3G Roaming – Pleasure with Pitfalls

Thanks to an international 3G subscription I have long ago given up searching for Wifi hotspots at the locations I travel to. Instead, as the Sprint guys put it, the 3G network is following me wherever I go. I see it the other way round. Wherever I go, the network is already there. While I have always managed so far to find good 3G or EDGE connectivity, there are some pitfalls which I wouldn’t have thought existed anymore three years after the commercial launch of the first UMTS networks. Here are some strange but true examples:

France: For a month or two now, Orange, the mobile operator that allows me to roam to its 3G network, seems to have a new software version running on their UMTS network in Paris. Since then, my almost brand new Nokia N70 behaves strangely and has trouble establishing a dedicated bearer during a packet session after some time of inactivity. This results in very long delays in the order of 10 seconds or more when I click on a link after some time of network inactivity. The only remedy is to trick the network into letting me have a dedicated channel continuously by constantly sending pings to a host on the network. While this helps for notebook use, I can’t use this trick while web browsing via the mobile phone. So I prefer using Organe’s EDGE network by forcing the mobile into GSM only mode. In other parts of the country things work flawlessly. This is probably due to the fact that Orange uses different UMTS access vendors in different parts of the country: Alcatel in Paris, Nortel and Nokia in other parts of the country. Well done, Orange!

Germany: Here, I have a greater choice of UMTS roaming partners: T-Mobile, E-Plus and O2. The first two work flawlessly with my Nokia N70. O2 also works well if the mobile can find the UMTS network. Sometimes, however, the mobile just refuses to see the network, especially after the mobile has lost coverage for some time like for example if I have parked the car in an underground garage. Switching the phone on or off does not change anything. Even a manual network search, which shows that the 3G network is available, does not force the mobile back into O2’s UMTS network. The only action that helps sometimes is to go back to the place where the phone has no GSM or UMTS coverage of O2’s network for a minute. It’s a repeatable phenomenon and I’ve only seen it in Germany and only with O2. Also, I have to restart the phone much more often than in other networks, about once per day, as after some time I can’t connect to the Internet anymore.

Austria: Again, I have several roaming partners for UMTS: T-Mobile, A1 and One. In the A1 network I have detected the strangest problem yet. With both my Nokia N70 and my somewhat older Sony Ericsson V800 I have problems to send data from the notebook to the network. An analysis with Wireshark, a network tracing tool, revealed that the network has problems with large IP packets in uplink direction. At first I thought it was a specific mobile problem in combination with the network components used in the A1 network. However, as two completely different phones have the same problem it seems to be a general network issue. What helps is to reduce the Maximum Transfer Size (MTU) of the notebook for dial up connections. After changing the MTU size to 480 bytes as described in this Microsoft bulletin, things worked a lot better. But quite frankly, I prefer using ONE’s network where things work as they should. Just in case I ever end up in a part of Austria where ONE’s network is not available, I still have my MTU jocker ready.

All of this is very strange as both of my UMTS phones are widely used in these countries. But I think it shows that 3G interoperability is still not where it should be. Nevertheless, things are not as bad as they might seem after describing these three cases for the following reasons: Even in the countries described above I have found at least one network in which things work flawlessly with my mobiles. In addition, here’s a list of countries where I didn’t encounter problems, at least not in the networks I used: Switzerland (GPRS and EDGE), Spain, Italy, Belgium (EDGE), The Netherlands, U.K. and Portugal.

CoComments Big Brother Plugin

CoComment is a great tool to keep track of comments left by you and others on blogs. Lately, they have added a Firefox plugin which makes the service even more comfortable to use. Unfortunately, however, the plugin raises a serious privacy issue about which CoComment does not inform its users about.

I’ve tried to get in contact with them to get a statement but got no response. O.k. so let’s discuss it in the blogsphere. While the plugin’s functionality is doubtlessly interesting, it contacts the CoComment server after each and every download of a web page. Part of the message sent to CoComment is the visited URL. In effect, CoComment is thus aware of each and every move a user takes on the Web. Cookies pale in comparison to this! What is so unacceptable to me is that CoComment does not inform users about this to let them choose if this is acceptable to them or not.

For my part I’ve uninstalled the plugin again and have reverted back to their bookmarklet. It offers less functionality but preserves my privacy except of course of keeping track of my comments, which is what it’s supposed to do.

Centrino WLAN vulnerabilities – Getting your virus with a malformed packet

Maybe it’s because we are used to getting patches to our PC every month or so now that the following story has not seen wide spread attention so far: Intel admits Centrino chipset driver issues: These allow attackers to send malformed wireless lan frames to insert and execute malicious code (read viruses).

This is scary for two reasons:

Firstly, no user interaction is required. This means that a user doesn’t even have to browse to a malicious webpage to get infected. It’s enough to have your WLAN card activated. Airports and conferences might become nice playgrounds for past time hackers and self replicating viri once an exploit for this hits the net.

Secondly, the fixes have to be installed manually. There is no auto update functionality like for example for Microsoft Windows patches which are downloaded and installed by the operating system once available. I’ve downloaded and installed the patch for a notebook with a Centrino 2200BG card. A 129 MB (MEGABYTE!) download. Incredible! At least it installed o.k. and the driver was updated. Then I downloaded the patch for another notebook which has a Centrino 2100 chipset. A refreshingly short 13 MB download… When executing the file it installed an update for the helper program but failed to update the driver for the chipset. The program showed no sign that the driver, where the real problem sits, was not updated. Perfect, the average user will never notice that… So I manually installed the driver update from the hardware settings. To make the day perfect, many notebook vendors have chosen to write their own wireless lan configuration utilities that interface with the driver in some way. Of course they could be broken if you install the driver. Take a look at F-Secure’s blog. Once an exploit for this hits the wild, it’s going to be big.

Speculation: Could the same scenario happen in the cellular world, too? In theory I could imagine this happening in the cellular world as well. Imagine that somebody finds a bug in the IP stack of mobile devices or in the mobile browser that could be exploited in the same way. Downloading fixes on such devices is still a procedure most device manufacturers have yet to come to terms with. For the moment, though, I think such a scenario is unlikely. Unlike in the PC world with a dominance of Windows and Intel Centrino chipsets the mobile space is much more diverse which would prevent or at least slow down such a scenario. Nokia with their Series 60 phones might have a good approach to this. No buffer overflows possible as per OS design and software and patches can be pushed to a device Over the Air (OTA) starting with S60 3rd edition.

A Nokia N80 as WLAN Access Point – A Double Blow for Mobile Operators?

Every now and then I go and check which search phrases lead people to my site. Today I stumbled over "Nokia N80 as WLAN access point". Google was nice enough to lead the searcher to an earlier post of mine about the N80 WLAN and SIP capabilities. The person looking for this was probably a bit disappointed as the blog entry did not touch this particular idea. But it looks like nobody else has had it so far, either. I think it’s a pretty cool concept. When traveling in a group, your Nokia N80, E60 or other N- or E-Series phone put in the middle of the table could give Internet access for the notebooks of all people traveling with you.

Linksys has dome something similar with their WRT54G3G Wifi router which has a slot for a 3G card. In the office or at home, the access point serves all local wlan subscribers and offers Internet access via the 3G card. However, it needs a power socket and is bulky, so not the ideal device to set up an instant hotspot for a mobile work group. This is clearly something an N80 or similar device could bring into the game. So come on Nokia, how about putting some Access Point capabilities in your phones?

Opportunity or double blow for mobile operators? Some people will argue that operators might not have been happy in the first place about Nokia putting Wireless LAN capabilities into their phones. They might see this as a way to drive usage and thus revenue out of their networks. Putting access point capabilities into a phone on the other hand drives usage back into the network, but not necessarily additional revenue. But let’s look at it from a different point of view: Many people do not use 3G so far for various reasons. Now just imagine what would happen if you invite your co-workers to use your "3G-Wifi Access point" while they travel with you. Sure, they’d happily accept and see how convenient it is not to have to search for the next wifi hotspot but to have Internet access right when and where you need it. Next time they travel on their own, they might want a 3G card, a 3G phone or a "3G-Wifi Access point" with a subscription of their own. An ideal 3G marketing tool?

Update: Still no S60 solution for this but the story continues here.

Fon (almost) made a smart move in Germany

Today I read a post on Teltarif, a German website, that FON has partnered with a German DSL reseller to offer bundles for prospective Foneros. For those of you who haven’t heard of FON before, they are attempting to build the largest Wifi Hotspot network in the world with the help of private enthusiasts.

I like the idea of bundling FON with a DSL offer, as I think becoming a Fonero must be as easy as possible in order to be successful. At closer inspection, however, the bundle falls short of the "Buy our product and we will deliver a box which you just have to plug into your phone socket at home to become a hotspot". Instead, the offer only bundles a FON account, a Fon Wifi access point and a DSL data subscription provided by Interroute Germany.

What’s missing is the DSL line subscription and a DSL modem which you have to get separately. Apart from being too complicated for the average user who’s thinking about becoming a Fonero because he lives downtown or near a hotel, the hardware lineup is also not reflecting todays DSL landscape. While it was common to have a separate DSL modem and a separate Wifi router a couple of years ago, integrated Wifi routers + DSL modems are the norm these days.

Great first step FON, but now it’s time to put your software on an integrated Wifi router / DSL modem and find a reseller who is willing to do the final step and sell both the DSL data subscription and the DSL line in one package.

Upload Times For Mobile Video Podcasts

If you have seen my previous entry you might have noticed that I am starting to expand my web activities from blogging to podcasting. The next step could be mobile video (pod)casts as my Nokia N70 has excellent video capabilities. With a resolution of approx. 352×288 pixels, mobile video capabilities have advanced far advanced beyond the first stamp size videos of doubtful resolution and quality.

The downside, however, is the amount of data that is generated. A movie of 30 seconds generates around 2 MB of data. With 1 GB flash cards available for less than 30 Euros today, storage space is no issue. For mobile video-casting on the other hand, 2 MB data transfers require quite some time. Here’s a list of upload times for a number of different wireless technologies. Upload times are calculated for a 2 MB video file which is put on a blog or other web site via eMail. eMails tend to increase attachments by at least 50% due to the coding used for attachments which is why the times below are calculated for a total transfer volume of 3 MB:

  • GPRS: 19 minutes, based on an uplink speed of 25 kbit/s (2 timeslots)
  • EDGE: 5.5 minutes, based on an uplink speed of 90 kbit/s (2 timeslots, good uplink quality)
  • UMTS: 7 minutes, based on an uplink speed of of 64 kbit/s
  • UMTS: 3.5 minutes, based on an uplink speed of 128 kbit/s
  • HSDPA: 1.5 minutes, based on an uplink speed of 384 kbit/s (only few networks and mobiles support this uplink speed category today)
  • HSUPA: 45 seconds, based on an uplink speed of 800 kbit/s (no networks support this today but coming soon)

Quite obviously, mobile videocasting only makes sense with EDGE or UMTS, as GPRS is just too slow. What astounds me most is the huge difference in upload times between the oldest technology (GPRS) with 19 minutes and the latest technology (HSUPA) with just 45 seconds. Less than a decade is between these technologies. GPRS was introduced five years ago in 2001. HSUPA is not quite here yet, but expect it late next year or early 2008. Just 7 years between 19 minutes and less than one. Makes me wonder where we will be in 2015…

Podcast: US Wireless Carrier Landscape

As a frequent traveler, I’m amazed that each country seems to have it’s oddities when it comes to mobile carriers. In some countries, mobile voice and mobile data is very cheap while in others just accross the border, the world looks quite different. Debi Jones of MobileJones.com and Mediaslaves has been nice enough to do the first in a series of podcasts discussing the state of wireless in different countries. Debi lives in California, so the US mobile landscape is the center of our discussions:

Part 1:

  • Coexistence of GSM and CDMA in the US
  • HSDPA deployment
  • Closed down CDMA handsets
  • The Nokia vs. Qualcomm battle
  • The Korean CDMA/UMTS situation
  • Price of wireless data

Podcast, MP3, 26 mins, 12 MB

Part 2:

  • Muni Wifi and combination of Cellular and Wifi Data plans
  • Wireless traveling in the US
  • A hotspot detector (check it out at CanaryWireless.com)
  • Wireless Instant Messaging, Yahoo Go!, Skype and Trillian
  • Verizon’s Walled Garden
  • Mobile phone multitasking on Nokia N-Series phones
  • Mobile phone design

Podcast, MP3, 28 mins, 13 MB

No Wireless Killer Applications without a Killer Environment

For quite
some time now everybody in wireless has been trying to find „the“ killer application.
It is hoped that such an application will bring the breakthrough for the
wireless Internet to be as successful as the PC based fixed line Internet we
know today. While reading a book on the history of the Google, I realized that
it took much more than just applications for the big breakthrough of Google and
the Internet in general. In my opinion, even the early Internet offered an
ideal environment for both creators and consumers which in the end triggered its
own mass market success. The same is needed for the wireless Internet. However,
the worlds of the fixed and the wireless Internet could not be more different.
Here’s why:

1. The
Creator Side:
The Internet as we know it today was shaped by creative people,
many of them being students at universities all over the world. Students have
two advantages over people working for companies. First, they are not pressured
by quarterly results and business plans but they can use their time to
investigate and develop whatever comes to their mind and whatever they find
interesting. Second, no business model crosses their mind when being creative.
Many of today’s big Internet companies such as Google, Yahoo and Excite started
off like this. This is not happening for the mobile Internet for a simple
reason: At Universities, access to the Internet has always been free for
students as it is part of the universities’ infrastructures. The mobile
Internet, however, is not freely available to students and academics in most
cases. Hence, not many of them go mobile.

2. The User
Going back 10 to 15 years in time, services that form the core of today’s
Internet such as the world wide web, FTP, HTTP and eMail were used by academics
and students for their research. Search engines such as Google, Excite,
Altavista and others were created mainly to serve this clientele. On the
wireless side however, students and researchers are not the users as the infrastructure
is not provided to them for their work. In effect, this means that a vital part
of the user base is missing. As a consequence, an important part of the feedback
loop is missing that inspires creators to expand and develop new services.

3. The
Service Side:
The Internet became a big success as most services were free for
people at schools and universities to try them and use them over a long period
of time. Services such as search, eMail and web browsing were free, as was the
use of the network. This inspired people to try them. Over time, they became accustomed
to using these services and started to appreciate the added value they brought
to both their professional and private lifes. Again, the way things happen in
the mobile Internet is quite different. While many services are also free, such
as the mobile portals and services of Yahoo, Google, Shozu and others, access to the network
is not. Therefore, potential users are not even tempted to try out these
services as from their point of view, they can’t use the services for free and
in most cases have no idea what it would cost them if they tried.

4. The
Over the past two decades, the personal computer became an integral
tool for students and researchers for both offline applications such as word
processing and calculus tools, as well as for online Internet applications. In
the middle of the 1990’s, computers became cheap enough for home use. At the
same time, people started to see the value of being connected and of being
online and thus they also became willing to spend money for Internet access at
home. Thus, an Internet connection became a logical extension for a PC at home.
At this point the Internet left its free islands, i.e. the Universities, and
became an everyday tool in people’s homes as well. Yet again, things are very
different in the mobile Internet. Here, the mobile phone is the equivalent to
the PC at home. The advantage for the mobile domain is that most people already
have a personal mobile phone which is data capable. On the other hand, the
Internet is not a „natural“ extension of the main use case of a mobile phone,
i.e. voice telephony. Instead, it creates a new range of possibilities which
are not directly linked with the initial purpose of the device. While the
transformation from offline PC usage to online usage was a natural process,
moving from the use of a mobile phone for voice telephony to using it as a
device for Internet data services is a rupture in the evolution which seems to
be hard to overcome.

So where
does that leave us?
It is obvious that even in its early days the Internet was
not free. Somebody had to pay for the computers and the local infrastructure,
and somebody had to pay the telecom companies to build and operate the wide
area networks. Universities are either funded by nations or by tuition fees if
operated privately. While it was accepted that Internet connections are a vital
resource for research and academia, the wireless Internet is still seen as a
luxury good. I whished this attitude would change to create a similar ‘creator –
user feedback loop
‘ to kick start the wireless Internet in a similar way as what
has happened for the fixed line world.  

So where
to take the money from?
Well, I guess it’s already been spent on other things.
Just imagine: Five years ago the German government alone got 50 billion euros
as a result of the 3G frequency auctions. Paid in cash!!! Only 10 percent of
that, 5 billion, spread over 10 years would mean that there would be 500
million euros in Germany
alone each year that could be invested in research and development of wireless
technologies, services and applications. That’s 1000 euros for 500.000 students
to buy hardware and network access every year. I am not even sure if there are
that many students in Germany who’d want to benefit from this. Staggering numbers, just imagine what the
number would be when you count all the license fees paid in Europe and over the world. So what happened to the money? It was used to reduce
the huge budget deficit instead of being partially reinvested into the future.
Well done, German government and others, that’s how we keep our technical