Disruptive Tendencies

AT&T Wi-Fi: How long to a billion?

In October 2007, AT&T acquired Wayport for $275m, and its national footprint of 20,000 Hot Spots in such high-profile locations as Starbucks and McDonalds. For many of us in the industry, this was a seminal event, as it was the first time that a major licensed spectrum holder embraced Wi-Fi (and unlicensed spectrum) as a major piece of its service offering.

Since then, AT&T has continued to be the poster child for its use of Wi-Fi for 3G data offload in areas of cellular congestion. Generally speaking, these locations can be described as areas where people eat, shop and congregate, and includes not only traditional Hot Spot locations, but also high capacity venues, college campuses, special events, and high-foot traffic urban centres.

Today’s announcement of 106.9m connections shows impressive, historic year-over-year growth, but it’s interesting to project forward.

• With over 90m subscribers, this number translates into 1.2 connections/subscriber/quarter.
• With around 30m smartphone users, this number translates into ~1.1 connection/smartphone user/month.

And keep in mind that, with free Wi-Fi at venues like Starbucks, some of these connections are from non-AT&T subscribers, so the number of connections/subscriber is going to be lower.

The following chart shows the number of AT&T Wi-Fi connections, as reported over the past 11 quarters. It’s interesting to note that AT&T has been able to materially impact the slope of the curve. When AT&T and Apple enabled seamless, transparent auto log-on for the iPhone in July 2008, the slope clearly changed. In July 2010, AT&T opened up its Starbucks locations to free Wi-Fi, the slope changed once again.

So how big can it get?

Clearly, one can imagine an AT&T quarterly report touting 1 billion connections (PREDICTION: Q3/2011), a 9X increase from today.

IMO, the biggest obstacle is subscriber mindshare and user intervention. People walk into a Starbucks today, and they know there’s Wi-Fi, so they’ll turn it on. But unless they know there’s Wi-Fi in the football stadium, rocking out at the concert, or walking around Times Square, users are far less likely to have their Wi-Fi radios turned on.

Maybe iOS v5 will be able to sense (or know via GPS) that there is an AT&T SSID nearby, and automatically turn on Wi-Fi, removing user intervention from the Wi-Fi equation totally. Now that would be cool, and welcomed by service providers the world over.

Wi-Fi for service providers? Onwards, and upwards!

One billion connections? Here we come!

 

That’s my .02!

Martin Suter

 

Spectrum is Like Beachfront

(Originally posted at http://www.von.com/blogs/guest/2009/12/spectrum-is-like-beachfront.aspx)

Plato is credited with having noted that, “Beauty is in the eye of the beholder.” This is certainly true in any discussion about spectrum. As with beachfront property, the thing about spectrum is that they’re not making any more of it. Recently, Blair Levin, who heads the FCC’s broadband program, acknowledged, “There’s not enough of it,” referring to spectrum. It was also noted that “several Tier I carriers are arguing for more licensed spectrum to keep pace with customers’ bandwidth demands, especially for mobile broadband.”

The FCC’s 700MHz auction raised $19.6 billion, with $16 billion coming from just 2 operators: Verizon and AT&T (National coverage maps). While I’m sure the FCC would love to run more multibillion-dollar spectrum auctions, is it really necessary to do so?

There are lots of moving parts in delivering a mobile broadband service, but it’s interesting to look at them holistically.

Different frequencies have very different propagation characteristics. It has been said that the best thing about 700MHz spectrum is its propagation characteristics. It has also been said that the worst thing about 700MHz spectrum is its propagation characteristics. It goes forever, has good in-building penetration, but these attributes present challenges from a frequency re-use perspective. Great for coverage. Not so great for capacity.

Conversely, the propagation characteristics in the 2GHz and 5GHz ranges are very different. These have limited range and poor in-building penetration, but these attributes are advantageous from a frequency re-use perspective. Poor for wide-area coverage. Great for capacity.

With the explosion in mobile broadband, the question today is whether the delivery of wireless broadband is a coverage play or a capacity play? And if capacity, what does the network look like?

Cell size and capacity have an inverse relationship. Macrocells, by design, are built to cover wide areas. However, the radios, spectrum and backhaul are all shared elements. The more users you have within a cell competing for these resources, the lower the throughput to each user. To increase capacity over an area, mobile operators pack in more, smaller cells (micro and picocells, and at the extreme low end of the spectrum, femtocells are being contemplated).

Andy Germano, vice chairman of the Femto Forum, said it very well recently when referring to the deployment of 4G networks: “Rather than go from big to small, they can go from small to big,” referring to cellular topology.

We know that, even today, it is challenging to keep up with the data capacity demands being driven by smart phones and other mobile devices. This is due to three factors: user demand for bandwidth, spectrum availability and cellular topology.

AT&T has been at the forefront of addressing the issue of adding data capacity to deliver a superior user experience for its iPhone subscribers by embracing Wi-Fi and unlicensed spectrum. Recently, at SUPERCOMM, John Stankey, president and CEO of AT&T Operations, stated:

“I don’t think a single macro wireless network is sustainable over time, given the pace of spectrum availability and what’s actually out there in terms of fixed spectrum. I think the reality is that there will be macro and micro networks. We’re starting down that path at AT&T, focusing on how we make Wi-Fi and licensed spectrum a more seamless experience for customers.”

So back to the beach.

There’s 80MHz of free, unlicensed spectrum, at 2.4GHz, being used today to augment data capacity by a leading licensed spectrum owner. Very interesting, but it’s still a small and pretty crowded stretch of beach. However, there’s 450MHz of spectrum, in the 5GHz bands, that is a beautiful stretch of beach – much less crowded, and ideal for small cells and frequency re-use, i.e. high-capacity locations.

Great stretch of beach, but how do I get there from here? Wi-Fi (802.11b/g) supports 2.4GHz and there are no 802.11a devices. The answer is 802.11n. Devices, such as the new Apple iPod Touch, are shipping with chipsets that support .11n and both frequency bands. Although the 5GHz radios are currently disabled, it is not unreasonable to expect that any upcoming versions of the iPhone, and other Wi-Fi enabled devices, will offer dual-band capabilities.

So, we will see 4G services rolled out at 700MHz, which will bring LTE to a big chunk of the country. But instead of waiting years (and paying billions) for more licensed spectrum to add capacity, I would say to Mr. Levin, “There’s a beautiful stretch of public beach up at this end, just waiting to be used.”

That’s my .02!

Martin Suter

Martin Suter is vice president of business development at BelAir Networks, the global market share leader in service provider Wi-Fi, enabling 3G data offload for wireless carriers, quad-play for cable operators, and managed WLAN business services. Previously, Martin was the CEO at Cohda Wireless, where he raised the company’s profile and negotiated a licensing deal with a Fortune 100 vendor in its core franchise. Prior to Cohda, he was vice president of business development at MeshNetworks Inc., a classic tech transfer/disruptive technology success story that achieved a major liquidity event for its investors in Q4/2004 with its acquisition by Motorola. Martin also was responsible for building several high profile alliances with and for leading technology companies, including Fujitsu, Microsoft, Netscape, Sun Microsystems, and Teradata. Additionally, Martin has successfully negotiated technology transfer, distribution and/or licensing deals with companies like 3Com, BioChem Pharma, Dow Chemical, Exodus, Fujitsu, IBM, Microsoft, Motorola, Netscape and Sun.

 

White Spaces: Wi-Fi on Drugs?

(Originally posted at http://www.von.com/blogs/guest/2009/02/white-space-wi-fi-on-drugs.aspx )

Firstly, I’d like to say that I applaud the FCC for opening up additional unlicensed spectrum for Internet use, but “white space” has set a new record for going from zero to the Peak of Inflated Expectations!

When Larry Page speaks of Google’s vision of “Wi-Fi on steroids,” it has a nice ring to it, but where have we heard that before? Hold on. Maybe it was three years ago, when Intel touted that WiMAX “was like Wi-Fi on steroids” and demonstrated a point-to-point link streaming video over 12 miles. In the same article, Clearwire had “introduced a precursor to WiMAX in four cities in Florida, Texas, and Minnesota, and hopes to use Intel's technology to add 16 more cities by year end.” Fast-forward three years, and Clearwire WiMAX service is available in Baltimore and Portland, and promises download speeds of “up to 6mbps,” hardly steroid like performance.

What does it mean to be “on steroids” anyways? Does it mean higher data rates? Increased range? Greater capacity?

In wireless, things are seldom cut and dry.

Many have commented on the increased range of a radio operating in the 50-700MHz bands as opposed to the 2.4GHz or 5GHz unlicensed bands. Technically, this is true, and these propagation characteristics are of value if there is a wide area to cover, and a relatively few devices. To achieve this, the base stations will look more like cell towers than wireless LAN APs, due to the very long (6m at 50MHz!) wavelength at these low frequencies. However, propagation is a double-edged sword. In environments where capacity is more of a consideration, frequency re-use is an imperative, and white space propagation will work against you.

Many have commented on the ability of 50-700MHz to penetrate walls. The answer here is, “Yes, the free space loss in these bands will be anything from 11 to 32dB better than Wi-Fi, so outdoor range in open areas will be much improved. And yes, the attenuation through materials such as concrete will be improved (maybe up to 10dB) but the long wavelength will make aperture effects -- like going through windows -- worse than at 2.4GHz.” Bottom line -- the jury is still out!

With respect to data rates, it is important to appreciate that the channel widths in white space are 6MHz as opposed to the 20MHz channels found with Wi-Fi in the unlicensed bands today. For this reason alone, it will be impossible for user performance to approach that which is currently available today. Then when macro-cell architectures are factored in, it won’t even be close. So much for Wi-Fi on amphetamines, Mr. Commissioner!

Stephen Rayment, BelAir’s CTO, recently drilled down into many of the specifics around white space in an online webinar. Sitting on IEEE committees, Stephen has first-hand visibility into many of the issues related to standards adoption, chipset availability and so on. It’s definitely worth a listen.

The industry does itself a disservice when it uses sound bites to articulate a position. Had Larry Page or Kevin Martin said, “Television white space will allow the Open Internet to bridge the Digital Divide, by bringing the web to rural areas across America,” it would have been equally compelling and more accurate.

As it stands, the Gartner Hype Cycle lives on with “white space,” but we’d all be better off if more of an effort was made to temper expectations with realistic dialogue. The eventual disillusionment would be less acute.

Drugs make for bad similes. We tell our kids that steroids are bad. They’re bad for athletes and they’re equally bad for technology.

That’s my .02!

Martin Suter

Martin Suter is vice president of business development at BelAir Networks, a provider of broadband mesh solutions for Wi-Fi, WiMAX, 4.9 GHz Public Safety and 5.9 GHz ITS networks. Previously, Martin was the CEO at Cohda Wireless, where he raised the company’s profile and negotiated a licensing deal with a Fortune 100 vendor in its core franchise. Prior to Cohda, he was vice president of business development at MeshNetworks Inc., a classic tech transfer/disruptive technology success story that achieved a major liquidity event for its investors in Q4/2004 with its acquisition by Motorola. Martin also was responsible for building several high profile alliances with and for leading technology companies, including Fujitsu, Microsoft, Netscape, Sun Microsystems, and Teradata. Additionally, Martin has successfully negotiated technology transfer, distribution and/or licensing deals with companies like 3Com, BioChem Pharma, Dow Chemical, Exodus, Fujitsu, IBM, Microsoft, Motorola, Netscape and Sun.