Thursday, October 23, 2008
Beyond white spaces
Back in 1999 I wrote a column that envisioned the uses of digital wireless in the home. I compared two nascent, much-touted wireless protocols, Bluetooth and HomeRF. I completely, totally, slippery-dash missed Wi-Fi. There had been a public 802.11 spec since 1997. The first 802.11b devices, which made Wi-Fi popular, burst onto the scene in early 2000, just a few short months after my clueless insights. Today HomeRF is forgotten, Bluetooth is for ugly ear jewelry and Wi-Fi rulz.
Broadband over Power Line (BPL), once touted as The Third Pipe, is another dud. DSL Reports just declared 2008 The Year BPL Died, citing tech problems, performance problems and market problems. BPL's flaws have been obvious for years to those whose salary doesn't depend on its success. Rahul Tongia's 2004 paper, Can Broadband Over Powerline Carrier Compete? should have been the last nail in its coffin. I blogged BPL-RIP in 2005, saying, "BPL is so crippled there's no 'disruptive' in it." The mistake persisted for years longer than it should have, supporting the myth that multi-modal competition is a substitute for national policy. Today all who look see that fiber rulz.
There's so many things a new technology needs to get right if it is to succeed. First it needs to be significantly better than what's out there already. It needs to be cheaper out of the gate. It needs to open up new markets and prove itself useful. It needs to achieve economies of scope and scale. It needs economics that won't strangle it in its crib.
White spaces, the FCC proceeding about using digital TV spectrum that's only lit in cities far away, strikes me as YARV, yet another risky venture. My friends at Google and the New America Foundation held a meeting on white spaces the other day. The title of the meeting, "Pervasive Connectivity," seems an over-reach without a systems approach that includes pervasive fiber and already-deployed wireless protocols, such as Wi-Fi and LTE. Who's going to build devices in such scale that they'll out price-performance Wi-Fi? Who's going to offer service that'll out-pervade the cellular network? And what about that one critical factor that every wireless network must consider, backhaul?
The white space idea is fraught with unsolved issues. In its first incarnation, its radios will only broadcast at 40 milliwatts, which raises doubts about any advantage it might have over Wi-Fi. (Meanwhile, Wi-Fi itself keeps improving.) One wireless data expert I know suspects that propagation characteristics around 300-400 MHz could cause major problems. Another expert I know, considering all the issues, suggests that white space may find its niche in under-served, e.g., rural, areas. Neither was heard at the "Pervasive Connectivity" meeting. Such views, no matter how correct, are problematic when specific policy change is the over-arching goal.
The FCC may or may not consider white space in its meeting on November 4, which conveniently falls on election day. Hey, the FCC was all for BPL. Indeed, the FCC absolutely should allow white space experiment to proceed; sins of commission, pun intended, are better than sins of omission, especially when they allow honest experimentation.
Google's strategy is an incremental, multi-frontal approach. That's OK, we need to make progress where we can. Even if the progress is limited to, "yet another experiment." But, in addition, there should be another effort, a big, synoptic plan, that rises above specific technologies and specific policy agendas, that uses all of the expertise available, to craft a comprehensive vision worthy of the moniker "Pervasive Connectivity."
Suppose in the next few months we get the opportunity to propose a real plan for Pervasive Connectivity for our nation, could we rise to the occasion? Or would we remain conditioned to the mindset of the last eight years, when small increments counted as great victories. Naomi Klein cites Milton Friedman's idea that in a nodal moment, the ideas implemented are the ideas lying around. Rick Perlstein, in an essay called, A Liberal Shock Doctrine, points out that even progressive progress occurs in spurts, at opportune times. We shouldn't limit our vision to one specific technology or one tactically available sliver of spectrum. Now is the time to have a comprehensive plan "lying around" for the network we really want.
Broadband over Power Line (BPL), once touted as The Third Pipe, is another dud. DSL Reports just declared 2008 The Year BPL Died, citing tech problems, performance problems and market problems. BPL's flaws have been obvious for years to those whose salary doesn't depend on its success. Rahul Tongia's 2004 paper, Can Broadband Over Powerline Carrier Compete? should have been the last nail in its coffin. I blogged BPL-RIP in 2005, saying, "BPL is so crippled there's no 'disruptive' in it." The mistake persisted for years longer than it should have, supporting the myth that multi-modal competition is a substitute for national policy. Today all who look see that fiber rulz.
There's so many things a new technology needs to get right if it is to succeed. First it needs to be significantly better than what's out there already. It needs to be cheaper out of the gate. It needs to open up new markets and prove itself useful. It needs to achieve economies of scope and scale. It needs economics that won't strangle it in its crib.
White spaces, the FCC proceeding about using digital TV spectrum that's only lit in cities far away, strikes me as YARV, yet another risky venture. My friends at Google and the New America Foundation held a meeting on white spaces the other day. The title of the meeting, "Pervasive Connectivity," seems an over-reach without a systems approach that includes pervasive fiber and already-deployed wireless protocols, such as Wi-Fi and LTE. Who's going to build devices in such scale that they'll out price-performance Wi-Fi? Who's going to offer service that'll out-pervade the cellular network? And what about that one critical factor that every wireless network must consider, backhaul?
The white space idea is fraught with unsolved issues. In its first incarnation, its radios will only broadcast at 40 milliwatts, which raises doubts about any advantage it might have over Wi-Fi. (Meanwhile, Wi-Fi itself keeps improving.) One wireless data expert I know suspects that propagation characteristics around 300-400 MHz could cause major problems. Another expert I know, considering all the issues, suggests that white space may find its niche in under-served, e.g., rural, areas. Neither was heard at the "Pervasive Connectivity" meeting. Such views, no matter how correct, are problematic when specific policy change is the over-arching goal.
The FCC may or may not consider white space in its meeting on November 4, which conveniently falls on election day. Hey, the FCC was all for BPL. Indeed, the FCC absolutely should allow white space experiment to proceed; sins of commission, pun intended, are better than sins of omission, especially when they allow honest experimentation.
Google's strategy is an incremental, multi-frontal approach. That's OK, we need to make progress where we can. Even if the progress is limited to, "yet another experiment." But, in addition, there should be another effort, a big, synoptic plan, that rises above specific technologies and specific policy agendas, that uses all of the expertise available, to craft a comprehensive vision worthy of the moniker "Pervasive Connectivity."
Suppose in the next few months we get the opportunity to propose a real plan for Pervasive Connectivity for our nation, could we rise to the occasion? Or would we remain conditioned to the mindset of the last eight years, when small increments counted as great victories. Naomi Klein cites Milton Friedman's idea that in a nodal moment, the ideas implemented are the ideas lying around. Rick Perlstein, in an essay called, A Liberal Shock Doctrine, points out that even progressive progress occurs in spurts, at opportune times. We shouldn't limit our vision to one specific technology or one tactically available sliver of spectrum. Now is the time to have a comprehensive plan "lying around" for the network we really want.
Technorati Tags: Broadband, Bushco, DisruptiveTechnology, FCC, fiberoptics, Google, Infrastructure, Innovation, Organizational Culture, ShockDoctrine, WiFi, WirelessNetworks
Comments:
I think that David Isenberg is right in his analysis of the TV white spaces issue. However, reading his note prompted several thoughts.
First, for an interesting discussion of the HomeRF/802.11 evolution see http://www.iep.gmu.edu/documents/WLAN_History_Paper.pdf. Reading that paper gives a lot of insight into why 802.11 prevailed in the marketplace. It's written by someone who was on the HomeRF side—but it seems quite unbiased to and fair to me. In 1999, it was quite reasonable to expect HomeRF to prevail in the marketplace.
Second, there's a great book called "The Triumph of Ethernet" by Urs von Burg, Stanford University Press, which examines how Ethernet, which many thought inferior to Token-Ring, won in the marketplace. One factor that helped Ethernet win was the fact that IBM effectively controlled the token-ring standard and, if you were a small guy in the industry, you probably preferred to build products that IBM could not make obsolete overnight.
Third, some thoughts on the white space debate. A caveat, I'm a participant in that debate. I've taken the position that licensing the white space would deliver far more to consumers than would unlicensed white space.
Unlicensed white space devices will be inferior to current WiFi in most incarnations. They will probably be built to use 6 MHz channels—but 802.11 now uses 40 MHz. At short ranges, it's bandwidth that determines capacity. So, unlicensed devices at 2.4 and 5 GHz will outperform unlicensed devices in the TV white space. Relatedly, MIMO capabilities are expected to be better at 2.4 and 5 GHz than in the white space. That superiority will probably create another factor of 2 or 3 advantage in short-range capacity for 2.4 and 5 GHz systems. Together, these factors create a capacity advantage for 2.4 and 5 GHz 802.1 of about a factor of 10 or 20.
For long ranges, the power limits on the unlicensed devices reduce possible coverage and run up costs. Licensed services in the 700 MHz band and the PCS can use base station powers three thousand times higher than are proposed for unlicensed TV white space fixed stations. Propagation may be better in the low UHF (white space) than in the PCS band, but not 3,000 times better. (Actually, "better propagation" is a complex issue—and it is not necessarily the case that propagation is "better" at the low UHF. It all depends upon what the meaning of "better" is.)
It's hard to think of any engineering model in which one can provide rural coverage with unlicensed TV white space devices at costs anywhere those possible for the current licensed operators. For a quantitative exposition of the inability of TV white spaces to provide rural coverage see
http://fjallfoss.fcc.gov/prod/ecfs/retrieve.cgi?native_or_pdf=pdf&id_document=6520176878
Finally, there is a key issue in the white space debate that the advocates of unlicensed completely fumble. That is, "How do we put in place rules and incentives that will make it easy to transition the current TV spectrum to other uses when that transition makes sense?" A licensed approach creates mechanisms for doing so. See, for example,
Hazlett and Smith's op-ed in the WSJ.
http://online.wsj.com/article/SB122299012125700337.html
In contrast, if the white space is filled up with wireless X-box controllers and cordless phones, it will be far harder to transition the TV bands to new technologies and new uses.
I think that David's hint that unlicensed devices in the TV white space will be a dud in the market is correct. 802.11 at 2.4 and 5 GHz beats them for short range; the power is too weak for long range.
This message is already long enough, but if anyone wants more on the topic, feel free to contact me.
Chuck Jackson
Post a Comment
First, for an interesting discussion of the HomeRF/802.11 evolution see http://www.iep.gmu.edu/documents/WLAN_History_Paper.pdf. Reading that paper gives a lot of insight into why 802.11 prevailed in the marketplace. It's written by someone who was on the HomeRF side—but it seems quite unbiased to and fair to me. In 1999, it was quite reasonable to expect HomeRF to prevail in the marketplace.
Second, there's a great book called "The Triumph of Ethernet" by Urs von Burg, Stanford University Press, which examines how Ethernet, which many thought inferior to Token-Ring, won in the marketplace. One factor that helped Ethernet win was the fact that IBM effectively controlled the token-ring standard and, if you were a small guy in the industry, you probably preferred to build products that IBM could not make obsolete overnight.
Third, some thoughts on the white space debate. A caveat, I'm a participant in that debate. I've taken the position that licensing the white space would deliver far more to consumers than would unlicensed white space.
Unlicensed white space devices will be inferior to current WiFi in most incarnations. They will probably be built to use 6 MHz channels—but 802.11 now uses 40 MHz. At short ranges, it's bandwidth that determines capacity. So, unlicensed devices at 2.4 and 5 GHz will outperform unlicensed devices in the TV white space. Relatedly, MIMO capabilities are expected to be better at 2.4 and 5 GHz than in the white space. That superiority will probably create another factor of 2 or 3 advantage in short-range capacity for 2.4 and 5 GHz systems. Together, these factors create a capacity advantage for 2.4 and 5 GHz 802.1 of about a factor of 10 or 20.
For long ranges, the power limits on the unlicensed devices reduce possible coverage and run up costs. Licensed services in the 700 MHz band and the PCS can use base station powers three thousand times higher than are proposed for unlicensed TV white space fixed stations. Propagation may be better in the low UHF (white space) than in the PCS band, but not 3,000 times better. (Actually, "better propagation" is a complex issue—and it is not necessarily the case that propagation is "better" at the low UHF. It all depends upon what the meaning of "better" is.)
It's hard to think of any engineering model in which one can provide rural coverage with unlicensed TV white space devices at costs anywhere those possible for the current licensed operators. For a quantitative exposition of the inability of TV white spaces to provide rural coverage see
http://fjallfoss.fcc.gov/prod/ecfs/retrieve.cgi?native_or_pdf=pdf&id_document=6520176878
Finally, there is a key issue in the white space debate that the advocates of unlicensed completely fumble. That is, "How do we put in place rules and incentives that will make it easy to transition the current TV spectrum to other uses when that transition makes sense?" A licensed approach creates mechanisms for doing so. See, for example,
Hazlett and Smith's op-ed in the WSJ.
http://online.wsj.com/article/SB122299012125700337.html
In contrast, if the white space is filled up with wireless X-box controllers and cordless phones, it will be far harder to transition the TV bands to new technologies and new uses.
I think that David's hint that unlicensed devices in the TV white space will be a dud in the market is correct. 802.11 at 2.4 and 5 GHz beats them for short range; the power is too weak for long range.
This message is already long enough, but if anyone wants more on the topic, feel free to contact me.
Chuck Jackson
