WirelessMoves

Another Mobile World Congress has come to an end. Thursday is usually a bit strange as everyone is alreday quite tired and just waits for the closing bell to ring. To my surprise, my Thursday was quite different.

In the morning, I met with Ajit Jaokar of Open Gardens / Futuretext to prepare for our upcoming course on LTE Services at Oxford University. Lots of great energy flowing in that session and I went out in a very positive mood. More about that soon in a separate post.

Next I went to the Next Generation Mobile Network Presentation stream to listen to some network centric presentations. Julius Robson reported on the progress of the LTE/SAE Trial Initiative, which I already reported on in one of the first MWC posts. Great presentation, Julius, it’s good to see LTE making progress.

China Mobile had a presentation on common FDD and TDD chipsets for LTE, so like me, they also think that a common chipset, i.e. mobile devices being able to do both FDD and TDD are vital for TDD to become a success.

As usual there is also the odd presentation from a North American carrier bashing on the vendors and the users alike, calling European network operators stupid for their extensive 3G network rollouts and sale of prepaid SIM cards for voice and data. I wonder if they will ever wake up!?

Good that most other operators have quite a different view on that and out of the many other presentations I especially liked Seizo Onoe’s presentation for NTT DoCoMo on Deployment and Realities of LTE. The most interesting chart for me was how DoCoMo plans to deploy LTE. It seems that they already have a deployment in which the UMTS radio module sits on the mast and is connected via a fiber cable to the digital module of the UMTS base station. For LTE, they will reuse the existing antennas and hook their digital LTE module into the fiber link that is already in place. Quite cost effective. In the Q&A the presenter was asked by a representative of T-Mobile (who also push very hard for LTE and don’t want to invest much more into the HSPA network) what he thought of HSPA+. Seizo Onoe seems to be one of the “fathers” of LTE, if one can say so, so he was of course not quite in favour of HSPA+ and said with a smile on his face that he doesn’t like it, it makes things too complicated in the mobile devices and one should go to LTE straight away. I had to smile as well, that was brought over very well. But no matter, HSPA+ will go into handsets anyway, that’s my opinion.

On into the afternoon, I had some more meetings with companies doing testing equipment and I learnt a great deal about how the optic interface between the LTE radio module and the digital part, called CPRI, will be used by maeasurement and testing companies to simulate mobile devices to stress test the base stations and the network in general.

And finally, I went to Wiley’s again to pick up a book about wireless to read on the way home and to say good bye. More about that in a separate blog post as well.

There we go, another Mobile World Congress has closed its gates. Not the most splendid one I have seen but I hugely enjoyed it again, have learnt many new things and spoke to lots of people enthusiastic about what they are doing. What else can one ask for?

It’s Thursday morning and before I head of for the final day of the Mobile World Congress, here’s what happened yesterday.

In the morning I went to see a number of companies doing measurement and testing equipment for LTE networks. While in UMTS, a lot of info can be extracted from the Iub interface between the base station and the radio network controler, this is no longer possible with LTE as all functionality has been put into the base station. So other innovative ideas are required to analyze and debug the radio access network in the future. I’ve seen lots of interesting approaches and I am sure the hearts of many engineers while beat faster when they get one of these new testers into their hands.

The highlight for me on Wednesday was of course my author session at the Wiley booth in hall 2 to promote the launch of my new book. Lots of people came by to say hello, to have a chat, to see what is going on and to pick up a copy. At the end of the session we were sold out. Thanks to all who stopped by it was very nice meeting you and to see the different angles you have on the industry.

Later in the afternoon I went to see TV moderator Guy Daniels at the TelecomTV studio to see what he thinks of this years MWC and the state of the industry. It was good to see them being in Barcelona again in full force pumping out lots of great interviews that one could watch live, over one of the huge video walls at the Fira, and as a live stream on the Internet on the TelecomTV.com homepage. This year, the program was also broadcast via DVB-H, so people lucky enough to have a Nokia N96 or other DVB-H capable handset could watch this and a number of other channels for free. I also had a chance to get a backstage look at the TV equipment used for producing the shows, including the 10 Terrabyte disk array that stores the week’s production.

One thing I didn’t get to see was a demo of Nokia’s Home Control hard and software which seems to have been cancled at the last moment. A bit dissapointing, I was very much looking forward to that.

After the Symbian Foundation Party on Monday and the great dinner reception of Ip.Access on Tuesday I didn’t really get enough sleep so I opted for something more quiet and with less people last night. The food was good and the amount of sleep a bit more so there is enough energy left for the final day at the Fira.

You might have noticed I have not written anything about new mobile phones having been announced this weeks. Well, they have just not really been my prime focus. Too many other things to see here that can’t easily be seen on the web. Any suggestions from your side which mobiles I should look at before the Congress closes this year?

Day 3 at the Mobile World Congress and my impression is that it is a little bit less busy than in the days before. People also start to be a bit tired, long hours at the congress and activites in the evening leave their marks. Anyway, so here are some of the things I have seen today:

Last year it was only Nortel and LG who where showing a live and over the air LTE demo. This year, I was expecting more companies to do the same. To my surprise this was not the case. Hall 8 is full of LTE demos but no real over the air and no real mobile devices or prototpyes are shown except for LG and Nortel. What’s going on!? I am puzzled. Good to see that the size of the mobile device prototypes have shrunk from desktop PC size down to about twice the size of a typical 3G dongle today. At least at the LG booth. These were no mockups but were used for the demos.

The GSMA and Ericsson showed the evolution of HSPA in hall 2 and I was shown an HSPA live air demo with a throughput of 14 MBit/s with a 3G USB stick with a Telstra branding. It said 21 on the top which probably stands for 21 MBit/s, which I think is the highest speed with 64 QAM without MIMO. I couldn’t see where the antenna was hidden but it was probably quite nearby. In all fairness one has to say that at the Congress pretty much all antennas are nearby to give excellent coverage as reported earlier.

Another thing that I checked out in more details is who is doing the different layers of the protocol stack. I know that the area is very fractured and lots of smaller companies are working in this space. I talked a bit with a representative of MimoOn who are doing layer 1 to 3 LTE protocol stacks for chipsets from PicoChip and TI. They had good demos what each layer of the protocol stack does and how they can be separated. While some companies lincense their complete stack, others implement some of the layers on their own and just want one or the other from them. Pretty much a mix and match game. So for a product like a femtocell, silicon of vendor x is used, the protocol stack from y and z and the femotcell vendor then might put the rest of the protocol stack on top itself plus all the femto applications and there are probably partnerships for that as well. And finally the femto cells get sold to an even bigger company which then sells it under their own brand. Incredible how many companies are involved.

And finally for this post, I’ve spoken to a couple of people about dual mode TDD / FDD handsets and if they think this will happen. All of them said they weren’t sure and that it is a quite complicated thing to do. They pretty much all said “Let’s see how things develop”.

Network capacity and access to devices in the home from mobile devices are going to be important features in the near future. Femtocells are an interesting solution to these challenges and ip.access has shown some interesting solutions in the past on other tradeshows which I could only watch so far on Youtube. So at this year’s MWC it was great to finally meet some people I’ve been in contact with at ip.access over time for the first time and to take a look at their femto solution first hand.

It’s so simple these days to get a demo up and running at the Mobile World Congress compared to the days when you needed E-1’s and other non standard links to equipment located somewhere else. Today, a simple Internet link does the job, inexpensive even at trade shows compared to dedicated E-1 lines. Ip.access made good use of it and had a number of live femtocells at their booth, connected to their femto controller back in Cambridge in the UK via an Internet connection in exactly the same way one would do at home. Interesting to see that they have partnered with Cisco for their first femto deployments, that’s certainly going to raise awareness.

In retrospect it seems a bit odd that I didn’t ask them if I could do a voice call on their femto, but I just believe that voice calls work, it’s the core feature. Instead, I was much more interested how mobile devices can be included in the home network via the femto cell to interact with the video and music library, the television, act as a remote control, etc. I’ve seen it in a Youtube video before but seeing it live is always different. They wouldn’t exactly tell me how the mobile is included in the local network, thus bypassing the operators network, but after seeing the demo I think I have a fair idea. Thanks for that, very interesting!

In addition, they also showed an interesting Facebook plugin where users can leave messages for a femto user which are delivered by SMS when the user’s mobile finds the femto cell at home or at the office and performs a location update procedure. Great integration of social networking. The scenario can be expanded to things like getting a message when the children arrive at home after school or automated reminders to water the plants, something that would help me a great deal, as my plants never survive very long.

We continued the conversation in the evening at Casa Batlo at their dinner party and it was great to meet more people working on the technology and share their enthusiasm for their product and wireless in general. Thanks very much for that! During the evening, I was asked when in my opinion femtos will be launched for the mass market. Obviously, the first standalone femtos are very close to market readiness, the Cisco femtos at the booth are a clear indication for that. For a real mass audience, I think it will take a bit more as I see an integrated Femto/Wifi/DSL/Home Gateway box with applications such as IPTV and streaming server as an ideal platform to attract interest from consumers beyond voice coverage enhancements.

So, what do you think?

When you are used that a mobile network consisting of lots of boxes in a huge data center it's a strange feeling when you go to a measurement and testing company like Rohde and Schwarz and you find a whole network from the base station antenna to the servers in the core network condensed in a single box sitting on the table. I've seen devices like this in the past in some labs but never had the time to take a closer look. This year at the Mobile World Congress, I decided to pay Rohde and Schwarz a visit to have a closer look.

The level of detail that can be analyzed is breathtaking. It starts at the physical layer where the tester can veryify that the mobile to be tested modulates the signal correctly and goes up to messaging to the UMTS SGSN or LTE SGW for things like network attach, bearer setup, etc. The tester can even run of the shelf server programs like VLC so video streaming and other sorts of throuput tests can be made with real applications on top.

New network technologies such as LTE are developed step by step, usually starting at the physical layer. So far, I always wondered how one could already test some layers when the software for others has not yet been written. After the demo at the booth I have a much better idea of how that works in practice. So in the first stages, when only layer 1 and parts of layer 2 exists, you pretend on both sides that a bearer, IP address, etc. have previously been negotiated and set those to fixed values. After that you just pump IP packets over the interface to test mobile behavior, modulation, use of sub-channels, etc. Everyhting automated with testcases against the 3GPP test specifications of course. As the protocol stack evolves, more and more components can be included in the test.

Another thing I discovered at the R&S booth was a test environment for network based A-GPS. While the Nokia A-GPS in my N95 connects to a Nokia server on the Internet to download ephemeris information to kick-start the GPRS chip, there is also a network based variant to download the information from the base station. I was not aware that this was going forward but since R&S has a tester that simulates it and a public mobile that uses it, some operators and mobile manufacturers seem to implement it. I need to have a closer look and dig a bit once I am back from MWC.

And finally, I was also shown their mobile testers to debug live networks. Depending on what a network operator or other interested party is prepared to spend, they have mobiles with test and monitoring software (e.g. on an N95) up to boxes that record the complete UMTS band and decode a carrier down to the physical layer. Makes my mouth water.

Thanks for the excellent tour, I thoroughly enjoyed it!

From a network perspective, LTE is of course the topic of the Mobile World Congress. This morning I met with Julius Robson of Nortel, who is chairman of the LTE/SAE Trial Initiative, or LSTI for short. I’ve been following the activities of the LSTI for some time now as their aim is not to repeat the mistakes made with UMTS in the past by not properly testing networks and mobile devices with each other before pushing them out into the market.

For the congress first results from the proof of concept phase are becoming available from a lot of vendors and I will report more on the details in the coming weeks. Here are some personal insights I gained during the meeting:

I’ve been a bit sceptical so far of the chances of TDD-LTE (Time Division Duplex) becoming a big technology outside China as most network operators in Europe, Asia and the Americas mostly have FDD (Frequency Divison Duplex) spectrum. To my surprise, it seems there are more than just a few network and mobile vendors who have reported results back to the LSTI on first TDD-LTE tests. They still seem to be a bit behind FDD-LTE but it’s not the tiny trickle I was expecting. Julius told me Nortel has worked on TDD-LTE as well and since everything above the physical layer is pretty much the same in FDD and TDD, only a few people were required to come up with the additional base station software for TDD in quite a short time. Quite a difference to TDD-UMTS which is a completely different game to FDD-UMTS used today.

Also, some mobile operators, for example in Europe, have TDD spectrum in the 2.1 GHz band that they bought together with the FDD spectrum during the UMTS auctions back in the year 2000. So far that invstment lies dormant. So if mobile devices do come out early that support both FDD and TDD than I think there might actually be a chance for TDD to be deployed alongside FDD in major markets outside China. Whether it’s as easy on the mobile side to have FDD and TDD combined in one chipset is another matter and I don’t have a lot of insight into this. Maybe I’ll find out more this afternoon as I’ll talk with some silicon manufacturers then.

In LTE, there are only two radio interface states, ACTIVE and IDLE. The LTE requirements say switching between the two states should take less than 100 ms. Looks like while this sounds ambitious from a HSPA perspective, were the switching time currently is in the range of one second, the goal seems to be reached by vendors even in the early stages. That’s good news as it makes the tradeoff very small between interactivity and power conservation from the mobile device point of view. So it will be interesting in the future to observe if Discontinuous Transmission (DTX) will be used a lot while in ACTIVE state to conserve energy or if the mobile is just set into IDLE state right away.

Time passes very quickly here at the Mobile World Congress and it’s already morning of day 2. Here are some thoughts I scribbled into my notebook yesterday which you might find interesting, too.

Apart from the N97, I’ve also taken a look at a couple of other things while at the Nokia booth. The beta version of Nokia Maps has been out for a while and after looking at it at the booth, I think it is time to upgrade on my N95 and wait for the half a gigabyte of map data to reload in the process. New features that have persuaded me are speed warnings, radar warnings, finally a navigation route overview option and, most importantly, integration of maps with Ovi on the PC, so I can plan routes and favorites on the PC and synch them with the mobile. A very nice feature.

Sportstracker has been upgraded as well and now supports Polar heartbeat monitors. Currently only available as a bundle with a new phone but with some luck at some point also available separately. I noticed that the heart rate of the presenter was higher than mine 🙂

Next on my list of things to do was to pick up the tickets for the Symbian Foundation evening event at their stand and to chat with some of the employees of where things are going. It seems things are still very much in flux and the next year will be an exciting one for those involved with the foundation.

I also went to see my book publisher Wiley in hall 2 (2A130) to prepare for my book presentation session on Wednesday afternoon from 2-4. The book is already selling well and I had the chance to talk with someone just picking up a copy while I was there. It’s always good to talk with people to see why they are picking up a copy. If you are around on Wednesday, please come around and say hello!

As I said, time flies and once that was done I was already on my way to the MobileMonday Global Peer Awards. As the event is quite big it was outside the Fira as every year, this time in the beautiful Palau de la Musica. The venue was not quite perfect for the event as the conference room was underground and there was absolutely no public network coverage. Some people, including me, were quite nervous not being reachable. It must be the times… Rudy de Waele was nice enough to give me instructions how to log on to the installed Wi-Fi network so I could at least send out e-mails that I am not reachable. Lots of people were at the event and I think the format to give each startup three minutes to present themselves was a good thing. Great event, Rudy, I guess you have done magic again for networking and getting the talk flowing. One thing I particularly noticed compared to previous years was that for startup companies the iPhone seems to have overtaken S60 as THE platform to develop for. Not sure an application store and an N97 will be enough for Nokia to reverse this trend. But they are certainly not sitting still so we’ll see what happens.

And finally in the evening we chilled out at the Symbian Foundation Party. As always lots of interesting people there to talk with, mobile and otherwise. A great first day and not much energy left afterwards.

Day 2 has just started now, let’s see what it brings.

During the day I ran a number of traffic tests on the public 3G network of Vodafone on the Mobile World Congress exhibition ground and I have to say I am very impressed by the results.

Even around high time with the main hall packed with people communicating, performance with my speed test application was around 700 kbit/s with peaks at 1.2 MBit/s. And this with all the people around me communicating and showing mobile Internet applications. One presenter at the Nokia booth even told me that she has switched to 3G for her demos as the Wi-Fi network they have installed couldn’t cope for one reason or the other. Now that’s something…

On the technical side my trace mobile shows that Vodafone is using three 5 MHz HSPA carriers. If the other three network operators do the same, there should be 12 carriers or 60 MHz of bandwidth used at the Fira for 3G. Even in the busy Barcelona city center currently only one carrier is used by Vodafone.

The tracer also detected only a single neighboring cell on one of the three carriers used, so neighbor cell interference must have been low, another must for getting high throughput. It’s probably a micro cell setup, although I have to walk around a bit more on the exhibition ground to confirm that.

GPRS performance in Vodafone’s 2G network was equally good with Opera Mini downloading web pages very quickly. A good indication that enough timeslots were free for data use.

On the voice side: All calls I made today mostly to other people at the exhibition connected very quickly and I never got a network busy response or people complaining I was not reachable.

I don’t think there are many other places in the world with wireless data and voice traffic as high as at the Congress except maybe Manhattan. So it’s good to see that despite the increasing use for Internet connectivity, well planned and deployed networks can do the job.

Here we go, day one at the Mobile World Congress and one of the first places for me to check out is of course the Nokia booth. While lots of people have reported about the new E75 and other devices having been launched this morning, I’ve been looking at some details of other already announced phones and applications which are of particular interest to me.

So everyone speaks about “Touch these days” and my next phone also needs to be one, too, no doubt about it. So I played around with a Nokia N97 for a bit to get a feel for the device and the touch screen. I really like the form factor of the device and also the sliding mechanism which tilts the screen while in the open position. Not only is it good for working with the phone while holding it but the screen can also very well be read when the phone sits on the desk. No complaints about the small keyboard either, I could type quite quickly with two fingers. With the new idle screen widgets it makes for an ideal central place to keep me updated on news, e-mails, social networking events and other bits and pieces that come in throughout the day.

I also very much like the finish of the device and the display resistance, it makes scrolling through web pages very easy and natural. The experience is much better than on the 5800 Express Music on which I find the screen surface a bit to resistive when quickly moving the finger over it to scroll down a page.

On the software side I think Nokia still has some work to do. Some things still feel a bit slow and other things such as the Gallery integration into other programs such as e-mail was still missing in the demo device they showed.

The other thing that I would like Nokia to improve a bit before launch is the built in browser. It doesn’t (yet?) reflow text and so on some pages the text is either tiny or one has to scroll left and right when zooming in. Also, scrolling does not work while the page is processed, something that I found quite irritating. Definitely not an easy thing to get working properly but I think it’s essential.

So all in all still a bit of mixed feelings on the N97 for the moment but with some luck the firmware that the first devices will be shipped with deal with these things. Can’t wait to get one then.

When I get on a new mobile network I usually give it a thorough check for my research and to discuss results with interested operators. Most already do application testing on their own such as how many videos are viewed without interruption or how fast web pages load, etc. However, some do overlook some other important parameters.

One of the first things I noticed in Vodafone’s 3G network here in Spain at the Mobile World Congress is that sometimes when I log into the network with my mobile phone and remain connected, the battery runs flat within only 2 hours. A bit of research revealed that the reason for this is that they are assigning public IP addresses for transparent Internet connections (via the APN airtelnet.es).

So you might wonder what the battery drain has to do with a public IP address assignment!? Well in theory, not much, but in practice, quite a lot. When getting assinged a public IP address, all IP packets to this address get routed to the mobile device no matter whether they are wanted or not. Since IP addresses are dynamically assigned the new holder of an IP address potentially gets the ‘leftover’ from a previous user and all the junk emitted from viruses and other malware on the Internet. While the mobile device doesn’t really care and discards those incoming packets, the radio link is constantly active instead of falling back to a power conserving state while no user data is exchanged.

In my case I noticed that in some cases when I attached to the network the air interface link was constantly kept up as every couple of seconds an unsolicited IP packet was received. Most operators use network address translation (NAT), which assings a private IP address to the mobile and thus fixes this issue.

A somewhat crude fix for this problem from the user side short of hooking up the mobile to a power source once an hour is to fall back on the 2G GPRS network. The unwanted IP packets still come in but the power consumption in 2G mode is significantly less as the air interface link is kept for a much shorter time than in 3G mode after the last IP packet has been received and also requires less power.

So Vodafone, while I appreciate a public IP address for my 3G notebook dongle I think it’s a particularly bad idea for battery driven devices and user statisfaction. Also from a network point of view this is far from ideal as it wastes significant ressources on the radio link that would better be used for real traffic.

PS: I’ve noticed that there is also an APN with private IP addresses but that seems to be only good for web surfing as all my other applications are blocked.