WirelessMoves

The air interface of next generation wireless networks such as WiMAX is based on orthogonal frequency division multiple access (OFDMA). This technology divides a channel into many sub-channels which can be used by several terminals independently at the same time. Fractional frequency re-use, which I describe in more detail below, can reduce interference caused by neighboring base stations using the same frequency block and can thus help to increase user throughput and  overall network capacity.

Due to the limited frequency bands available and the high bandwidth per cell (e.g. 5 or 10 MHz) of future wireless broadband networks, base stations of an operator use the same frequency band. Using the same frequency band for all base stations, however, creates interference for subscriber terminals when they receive signals from more than one base station at a time. This is the case for example for a subscriber located just in the middle between base station A and base station B. If the subscriber listens to base station A, the signal of base station B is seen as unwanted interference. Thus, from the point of view of the subscriber, the output power of base station B should be as low as possible in order to create as little interference as possible. For another subscriber in a similar location but listening to base station B the situation is just the opposite. Thus, a compromise has to be found to adjust the output power of the base stations to a value which enables subscribers at the cell edge to still communicate at a decent speed while not creating too much interference in neighboring cells.

For distant subscribers a base station has to use more transmission power in order to reach them. Subscribers close to the base station on the other hand require much less transmission power to receive the signal. As client devices only transmit and receive on some but not all sub-channels of the frequency band, transmission power of sub-channels used by clients close to a base station can be lower than the transmission power of sub-channels used by clients at the cell edge. In practice, the reduced transmission power for sub-channels used by clients close to a base station thus creates less interference for users close to other base stations.

Using a combination of high and low power sub-channels can be exploited to increase the overall capacity of the network compared to networks which use the same transmission power for all sub-channels. Base stations can be organized in a way to use the same set of sub-channels to serve subscribers close to them with a low transmission power. The rest of the sub-channels are used with a higher transmission power and can be used by both distant and close subscribers. To minimize interference of high power sub-channels for clients of neighboring base stations the cells are further organized in a way that two adjacent cells do not use the same high power sub-channels. Thus, both close and distant clients of a base station will not see the high power sub-channels of a neighboring base station as interference. This approach is known as “fractional frequency re-use” (FFR) as all base stations use the same frequency band, the same low power sub-channels, but only a fraction of the high power sub-channels.

It is worth to note that fractional frequency re-use can not be used in UMTS, as all subscribers use the complete bandwidth instead of sub-channels. Thus, OFDMA networks implementing FFR suffer less from interference problems which translates into higher spectral efficiency compared to UMTS, i.e. the overall bandwidth available in the network is higher.

Currently the only indication the 802.16e WiMAX networks will use this technique is a whitepaper written by the WiMAX forum (changes place frequently so no link given here). I performed a search on the net but found no vendor who is talking about it yet. A bit strange as FFR seems to be an interesting technology.

Further information on next generation wireless networks such as LTE, WiMAX and EVDO Rev C can be found here.

In just about a week the 3GSMWorldCongress 2007 will open its doors for a week of excitement, fun, new ideas, challenges and good discussions in Barcelona. I’ll be there all week enjoying both the congress and all Barcelona has to offer. I am also looking forward to meeting friends again. If you would like to catch up send me an eMail to gsmumts@gmx.de and propose a spot and date. Viva Barcelona!

One of the great things about blogging and book writing is to make the acquaintance of inspired people working on wireless and other hot topics. Due to recent events, let me introduce you to two of them:

Right in time for the release of Windows Vista, "The Unofficial Guide to Windows Vista" co-written by Stuart Mudie, a friend of mine, has hit the book stores. Very well done Stuart, looking forward to buying a copy at the Wiley booth at the 3GSMWorldCongress in Barcelona. Once I am in Paris, I’ll come by to get a dedication!

And while I am already at it, take a look at Gabe Frost’s new blog about Windows Vista. Gabe, working with Microsoft has already written a lot of interesting articles on Vista’s wireless capabilities on the official Microsoft blog which have greatly helped me greatly in my own research. Now it looks like he thought it was time to additionally start a personal blog to discuss his own view on Vista. A treasure chest of tips and tricks around Vista.

Great stuff guys, hope to see you again soon somewhere!

I am sitting in a train right now with some time to spare to write five things about me now that it’s been kindly requested by three people. So here we go:

1. I used a mobile phone for the first time in 1996 (wow, only eleven years ago) while I was living in Los Angeles. At the time I was working for a small computer company. Every now and then I had to go to clients to install some software and hardware. To keep in touch with the ‘mother ship’ I was given a mobile phone. It was already quite small but still one of those analogue cellular phones already long forgotten today. I was amazed at the time but as these things were still quite expensive and I was still a student I didn’t see the immediate benefit for my personal life. I hardly had enough money to buy food and keep the car running (indispensable in L.A.)

2. People in my previous company back in 2000 speculated if I quit my job and went to a new company because I was tired of being ridiculed for having more than one mobile phone lying on my desk (see the radical change from 1996…?). While quite normal today, 6 years ago a mobile phone was still a novelty and having two almost bordered insanity. 

3. Italy and nordic countries hold a special attraction for me. Italy because I like the people, I like the language, the food, the cafe and Don Camillo and Peppone. Not sure what fascinates me about nordic countries, but when I hear Stockholm or Helsinki I feel a desire to go on an adventure immediately. Quite strange as it is quite cold and dark there in winter, two things I don’t really like at all. So where does the attraction come from? No idea, really.

4. I like Opera very much and take every opportunity to go. I already reserved tickets for Don Carlos at the Liceu in Barcelona, where I will be for the upcoming 3GSMWorldCongress. A perfect combination of technology and culture during that week.

5. I’d like to work for the multimedia division of Nokia or S60 one day. In my opinion they are working on the best products for my very mobile lifestyle and working and contributing my ideas to these products should be a lot of fun.

At this point I should name another five people for their turn to tell five things about them. However, most people in my bloggroll have already done so. So  instead I recommend to read Tommi’s five things on his S60 blog, they are truly the strangest and craziest ones I have read so far!

Oh yes, and just as Tommi, ‘I demand my cup full’ as well 😉

Pretty much after every major conference these days people report that the Wifi network was hopelessly overcrowded. One of the possible causes is uplink congestion. Usually, Wifi hotspots are connected to the Internet via a DSL line. Thus, the available uplink bandwidth is much smaller than downlink bandwidth. This becomes a problem when people use VoIP applications or when they upload pictures and other bandwidth demanding content. Such kind of traffic quickly saturates the uplink. A saturated uplink, however, has a big impact on the downlink as well.

Most IP applications use the TCP protocol which acknowledges data received in order to detect missing packets and to throttle the transmission speed. Once the uplink of the Wifi hotspot becomes congested, TCP acknowledgments are delayed because they are queued behind other packets. This has a severe impact on downlink speeds of other users up to the point at which most of the downlink bandwidth is unused due to the congestion in the uplink. The graph in the first figure (click to enlarge) below shows this behavior. At the beginning of the graph you can see a red curve which represents a downlink transmission. After some time, an upload starts and the graph becomes green and yellow. Green represents the uplink speed and yellow the downlink speed. As can be seen the uplink bandwidth is much smaller than the downlink bandwidth. It can also be seen that as soon as the uplink starts the downlink speed is greatly reduced.

The second figure just below shows how the downlink transmission returns to normal once the uplink transmission is over and the TCP acknowledges are sent through the uplink again quickly.

Counter Measures

Some routers are capable of traffic shaping in the uplink direction by preferring TCP acknowledgment frames before any other traffic, i.e. TCP ack frames are not queued behind large uplink frames. This way even a congested uplink does not slow down other downlink connections.

One of the few routers capable of doing this today is AVM’s Fritzbox, a quite inexpensive DSL/Wifi/VoIP router with an incredible feature set. Take a look here for details (sorry, the page is in German, but the graphs are nevertheless quite interesting). The latest firmware release of the Fritzbox also supports the Wireless Multimedia Extension (WMM), a subset of the 802.11e standard which introduces Quality of Service on the wireless interface for VoIP and video streaming. More about WMM can be found here.

I recently visited a friend of mine in Paris who is now working for a N2NSoft, a company with quite an interesting product: Before deploying networks or even before applying for a license, operators need to know how much to spend on the wireless and fixed infrastructure to see if their business model is sound. The cost and number of base stations, routers, fiber links, etc. mainly depends on the number of users, what kind of applications their are using, if they are mobile or not, how good the coverage of the network is, etc. etc. As there are so many variables, simple calculations might give a ballpark estimation of the costs to come. However, this is not good enough for financial calculations.

If one could simulate the user behavior, however, a prediction could be much more accurate. This is just what the company has set out to do with their flagship product called NetScale. Simulating the behavior of a couple of users is simple in terms of computing power. So far, however, processing time of most algorithms used grow exponentially with the number of simulated users. So simulating more than a few users gets tricky very quickly. It looks like N2NSoft has found a solution to this with an algorithm developed at the INRIA, one of the top French research institutions. Instead of computing time growing exponentially, their algorithm scales linearly. This is the dream of any mathematician.

Over lunch, we played with a network simulation of 250 nodes and 50.000 users. Each user was simulated individually and the simulated network had all kinds of different users from stationary notebook usage to moving hand-held eMail retrieval and ring-tone downloads. For each user, the simulation shows the TCP/IP traffic flow in real time, packet losses, retransmissions and overall performance on every link. To see the effect on end user throughput when capacity on a link is reduced or increased is simple as everything is computed in real time. Click on the picture on the upper left, it shows a bit of the GUI and simulation output. More pictures and info can be found here.

Thanks for the interesting lunch break, Pierre, I keep dreaming at night of 50.000 mobiles all accessing the network simultaneously 😉

In the previous entry on this topic I’ve given an overview of how to use a Linksys WRT54G Wireless LAN router OpenWRT, Kismet and Wireshark for Windows for wireless network monitoring and analysis purposes. This blog entry takes a look at the steps necessary to install the required software on the router.

The first step of the process is to buy a wifi router which supports OpenWRT, an alternative and free of charge Linux operating system for wifi routers. The routers supported by OpenWRT can be found here. I decided to buy a Linksys OpenWRT54GL as it offers enough RAM and Flash to run OpenWRT and has enough space for additional software to be installed later on. The price of the router was about 60 euros.

Installing OpenWRT

Once the router arrives, I suggest to use the built in software first to configure it as a Wireless Access point to verify the hardware works as expected. Once done, the router should be reset to run with the original out of the box settings. There are a number of different OpenWRT images depending on the router and router version. This page on the OpenWRT Wiki explains how to select the right image, from where to download it, and how to install it on the router. I chose to install this standard image for the Linksys WRT54G(L). In essence installing the OpenWRT firmware image is very simple by using the ‘update firmaware‘ functionality of the web interface of the original firmware. Be sure to read the installation page linked above for the fine print.

Installing a better Web Interface

After a router restart, open another browser window and access the router again. In a first step, select a new password. Afterwards, configure the router via the web interface so it is able to access the Internet. This is necessary as the software packages which are installed next are directly installed via shell commands on the router.

As the OpenWRT web interface is quite bare I suggest to install X-WRT next, a much improved web interface. This is done via a secure shell (SSH) session. This requires a Telnet/SSH program for Windows. I recommend using Putty.

Important note: Use SSH (port 22) to log into the router, telnet is not supported for security reasons. To log in use ‘root’ as user name and whatever password you’ve initially configured via the web access, even if that password is associated to a different user name.

Once you have shell access to the router, ensure the router can reach the internet by pinging a public IP address (e.g. ‘ping www.google.com‘).

Next, install X-WRT as follows: First ‘ipkg update” has to be typed in to load descriptions for additional software to the router. Once done, ‘ipkg install webif‘ installs X-WRT. Once the command has run successfully, the new web interface is available.

Before proceeding to the next main step one further thing should be done: On the web interface, select Wifi configuration and install the NAS package required for WPA/WPA2 encryption later on. This is  done by pressing the corresponding button on the wifi configuration page. Once done, the wifi setup page looks like shown in the picture on the left at the beginning of this paragraph.

Installing Samba/CIFS package

The next major step is to install a driver on the router to be able to mount a Windows directory. This is needed as the Kismet dump files which contain the intercepted packets can get quite big and there is not enough space on the router to store them. This OpenWRT wiki entry describes the details. In short: ‘ipkg install kmod-cifs‘ downloads the driver. ‘insmod cifs‘ then install the driver into the Linux kernel. Both commands only have to run once as the driver insertion is permanent and survives reboots.

Once done, create a new directory on the router (e.g. ‘mkdir /martin‘) which will be used in a next step as a mount point. Afterwards, share a drive on the Windows PC connected to the router and mount it on the router with a command in the shell that should look like this: ‘mount -t cifs //192.168.70.63/testshare /martin -o unc=\192.168.70.63testshare,ip=192.168.70.63,user=martin,pass=xxxyyyzzz,dom=workgroup‘.

192.168.70.63 is the IP address of the router and ‘testshare’ is the directory I’ve opened for network sharing. Make sure the user name and password given in the command are correct. If you don’t know the domain of your PC it’s probably the default domain name ‘workgroup’. If the command succeeds a new command prompt without a further message appears.

Change the directory to the mount point and verify the share is mounted correctly by putting a couple of files into the Windows directory with the file explorer and by doing an ‘ls -l‘ on the router. They should then be visible on the router.

Note: ‘mount’ doesn’t give very detailed error messages in case something goes wrong. In case you encounter problems, make sure first of all you have set the Windows firewall to allow network and printer sharing on the local LAN.

Note 2: The share has to be remounted after the router has been rebooted.

Installing Kismet

Hold your breath we are almost there now. The final step is to install Kismet. This is done by typing ‘ipkg install kismet-server‘ and ‘ipkg install kismet-client‘. Once done, change into the directory in which the share of the PC is mounted and type ‘kismet_server‘ (note: underline!). Kismet will then start and output a number of messages as shown on the left. Open a second ssh window and start the client by typing ‘kismet_client‘. The client automatically connects to the server and displays an overview of the number of networks found, packets received, etc. On the PC take a look at the shared directory. You should see a number of new files created by Kismet. The most important one is the .dump file which contains all packets received by Kismet. Once enough packets are collected, press ‘CRTL-C’ in the kismet server ssh window to stop the server. Afterwards, rename the ‘.dump’ file into ‘.pcap’. If you’ve already installed Wireshark on the PC you can now double click on the file to open Wireshark to display the captured frames. If you haven’t installed Wireshark so far, download it from here.

Here we go, that’s it! Not quite a solution that works out of the box straight away but with some prior Wifi configuration experience and a bit of Linux experience not to difficult to do. Have fun!

I read an interesting article today about the current power consumption of the worlds mobile networks. The number is quite staggering. According to the article, current mobile networks consume about 43 billion kilowatt hours of energy a year.

If one assumes 2.5 billion mobile users, that would be 1.4 kWh a month or 46 Watt hours a day or 2 Watt of average power per user. To set this into relation: My personal power consumption at home per month is about 200 kWh. Thus, depending on how you look at the 43 billion kWh the number is quite big or small.

Here’s another number from another source that says that German mobile networks currently use about 1 billion kWh per year. Divided by let’s say 70 million subscribers the numbers per user per month match quite well.

So how many power plants do you need for all networks worldwide? 43 billion kWh broken down to average power is 43 billion kWh / (365 days * 24h) roughly equals 5.000 Megawatts = 5 Gigawatts. That’s the energy which is produced by about three to four big nuclear power generators (source)