From: Michael Kende <[EMAIL PROTECTED]> Date: May 22, 2007 10:30:03 AM EDT
Dave, The CWA in the cited congressional testimony have provided an apples and oranges comparison. For the US, they did extensive tests of actual speeds to get the 1.9 Mbps number, but for the other countries they went to a report that used advertised speeds. What is disingenuous about this is that in the ITIF report that they cited (http://www.itif.org/files/BroadbandRankings.pdf), the US had an average advertised download speed of 4.8 Mbps, but they ignored that in favor of their lower actual speed. We did a study (the results of which you posted to this group several months ago) that gathered data on actual download speeds in other countries, and found that the higher the advertised speeds, in general the bigger the gap between actual and advertised speeds, so some of the countries mentioned by the CWA have significantly lower download speeds then the headlines listed here. I raise this simply to point out that this appears to be an attempt to mislead the audience about this issue, and is particularly ironic in a testimony in which the President of the CWA states that "Good data is the foundation of good policy." (in support of a bill on broadband census taking) Michael From: Daniel Weitzner <[EMAIL PROTECTED]> Date: May 22, 2007 2:02:16 PM EDT I'm sure these numbers are real, but it's less than clear what we should infer from them. I'd like to know answers to questions such as: -what is the average upload speed? -how many blog/YouTube posts are made from these various countries? -how does the overall volume of upstream content posted correlate with the average download and upload speeds? -what regulatory and ownership structures come with these network configurations? As far as I know, some of the countries on the list have made these impressive broadband deployments happen either by direct public subsidy and/or by continuing monopoly control of the networks. I'm not condemning either option necessarily, but they do have baggage that comes with them once you've gotten your bits fast. As much as I like high bandwidth downstream, good upstream bandwidth and non-discriminatory carriage are as important of not more so than raw bandwidth. Danny From: Dave Burstein <[EMAIL PROTECTED]> Date: May 22, 2007 5:28:44 PM EDT Dave, Daniel Jumping in because Daniel's comment "I'm sure these numbers are real," isn't so. Actually, some of these CWA figures are unsupportable, based on unrealistic "up to" speeds. Real average in France is closer to 8-12 than 17. Japan has 7M fiber connections at 50 meg plus, but with the DSL and cable included the "overall average" is probably less than 15. Canadian speeds are similar to the U.S., although prices are still lower. The UK and soon Spain are rapidly moving to faster DSL's, advertised as "up to 24" and averaging 8-15 down. France has fiber coming fast starting now, and Germany, Holland and others VDSL from the curb for a real 30-50 down, 10+ up. Kende's comment that the 15-25 DSL is typically half that is on target. Still a heck of a lot better than the typical U.S. downloads of 1.5 - 6, with an average of 3-4. Weitzner is on target noting upload is very important, and the ADSL/cable current limits of about a meg not "a good thing." Perhaps time to write an essay "up to considered harmful." >Dan also wondered how many blog/YouTube posts are made from these >various countries? Loads. Blogging is incredibly active in China and France, I know, and in France Daily Motion is providing a very active video service winning share from YouTube, many in French and other European languages. >-what regulatory and ownership structures come with these network >configurations? Overwhelmingly, the countries with faster broadband have more competition than the U.S., with subsidies generally a minor factor. With the exception of China, soon to pass the U.S. in broadband totals, monopolies have done badly and duopolies (U.S. telco and cable) not well. Leaders Japan 4 (had 6 in the major growth stage) France 4+ (has 7, but rapidly consolidating to 4) Korea 4 (had 7-8 in the major growth stage) Germany 4+ (just the last two years, but inspiring 50/10 to millions) Overwhelmingly, the countries with 4 or more are getting much better prices and higher speeds. This is typically accomplished by efficient "local loop unbundling", in which 2-4 competitors have built DSLAM networks and backhaul reaching 50-70% of the nation, sharing only the wires under the streets. The UK is starting to see a similar boom and improvement in offerings. Countries like the U.S. with two successful carriers typically are behind. In turn, the countries with only one major carrier (Greece, Ireland, Germany until 2005, Spain) are behind the U.S. The main exception is China, where China Telecom and China Netcom have virtual monopolies in their territories but strong government guidance. Here are some basics about broadband speeds. Based on eight years of reporting at DSL Prime, including many real world data reports. ADSL2+ 3-15 meg down for most people, less at extreme distances. Peak of 26 meg down is good for less than 100 meters. Significant numbers (10% in U.S.) may not get even a meg. 700K to 1 meg up, commonly, Cable 5-10 down for most people. (36 meg shared) A well run cable network has surprisingly little congestion, a bad one extensive slow periods. Upstream typically well under a megabit, with some networks geared for 200K. DOCSIS 3.0 is a real 25-100 megabits in both directions, and early versions are shipping in Japan, France, and the U.K. Expect big change in 2008-2012. Fiber 50+ meg in both directions (including VDSL from the curb/basement) Wireless, including WiFi, Wimax, EVDO, 3G CDMA/GSM, etc. 200K-2 meg typical downstream, slower upstream. I am not a wireless expert, but my opinion is that wireless will generally peak out in single megabits as video demand increases. That's not inevitable, but far more towers/access points (or spectrum) will be required to reach higher speeds as video over the web increases. Satellite is a form of wireless, but currently even slower. Any operator can do a cheap, lousy job and foul things up, no matter the technology. Essentially every network is shared in the backhaul and the connection to the wider Internet. This "contention" or "over-subscription" works fine 99% of the time in a robust network, but if overdone means user speed plummets at peak times. Bandwidth isn't free, but in a large volume network decently connected it's darn cheap. The difference between a good network that delivers peak speeds reliably and a congested one is less than 5% of the typical retail charge. But some operators cut corners on this, and smaller ISPs that have to overpay for backhaul to incumbents are particularly challenged. The physical network needs to be maintained, with reflections, bridge taps, etc cleaned up when needed. If I had more time, this post would be shorter. Also sorry to review why U.S. unbundling policy failed (1998-2003, or the comparative subsidy numbers. Dave Burstein Editor, DSL Prime From: "David P. Reed" <[EMAIL PROTECTED]> Date: May 22, 2007 6:19:03 PM EDT I'm a little troubled by the promotion of Verizon's service over RCN's and Comcast's. RCN at my home is rather good service, with 20 Mbps down (peak is somewhat higher). Comcast is offering such levels of service in some places. But the key thing is that FIOS is hardly widely deployed! In fact Verizon sold off its businesses in a number of rural states, perhaps because that would make their percentage of deployment look better. Let's be fair. In homes passed, all US broadband offerings pretty much suck compared to, say, S. Korea. Don't write subject lines that pretty up the picture as if Verizon is some kind of special case. Most places Verizon has no near term plans to do FIOS at all, and that's even if they get special national franchising overrides on the real cash cow - subscription TV. --------------------------------------------------------------- WWWhatsup NYC http://pinstand.com - http://punkcast.com --------------------------------------------------------------- _______________________________________________ Discuss mailing list [email protected] http://lists.isoc-ny.org/mailman/listinfo/discuss
