The Truth About Japanese Broadband
This week, we’re taking a look at flaws in the OECD broadband study and the problems with relying on it to justify broadband policy changes. One of the oft cited examples of countries doing broadband “right” is Japan. Ironically, Japan usually resides just below, or just above, the U.S. in the OECD rankings (in the latest rankings, they’re just below us). Yet Japan still commands outsized adoration.
However, a closer look reveals that the nirvana of Japanese broadband has been greatly exaggerated.
The Miracle of Japanese Broadband
If Japan has been held up as the shining example of low-cost, high-speed connections available to all, how could they slip behind the US?
The short answer is they never were the beacon of broadband they were held out to be. While the maximum speed available to “some” in Japan is 100 Mbps, the fact is that, in most cases, that is a theoretical top speed. The problem is these studies often use “advertised speeds” rather than actual speeds. Advertised speeds sometimes do not meet the promise. Most Japanese FTTH providers also include caveats that such a speed is not guaranteed and will, in fact, be lower during times of congestion.
In reality, roughly half of Japanese broadband connections are carried over DSL at speeds nowhere near 100 Mbps.
Surely, though, even with the speed being less than claimed, the fact that it’s so much cheaper makes it okay, right? Well, not exactly. The dirty little secret that goes unspoken in discussions of Japan’s low-cost ISPs is the fact that you have to subscribe to two different services to get an equivalent product to that offered in the US. For example, not only would you subscribe for the network connection (essentially the equivalent of getting the physical line), but you also have to subscribe to an ISP (the network in Japan is divorced from the content).
For access to a 100 megabit connection and Yahoo! ISP service, you’ll be forking out between $55 and $60. You may reach peak speeds at 2 a.m. when nobody is surfing, but the cost is not significantly different as a monthly outlay. While some studies look at this calculation as the cost per 100,000 bits per second per month, it’s hard to actually quantify that if your speed is not guaranteed (or even static).
Looking at the OECD stats, we’ve already uncovered problems with the definition of subscription and the mathematical issues that result from their unit of measurement. When other parties disseminate faulty and misleading information about cost and speed metrics, one realizes that most of the discussions about global broadband are largely anecdotal and not at all scientific.
As we have said throughout this series, there is a conversation to be had about ways to encourage broadband adoption. That conversation, however, should be driven by facts, not anecdotal evidence and faulty math.
In our next installment, we’ll look at the challenges the U.S. faces getting people online.
::
Send to a friend
February 8th, 2008 at 5:22 pm
I think I am in complete agreement with you here Michael.
Advertised vs actual speeds are a real issue everywhere. Personally I pay $35.99/month for an 9-15Mbps service. That is a pretty large swing is estimated speed. I am usually lucky to get 10Mbps. My ISP is a cable company and I do understand the complexity, as much as my mind can hold, of the share circuit cable uses. However, ISPs should be able to regulate or estimate more accurately their actual speeds.
I would welcome 100Mbps service for $60/month. Were you referring to advertised speeds in U.S. Dollars in your Japan example? At my current rate I would be paying $350/month for such a service and only get 90Mbps. Sadly I would probably go for it if I had the money to spend.
February 12th, 2008 at 11:43 am
The OECD study relies on advertised speeds for all of its statistics, yes. It’s one reason that cost per megabit figures are largely unreliable.
April 9th, 2008 at 8:00 pm
So if the theoretical 100Mbps Japanese connection isn’t really that fast, how fast is it? 1/2 that? 1/4 that? Even if it’s 1/4, it still pales in comparison to my 6Mbps connection, for which I pay almost as much.
If you were trying to convince me that American residential broadband isn’t a terrible deal compared to Japanese residential broadband, you succeeded (barely).
It’s not a terrible deal in comparison, just a very, very poor deal.
April 9th, 2008 at 9:05 pm
Michael –
Your comment assumes that 1/4 would be a good deal. Since DSL lines typically vary between 512k and a top speed of about 23 1/2 Mbps, I think you’re being generous assuming most people are getting that top speed (1/4 of 100 being 25).
I haven’t seen any studies of actual speeds for Japanese DSL, but I imagine it could be 1/8 or possibly even lower.
May 15th, 2008 at 10:14 am
I really have a hard-time why you are debunking the OECD broadband stats so hard. What are you trying to tell us? Are your trying to say that the market in the US is more saturated than we currently know. I understand why politicians dislike it when they hear that their country is lagging, but as an industry, why would you care…. Just means that there is more room to grow.
Now coming back to what is advertised and what is offered. You rightly say that speeds as advertised are not always the same as customers receive. But then again, none of your members is claiming the same. Let’s have a look at Docsis 3.0. It’s a great technology, but it is a shared medium. I do hope your members are not claiming that end-users can all watch HD-TV over IP when using Docsis 3.0 because that will breakdown pretty quickly. 160mbit/s shared over multiple users is not alot.
However, in Japan local loop lengths are alot shorter than in the US. Many times the DSL equipment is within 100 meters of the end-user, since it is installed in the basement of the appartment builings many Japanese live in. So the advertised 100Mbit/s can be easily achieved using VDSL2 (Michael Turk ADSL is a primitive technology, we’ve moved on)
Now if you want’t to cite something that the US is leading in, name the Fiber deployments. You’re doing better there than most of Europe. (oh wait, that’s Verizon)
But what does Hikari One cost this is the FTTH service of KDDI: Well you can get it for 73 dollars, 100mbit, telephony and television. http://www.kddi.com/english/dion/connection_services/ftth/hikari-one.html
Even more fun it’s to look at what the French are paying. Free.fr has an Internet, TV, and telephony offer (with HD-TV, remote viewing of tv, unlimited calling to 70 countries etc) for 30 euros.
(When comparing prices assume a 1 on 1 price parity between dollars and euros, the dollar has devalued quite alot in a small period of time, or did your income also rise with 60%?)
All in all, the conclusion is quite simple. The US is lagging, using any and all available statistics. None of the argument brought forward here change that fact in any significant way.
June 26th, 2008 at 3:49 pm
Does anyone else see something wrong with an American company badmouthing Japanese Internet providers? OCN, an ISP in Japan, is imposing 30 gig A DAY UPLOAD caps on their fibre connections. that’s 900-930 gigs a month, just upload. With unlimited download.
The ONLY large ISP in America that comes even close to competing with that is Verizon, with their FiOS service, which as it sits currently has no bandwidth caps. Time-Warner, Comcast, Cox, all have limits much lower than OCN.
Sure, you can go on about how the japanese consumers might not be getting full 100mbit service, but the last time I called my cable provider about not recieving the full advertised speed I was subscribed to, I was quite clearly told that this was an up-to number.
June 26th, 2008 at 4:22 pm
Larry –
Nobody is badmouthing Japanese Internet providers. What the original post attempted to do is point out that Japanese broadband a) doesn’t quite live up to the hype it gets in some circles, and b) is actually more expensive when you consider the fact that you’re paying for two different services. If you wanted to, you could also include c) suffers from many of the same troubles that plague American broadband.
As an example, you are correct about your “up-to” number. That’s really no different in Japanese ISPs. Shared node networks will always have difficulty in guaranteeing throughput because heavy users put a drag on everyone else’s connection.
That’s what OCN was attempting to deal with in imposing the cap and what American ISPs have attempted to deal with through network management practices and caps.
Do the Japanese benefit from population density that makes running fiber more cost-efficient when compared to the long hauls US ISPs have to run? Yes they do. Does that enable them to provide upgrades to their core network in ways that are more expensive in the US? Yes it does.
However, the Japanese have clearly demonstrated that you can’t simply build your way out of bandwidth constraints. Once they upgraded their network, traffic expanded quickly to consume all the available capacity. They’re now being forced, despite that increased capacity, to implement the same types of caps you see in the US.
June 27th, 2008 at 12:35 am
I bet anyone–if you gave the public of the choice of paying japanese rates for Japanase servic vs American rates for american service, what do you think most would prefer? Hummmm….
$50 for 100mbs line in Japan….US Telcoms would have NO business. The only reason that US telecoms make money is that they are a monopoly and really don’t compete. Guess I live in LA—I want ATT service, but cannot get it since ATT will not compete with guess who–Verizon. ATT will service 1 block over, but not my block. I can see their ads but cannot get them. So my real choices are cable vs. verizon–some choice–NOT! Maybe in the far future when there are four or more choices, Verizon will increase my speed without me paying more or give me more services. I talk to my friends in Holland—for what we pay for, it is pretty bad for what parts of W. Europe get.
June 27th, 2008 at 9:54 am
Honestly, I don’t think it’s any surprise that people would want to pay less and get more. That’s just common sense.
There are a large number of companies looking at new technolgies for delivering broadband. There are companies looking to serve broadband over power lines, via satellite, through wirelesss technologies, etc. etc. All of these will have different constraints.
With cable, the constraint is generally due to analog television channels carried on the same wire. For every channel migrated from the analog to the digital tier, cable can free up bandwidth for other things. As they have moved toward digital delivery they have increased bandwidth speeds, but they have also used that capacity to roll out services like OnDemand.
The Comcast pilot underway in the Twin cities is using that recovered capacity to test the DOCSIS 3.0 cable modems and delivering 50mpbps. Using the DOCSIS 3.0 channel bonding technology, cable companies are looking at increasing speeds up to 160mbps and beyond. That’s going to take some time to roll out nationwide, but it’s coming.
Now, not all cable operators have the same capabilities, but as they continue to upgrade their facilities they’ll be able to provide different and faster services.
However, you can’t just build your way out of the problem. As cable adds more capacity, application developers will write programs that consume more bandwidth. That’s why the industry is also working with developers to identify ways they can optimize their applications to strike a balance between network performance and not placing burdens on others.
June 27th, 2008 at 1:29 pm
I Only have one argument to prove that this ENTIRE post is completely false and misleading information…
http://www.dslreports.com/shownews/Long-Awaited-Japanese-Caps-Arrive-930GB-Per-Month-95580
Have a nice day.
-Dustin
June 27th, 2008 at 2:34 pm
Dustin –
If that’s your one argument, perhaps you should go back and read the many arguments that people in the comments on that thread made dismantling it. For instance:
July 1st, 2008 at 5:45 pm
I don’t see how you can compare Comcast’s 250GB/400GB+ MONTHLY DOWNLOAD caps to NTT’s 90GB DAILY UPLOAD caps. It’s also misleading to use percentages to compare such dislike bandwidths.
Yes, with that cap they could only use their full up speed for just over 2 hours a day, BUT your average high end computer is going to have a hard time pushing out that much data so quickly without a RAID storage system. If you even distribute that 90GB cap over the whole day that works out to almost 9Mbps! We’re just starting to see 1Mbps upload speeds become common on mid-grade tiers outside of Fios completive areas.
Even downloads at 1/10th of 100mbps would still be faster than over 90% of US connections at their peak OFF times.
Yes, unlike the US they have the line and the internet service separate, BUT this is good. You have a perfectly good fiber line running right to your house, you get to choose your internet provider separately from the type of network. This makes the ISP’s compete on thing like customer service, which has been almost completely forgotten about here in the states. Looking at the NTT’s prices and including the line cost, and tax, it seems to be between $40/m and $80/m.
September 1st, 2008 at 4:26 am
“These [900GB] caps are much, much worse” “than Comcast’s”.
Hilarious! You’re a funny guy Michael.
February 6th, 2009 at 11:10 pm
The author has clearly never witnessed the ridiculous speed of peers on a Japanese P2P filesharing program…
and TAD:
Your average high end computer today (and for the past several years) would have no problem pushing out 90GB per hour…much less per day as you speculate. Welcome to the 21st century.
US broadband is weak, borderline shameful, but getting better.