Communications

Packet inspection renders Internet connection unusable for blogging

Here's one I hadn't run into before.

The typical hotel broadband access service checks the MAC address of your computer. When you first connect, it sees the new MAC address, interrupts whatever access you were about to make and instead delivers a sign-up screen.  After you've signed up, your packets are routed directly to and from the Internet until your sign up period expires.

Of course this means every packet is being inspected to verify the MAC address is for a computer that has paid for access.  This is very light packet inspection, nothing like the DPI that various ISPs have been called to task for.  It's also very fast so it doesn't introduce noticeable latency, at least until now.

Last night I was staying in the InterContinental Nehru Place in Delhi and I observed a strange effect on perhaps 1 out of 3 or 1 out of 4 page loads.  First I would see a very brief splash screen from the hotel's ISP, then my desired web page would load.  This was unusual but not a bother.  Next, while using Bloglines, every now and then the left screen list of unread posts would fail to refresh, instead displaying a message about access had timed out.  However the next refresh always worked, so again, I can live with that.

Finally I tried to create a blog post but I lost the text as I went to post it.  Ugh!  Apparently, Typepad's blog author's content posting page requires reasonably fast turn around time for update messages.

Clearly MAC address redirection was beginning and then being canceled.  I'd never run into anything like that before -- a case where even shallow packet inspection is a problem.

Visualizing African Internet Connectivity

Here is a great illustration of the not-so-fast or reliable connectivity at African Universities as seen from Trieste, Italy.  Only 46 seconds.

Thanks to Marco Zennaro at the Science Dissemination Unit (via TIER at UC Berkeley).

Similar motivations for mobile and Internet usage in rural Zambia

Over the years I've uncovered a number of serious studies of the social and economic impact of mobile phone adoption in emerging markets.  In fact, my list of links (embedded in various blog posts: 1, 2, 3, 4, 5, 6, 7, 8) has become a reference for several people, for example here. I've been somewhat less organized regarding the social and economic impact of Internet access, but I've resolved to track such references as well. 

Here's something I stumbled on this weekend -- a study by Paula van Hoorik and Fred Mweetwa of TNO Information and Communication Technology in the Netherlands, entitled "Use of internet in rural areas of Zambia."

What struck me in this study was the parallels between the social and economic benefits they found and those conventionally associated with the adoption of mobile phones.

Based on our study, the most important social benefits are:
• Internet enables people to keep their network and enlarge it by communicating with friends, family and others
• Internet enlarges the world of people in rural areas by giving access to information
• Internet brings knowledge and supports education

The most important economic benefits are:
• Reduction of commute time
• Saving money in a lot of different ways, such as only traveling to pick up something when you know it is actually there instead of having to return multiple times
• Bringing new opportunities and using them, such as learning new farming methods or opening an internet café to make a living.

Most studies of mobile phone adoption in emerging markets, especially amongst the poor, show that social reasons - communicating with friends and family - come first, economic advantages come second.  And among both social and economic advantages, the ability to eliminate journeys, i.e. save time by not having to walk to the next village to ask a question, is extremely important.

I suppose I shouldn't be surprised.  At root, they are both communications technologies and our use of such technologies is driven by basic human needs that are remarkably similar, for everyone, everywhere. 

How tiered Internet pricing could actually facilitate P2P

Time Warner Cable's planned experiment with tiered charging for Internet access has generated a flurry of coverage in the blogsphere, but no new insights (at least that I've seen).

The primary problem ISP's complain about is that 5% of their customers use 90% of the available bandwidth and when they examine this traffic, it's mostly peer-to-peer file sharing.  A reasonable question is how to allow as much of this traffic as possible without increasing an ISP's variable costs or slowing down their other users.

This may not be as difficult as it appears.   Indeed if Internet access was as competitive as mobile telephony, we might already have seen what I'm about to propose — a combination of bundled pricing equivalent to mobile's "free nights and weekends" and "free on-net calls" with a way to facilitate P2P traffic that leverages exactly these "free" periods.

An ISP's costs

ISPs have some costs which are relatively fixed and others that are tied to usage.  A network is a relatively fixed cost and when it's not full, the incremental cost of adding traffic is zero!  This is the reason mobile operators give away free nights and weekend.  They've built their mobile network for the peak daytime traffic, so it costs them nothing to run promotions that add incremental traffic at off hours.  Peak hours and off hours may be different for an ISP, but the concept is the same. When a data pipe is lightly loaded the ISP's cost of adding incremental traffic is zero.

On the other hand, some ISP costs are usage based, for example "IP Transit" or more properly, Internet Transit.  This is the ISP's upstream cost to send and receive traffic to/from the rest of the Internet.  However, even here, usage-based costs occur at heavy usage.  Light usage periods don't save money.  To understand what's happening, it's worth a digression on Internet Transit.

Internet Transit

Internet access is monopoly or duopoly or a heavily regulated industry.  The middle mile connections from the local network to the Internet backbone may or may not be competitive depending on where you are.  But the Internet backbone itself is extremely competitive.  If you can get to a major Internet Exchange Point in the US or Europe, there are many providers offering extremely competitive rates for Internet Transit.  Typically these services are priced on a megabit per second per month basis (Mbit/s/Month) with lower rates for higher volume commitments.  The other key idea is that charges are based on the 95th percentile of all the five minute data rate samples taken during the month.  So an ISP can have a few bursts above their typical rate, as long as they represent less than 5% of the sampled intervals.

But this also means there is no extra cost to run at or near the typical rate at all times.

Local traffic

Even more important, if file sharing is done with other computers on the same ISP's network, then there is no need to pay for Internet Transit at all.  The question is how to figure out which potential peers are "on-net" and which are "off-net."

Sending signals to P2P software

Most P2P file sharing software has relatively little knowledge of locality.  Some P2P software practices "prefix awareness," for example, Joost gives preference to peers in the same /24 IP address block when they are available.  But if a major operator provided an automatic way for P2P client software to determine whether a prospective peer's IP address was currently reachable "for free", it seems likely the file sharing community would leap on it, and if there's money to be saved, active file sharers would download the new clients immediately.

A standard way to present such information might be via an extension to the XML-based response codes in one of the whois information exchange proposals, e.g. from ICANN or from APNIC.  Also, while what I'm proposing might start as a pricing plan rather like a mobile operator's "free nights and weekends" and "free on-net calling," it's not hard to see extensions where an ISP could offer dynamic access to underused capacity to those programs that were prepared regularly interrogate an ISP's server and use just the advertised off-hours capacity.

In closing

People liked fixed price deals.  Unlimited is great, but there's plenty of experience with bundles of minutes and the idea of data bundles has already showed up in 3G mobile data plans.  The combination of several tiered data bundle prices with the availability of "free" connectivity for "on-net" peers and during off peak intervals is likely to appeal to file sharers and produce better results for both the sponsoring ISPs and file sharers alike.

The Moral Instinct, cyberjustice and Internet law

There's an excellent article in today's NY Times Magazine, The Moral Instinct, in which Steven Pinker summarizes recent neuroscience research on the human moral sense.  This bears directly on how we think about justice and morality in cyberspace.

He covers the enormous progress (at least since I was in college) in our understanding of where the human moral sense comes from, including what is universal, what is cultural and how we can be easily mislead by "an aura of sanctity, distracting us from a more objective reckoning of the actions that make people suffer or flourish."

Illusions are a favorite tool of perception scientists for exposing the workings of the five senses, and of philosophers for shaking people out of the naïve belief that our minds give us a transparent window onto the world (since if our eyes can be fooled by an illusion, why should we trust them at other times?). Today, a new field is using illusions to unmask a sixth sense, the moral sense. Moral intuitions are being drawn out of people in the lab, on Web sites and in brain scanners, and are being explained with tools from game theory, neuroscience and evolutionary biology.

While the article doesn't mention the Internet, a better understanding of our moral sense can only help temper our approach to law and culture in the new social relationships facilitated by the Internet.  Depending upon the cultural setting, subjects as diverse as ethnic humor, pornography and religious discussion can be completely acceptable or morally outrageous and yet, except for some language effect, the Internet cuts across cultural boundaries.  Then there are Internet-related tragedies like the death of Megan Meier which provoke discussion and calls for sweeping laws to regulate cyber-behavior.

Undoubtedly, our laws and culture will evolve in response to the Internet age, but hopefully we do this with care, taking advantage of what we can learn from history and from the new science of the moral sense.

Here are a few short quotes from Pinker that struck me:

There are many other issues for which we are too quick to hit the moralization button and look for villains rather than bug fixes.

Our habit of moralizing problems, merging them with intuitions of purity and contamination, and resting content when we feel the right feelings, can get in the way of doing the right thing.

Far from debunking morality, then, the science of the moral sense can advance it, by allowing us to see through the illusions that evolution and culture have saddled us with and to focus on goals we can share and defend. As Anton Chekhov wrote, “Man will become better when you show him what he is like.”

If you are at all interested in morality, neuroscience or evolutionary biology, I highly recommend the Pinker article.

edushi.com – Possibly the best city maps in the world?

After my last post (on sanyamap.com), Tariq Mufti pointed me at an even better source of cities maps, also in China.  It's Edushi.com (roughly E-City in Chinese) and they currently cover 24 cities. The title of this post is copied from Dr Andrew Hudson-Smith who's written up edushi.com in some detail.

Edushi.com is owned by and uses technology developed by Hangzhou Aladdin Information & Technology Company.  Their English language description of the technology includes this:

Aladdin has first put forward the new concept of Internet application: Edushi(or rather, E-City) which is a corresponding result of the local governments’ encouragement of building “digitalized cities”. As a platform for 3-D emulation online-interaction, E-City is based in the WEB GIS and virtual-real technology.

"Digitalized city" integrates E-map, E-yellow page (city telephone directory), E-business, virtual community etc. functions.

"Edushi" allocates various information of the city based on geo information of the real city.

"Edushi" does not only emulate the architectural shape, geological status of the real city, but also vividly reproduce the entire city as well as the social activity and economical activity on Internet.

They also mention support for mobile phones, in-car navigation, a community information platform and a digital TV terminal application.

While I don't read Chinese, if you float your cursor over the list of cities at the top of this page, you can see their URLs in standard ASCII.  Using this approach I was able to go to the city view for Xian and navigate to some tourist sites that I have visited.

All and all, a fascinating combination of 3D maps, virtual reality, directories and current city culture.

China's Internet, at least as interesting as anything in the west

A common misunderstanding in the west suggests Internet entrepreneurs in China just copying whatever works in the west.  I'm sure that's true in some cases, but I've written in the past about peer-to-peer TV where Chinese companies were several years ahead of Joost and others in the west.

Today, I noticed a fascinating post by a blogger operating under the name Sun Bin.  He points to a mapping/ virtual earth website for Sanya, a resort city in Hainan province in southern China.  If you are proficient in Chinese,  try the website yourself.  Otherwise, grok this:

Yes, this is just one city so it's hardly on the scale of Google Earth or Microsoft Virtual Earth, but the user interface looks very interesting as it combines satellite photography, street maps, SimCity graphics and actual photographs of buildings, all in one beautiful graphic.

My point is not that Chinese entrepreneurs are ahead or behind, but that there are more Internet users in China than in the US, and the entrepreneurs that cater to them are smart and are doing interesting things that we seldom hear about in the English language blogsphere.

I follow a few bloggers, like Sun Bin and Gang Lu, who cover China in English (at least part of the time) and I try to maintain contacts with individuals I know in China, but it hardly makes up for not being able to read Chinese.

Rate of growth of traffic at the Amsterdam Internet Exchange

Here's a great historical graph, from page 28 here, showing the growth in traffic at the Amsterdam Internet Exchange.

By my take off from this graph, there was 1000x growth in nine years.  That means traffic doubles every 11 months.  While this is only one data point, it's new to me and fairly up-to-date.  It's also roughly consistent with the 12 month doubling rate mentioned in my earlier post.

Connect 2007 Madrid opens with panel on the mobile industry

The session was entirely Q&A (no slides) which resulted in a great discussion – broad ranging and much better than talking heads reading slides!  Luca has already written up his reactions.

I'm writing this after the fact as, at the last minute, I was tagged to participate, filling in for Vincenz Wagner of Jamba who's arrival has been delayed.  The opening session at Connect 2007 in Madrid was entitled "Industry Overview" with Joel Hughes, VP & GM of our Mobile Applications business moderating.  In the end the panelists were:

• Philip Kelley, Director, Mobile TV Standardization, Alcatel-Lucent
• Kari Lahtinen, Business Development Manager, Elisa Corporation
• Peter Karney, Senior Technical Marketing Manager, NEC
• Brough Turner, i.e. yours truly :-)

Since I was participating, I have only a few interesting items (at least interesting to me) that I noted during this discussion:

Philip commented that the predominate use of 3G is to connect PCs to the Internet, i.e. dumb pipe mobile Internet access.  Philip also mentioned Triple Play, which seems old hat to me, as innovative in (parts of) Europe.  I argued that this was marketing innovation (bundling), not really a new service.  We agreed the innovation was in cost and convenience of the services.

Several panelists seem to think that innovation would come by porting Internet applications to the mobile space.  I argued that was currently true, but only because the Internet was open and mobile was still closed.  There are many characteristics of mobile (like mobility and intimacy) that will foster new applications but we don't yet have the open environment that allows zillions of developers to experiment.

At one point, I made a derogatory comment about most VoIP being just digital POTS.  The ensuing discussion brought out the parallel between Skype (which combines voice and IM) with what's happened in mobile telephony, i.e. the combination of voice and SMS to achieve the same objectives.  Either way, people want to determine the actual availability of the person they are about to call and people need a way to communicate when they can't talk.

The day the routers died (to music)

The RIPE 55 meeting in Amsterdam has ended and Raindeer just uploaded this summary to his blog Lunatic Thought:

"The RIPE 55 meeting has just concluded. There was much debate on what to do on the imminent depletion of the unallocated IPv4 pool in 2010. We could do nothing or we could create a market place and facilitate transfer of IP-adresses, but it's all a train wreck waiting to happen. This is best shown however by a beautiful song "The day the routers died" also available on Youtube written and performed by Gary Feldman. So please all upgrade to IPv6 soon, or else you will not get 40Gbit/s to your mother."

and points to this wonderfully funny video: