz for capacity distribution requires the arduous process of identifying and acquiring key high sites throughout the City. The density of buildings and height differential in a downtown corridor often lead to situations where it is extremely difficult to identify viable high sites from a radio frequency (RF) propagation perspective. This is exacerbated by the fact that 5GHz is inherently a line of sight (or near line of sight) wavelength.
Many of the challenges noted above are specific to the common architecture used for municipal Wi-Fi 1.0, but many are inherent in the constraints of the technology itself. As next-generation, 2.0 models emerge, engineering approaches and deployment models (e.g. increased use of multi-radio solutions, organic ad-hoc networks and so on) may address some of these challenges, while at the same time introducing new ones that architects and engineers will have to resolve.
Posted by jason at 09:27 AM | Comments (0)
May 19, 2008
Civitium Analysis of Wireless Philadelphia Developments
On May 13, EarthLink Corporation issued a press release stating that it would discontinue its municipal Wi-Fi operations in Philadelphia. Due in part to Civitium’s past affiliation with the project, and our status as an advisor for public communications projects, we have received numerous inquiries from clients, partners and the press since last week. This memo seeks to provide our perspective on the factors that led to the breakup of the Philadelphia-EarthLink partnership and how these developments may be used to inform decision-making for cities involved in similar initiatives.
Posted by Greg at 08:45 AM | Comments (0)
May 10, 2008
On the Issue of Free
This is a post that I’ve wanted to write for a while. With the market correction in municipal Wi-Fi, the exiting stage left of EarthLink, and the recent announcement by Cablevision, my random thoughts and observations about the issue of free vs. paid finally took shape. So, here goes.
There are lots of industries and market segments that have grappled with “the economics of free” over time. This has certainly been the case in the music industry. In fact, most content-related industries are in some way continuing to sort through what should be free and what should be paid.
As the municipal Wi-Fi market has evolved, and especially since the correction that began in mid 2007, free vs. paid has continued to be one of its most debated issues. Plenty of consultants, analysts, bloggers and pundits have piled into this debate with a seemingly reasonable argument; that the correction was caused in large part by too much focus on free. I will take the risky position that 1) its failure was inevitable, and 2) the only way that large-scale, particularly urban, Wi-Fi will ultimately succeed is IF it is free.
Conventional Wisdom
There is something about the concept of free that elicits doubt and cynicism from anyone it’s presented to. Think about all the negative associations with the term free; nothing comes for free; you get what you pay for; wanting a free ride; freeloader.
This has been apparent in coverage of the municipal Wi-Fi market. An equipment provider in this article says "the [municipal Wi-Fi] market is very much in transition from that irrational exuberance around free WiFi, which just frankly is silly. The business model doesn’t work."
And in this article, an industry pundit says "the utopian world of free wireless everywhere was just silly." Editors note: I commit to readers of this blog that I will resist the urge to use the word “silly.”
Another of my favorites has been the argument that “nothing is free.” A research director at ABI was quoted in this article as saying “the shifting gears that you’re seeing is the realization that nothing is free. You can’t promise free Wi-Fi for everybody. That model doesn’t fly at all.”
Well, I can easily challenge that statement. Many things are indeed free, assuming that one accepts a definition of free as a product or service delivered without a discrete fee, and accepts that it’s still free even if something is gained by the provider, or offered of value by the consumer. The important issue is whether I, as a consumer, perceive something as free, and whether I am willing to accept the trade-off (e.g. my attention to view a non-organic advertisement when doing an Internet search, or my loyalty to stick with a provider because of a bundled service) that comes with using something that has no discrete fee attached to it.
Emerging Frameworks for Free
Fortunately, there have been some very smart people engaging in more critical and objective thinking about the economics of free, its history in other markets, how it might be applied to actually strengthen competitive positions and so on. Some of the best (and original) work in this area has been done by Mike Masnick over at Techdirt.
Masnick summarizes a great deal of his thinking, writing and interaction with his readers in this post entitled The Grand Unified Theory On The Economics Of Free. For anyone interested in the dynamics of free vs. paid, this is an absolute must read. Much of Masnick’s work centers on the digital media space, specifically on the love-hate relationship that the music industry seems to have with the transition from CDs to digital music. So, it gets into the implications of free on copyrights, IP protection, DRM and so on, but it still provides great insight for thinking about broadband access.
And of course, Chris Anderson, Editor in Chief at Wired Magazine, and author of the book The Long Tail, has a new book on the drawing board appropriately entitled Free. Anderson tackled the issue of scarcity vs. abundance in The Long Tail, and his work on Free seems to be a natural extension to dig into the infinite abundance that Free represents in a market.
Anderson proposes free as a legitimate business strategy in saying that “when the marginal cost of producing something trends to zero, the smart thing to do is to treat it as zero and get ahead of the competition: give it away for free in order to sell something else."
Can Access/Distribution Fit into this Framework?
While Masnick and Anderson’s work is fascinating, and they lay out some great frameworks to think about what free is, what it means, how it can be applied as a business strategy, etc., I find that neither really tackles the issue of whether these frameworks can/will apply only to digital content and media (inherently abundant resources, once they are composed/conceived), or whether it can actually be applied to the distribution of content and media; the underlying access networks, and in this context, Wi-Fi. Anderson does reference in this Nokia speech something that comes close to this. He refers to the common tactic by mobile operators of “giving away the phone to sell the minutes.”
The build-up of the municipal Wi-Fi market, and its subsequent meltdown, provides us with an excellent case study to test out, and think through Masnick and Anderson’s theories. But first, we need to take a brief trip back in time.
Attacked by Free – The Netscape Story
Remember the Netscape story? Emerging in the early 1990s, Netscape became the largest IPO in history, an IPO that was viewed as a watershed event of the Internet era. As you recall, Netscape’s business model was heavily, if not entirely, dependent on consumers paying for a licensed copy of Netscape Navigator, the company’s flagship browser software.
What does Netscape have to do with Wi-Fi though? Well, I find important similarities between Wi-Fi and the Internet browser, at least from the standpoint of how their markets evolved, or are evolving. Sure, one is a software interface to content and the other is a network access technology, but important similarities exist nonetheless.
- Ubiquity - The browser is ubiquitous. How many devices can you think of that do not have a browser embedded when purchased? This is of course true with Wi-Fi.
- Consumer expectation - Consumers expect browser software to be free. Who could imagine pulling out a credit card to pay for a download of Firefox? This is also true with Wi-Fi access in many settings.
- Standardization - Browsers, or more accurately their content handling and APIs are highly standardized. Sure there are nuances with Java support, scripting and other things, but for the most part, browsers owe much of their success to being standards-based. Once again, this is true with the family of Wi-Fi access standards.
- Consumer reliance - Consumers couldn’t function without a browser. Who can imagine going through an average day and not opening their browser? Again, at least in a home setting, and increasingly in an office setting, one could argue that this is true with Wi-Fi.
If you buy the basic similarities, the next step is to look into what this might tell us about the future of Wi-Fi, and especially its latest (and most uncertain) domain, municipal Wi-Fi. Going back to the evolution of Netscape, it appeared that by 1995, Microsoft perceived Netscape as a threat to its non-Internet-enabled Windows franchise. Microsoft got serious about integrating Internet standards across their product lines, and more importantly, decided to make Internet Explorer free for download. You may remember during the USA vs. Microsoft antitrust case where it was alleged that Microsoft “made a deliberate move to ‘cut off the oxygen supply’ to Netscape” by giving away Internet Explorer.
Sound familiar? Well, if you replace EarthLink with Netscape it might. Like Netscape’s dependence on browser licensing revenues, EarthLink’s entry into municipal Wi-Fi was almost entirely dependent on access fees, especially given that it was mostly unsuccessful at getting revenue assurances from its municipal partners. And as industry veteran Marty Hahnfeld points out on his blog, since EarthLink had a relatively narrow portfolio of other products to sell, the idea of bundling Wi-Fi as a differentiator to improve the value of other core products, or to reduce their churn, or to pull customers away from competitors, was really not of much use.
Combine this with the fact that promotional prices for DSL were being introduced. Add to this that few barriers to competitive entry were possible for EarthLink - since the networks they were building used unlicensed spectrum and standards-based (some would argue, commodity-based) equipment. And one more factor may have sealed the fate; the ubiquity of Wi-Fi on client devices ended up being of not much value in a business model like this, since residential CPEs were still costly. It didn’t matter how many residential customers you could get, since the cost of acquiring those customers exceeded their lifetime value (due to the low selling price of the service).
In short, at the time EarthLink was trying to introduce Wi-Fi as a “good enough” alternative to disrupt DSL and cable, other alternatives were being introduced - many of which were free - that were “good enough” to prevent its paid access networks from even getting close to a critical mass of paying subscribers. So, in our case study of the Netscape of Broadband, there was not a single attack from a Microsoft, but rather a whole range of industry dynamics combining to “cut off the oxygen supply.” A visit from Clayton Christensen or re-reading The Innovator’s Dilemma might have been in order, but that’s all water under the bridge by now it seems.
Wi-Fi Wants to be Free
Now where are we? Well, since the market correction, we have seen the emergence of organic Wi-Fi solutions and models like FON, Meraki and even Open-Mesh; we have seen AT&T open its nationwide hotspot network for free to AT&T customers; we have seen AT&T again offer free Wi-Fi access to its iPhone customers at Starbucks and beyond; and most recently, the big move; Cablevision announces that it will commit $330 million to deploy mesh Wi-Fi over the next two years. And as a side-note, Cablevision’s move fits perfectly within the frameworks from Masnick and Anderson, since 1) access will be free for existing customers and 2) its value will be in the differentiation it brings to their core products. $330 million sounds like a big number to me, but apparently it’s “a marginal cost nearing zero” on a cost per home basis to them.
While we don’t know if Wi-Fi will ultimately succeed on a municipal scale (specifically in dense urban markets; it’s already doing just fine in mid-American cities), what we do know is that larger-scale private investments in dense urban Wi-Fi environments, with the primary purpose to serve commercial needs, will likely re-emerge.
The Implications for Suppliers
The above position is counter to almost everything you will read (much of it from the vendor community, and the blogs they advertise on) about how “municipal governments must commit anchor tenancy and/or fund public-owned Wi-Fi networks” in order for large-scale Wi-Fi networks to succeed. When I read those things, I often assume the author is trying to defend something that actually doesn’t need to be defended (the mid/long-term prospects for municipal Wi-Fi), or knee-jerk-reacting in the hopes of convincing someone to buy radios that are stuck on the shelves. I find myself wanting to say to them “this is a marathon, not a 100 yard dash!”
Having a level-headed view of where the industry is going is critical for vendors and suppliers. With all the mixed signals that exist in this market, that’s an extremely difficult thing to accomplish. The implications of knee-jerking are likely to be the most damaging to the vendor community. For example, imagine an equipment provider who perceives the operator market for municipal Wi-Fi is dead, and advances a position that “municipal ownership and Wi-Fi as an extension for government IT” are the only opportunities to create value in the market with their solutions. Those positions can put them at a substantial disadvantage if/when operator investments reemerge.
On the other hand, vendors who “swing for the fence” on large operator deals, at a time when the only credible buyers are small municipalities, can find themselves out of cash and time. Once again, being able to see just over the horizon is the only way to mitigate these risks, strike the right balance of message, and achieve the right focus at the right time.
The Implications for Major Cities
Before I continue, a couple of disclaimers are necessary. 1) I prefer to stay out of the ideological debate over ownership and control of these or other networks, and 2) I have no pre-conceived policy positions to advance; only an interest in my client’s goals and an objective view on the evolution of this market.
I believe that when its history is written, municipal Wi-Fi will be chalked up as one of the most amazing examples for how local government intervention and action changed the fabric of the broadband market. Follow the chain of events from early 2004, to the Philadelphia project, to hundred of projects being launched in mid-America, to the swell of major city initiatives, to EarthLink’s rise and fall in the market, to the inflows and drying up of venture capital to equipment, software and service providers, to Apple’s integration of Wi-Fi into the iPhone, to AT&T’s renewed energy for its hotspot networks, to Cablevision’s announcement yesterday. What some may call failures in different parts of this chain, I would call building blocks.
In this context, the number of Wi-Fi radios installed in Metropolis, USA becomes largely irrelevant to the overall market’s evolution. While you can pick out mistakes and failures along that entire chain of events, that’s really missing the forest for the trees. I believe that major cities understand the impact of their involvement in this area of the past few years, despite whether the outcome was all they had hoped for. And now, with Wi-Fi having a mind of its own, I believe they are just getting warmed up in terms of their influence on the broadband market in America.
Implications to National Policy
I propose that there are two lessons resulting from the past few years of intense activity, and that these lessons should inform national broadband policy.
First, that municipalities can and do play a vital role in strengthening the broadband market. Their grass-roots, intense experimentation with Wi-Fi was a critical link in the chain above, and one could argue it was in fact the genesis for the deployment of new facilities and services (by .com, .gov AND .org) that may be reemerging. Even massive failures and reversals of direction in a market can create “shoulders of giants” for others to stand on.
Second, that unlicensed spectrum was the underlying force that allowed this to happen.
While Congress and the FCC apply scarcity economics to stuff the U.S. Treasury with spectrum auction proceeds, they would be wise to also understand and begin applying abundance economics. How could this be done? Well, in addition to getting serious about opening the white spaces spectrum, how about allocating a sliver of the proceeds from the recent 700 MHz auction as R&D credits for smart radio technology? Wouldn’t that be a way for the FCC to have its cake (generate the required funds for the auction) and eat it too (plant the seeds for future competition, promote more efficient use of spectrum, and mitigate the consolidation of power by incumbents).
And finally, the implications of all this for competitive access providers can be summarized in one sentence. Trying to sell Netscape Navigator licenses in the Firefox era might be, well… just silly.
Posted by Greg at 10:11 AM | Comments (0)