The Paradox of Open

Yesterday, Ajit Jaokar over at Open Gardens published a thought piece he called "The Paradox of Open: What we can learn on Open from Apple and Microsoft". He argues that despite Apple and Microsoft having closed mobile operating systems they are successful. The conclusion he reached was the following:

"I believe that: A 'closed platform' works provided you have an 'Open ecosystem' BUT an Open platform (open source and / or open standards) without an ecosystem (open or closed) does not work."

An interesting statement I agree with and would like to extend a bit to show further influences:

A couple of days ago I met with a friend who's had an iPod touch for some time now and is a glowing part-time developer. When I asked him why he picked up programming for the iPhone / iPod touch and what he thought about the development environment, he said that:

  • he picked it up because Apple is doing great things
  • because he likes the UI and how easy it makes things for users.
  • the development environment for the iPhone is not so nice with the license you have to get and all the restrictions that are put in place with developer codes, distribution, etc.
  • he only works on it because the product on which the application will run is so great. Otherwise he would not bother.

So I think in the end it doesn't really matter if an OS is open or closed. If it is not liked by users for whatever reason and is not easy to use and there is a better alternative available, then nothing will come of it.

Speaking of ecosystems: I think for a free and open source OS it is much simpler to get an ecosystem in place because the community can help vs. a closed source OS where the burden of fostering an ecosystem lies in the hands of the company that owns it.

A good example is Nokia's Memo platform for their tablet devices. It's for the most part free and open source and yet, only few users have bought one. In my opinion the handling and UI is just nowhere near the iPhone or similar touch devices yet.

In other words: Part of the ecosystem, a part that one can't touch, is the success of the product, its ease of use and in turn how many people use it. So Open alone is not warranting success on its own. In that respect I am not sure if there is a paradox with Open? As always, comments are welcome.

LTE TDD in China and Europe?

Even though LTE is a global standard there are two different air interface flavors: The first one is FDD and pretty much destined to be used around the world. And then there is TDD, for the moment mostly foreseen to be used in China. However, many European operators have acquired TDD spectrum in the 2.1 GHz band during the 3G auctions back in 2000. So while I haven't heard anyone talking about this so far, I wonder if LTE TDD might be something of interest to European carriers!? In that regard, does anyone know if the TDD band in China is anywhere near the TDD region that was auctioned in Europe? Might be a great push for dual mode FDD and TDD devices.

How to Explain the Thoughts Behind BICN

Bicn-stack There are lots of things in this world that don’t make a lot of sense unless you know how they have evolved to their current state. One of those things is the migration of circuit switched telephony to the IP world with the Bearer Independent Core Network concept specified in ITU Q.1901 and introduced in GSM and UMTS starting with 3GPP Release 4.

Here’s my take at it:

With BICN, the circuit matrix of the MSC (Mobile Switching Center) that creates a physical voice circuit between two subscribers is replaced by a media gateway. The media gateway maps the concept of a circuit connection to an IP stream between two parties. The stream is then transmitted together with many other streams over a shared packet switched link, which is for example based on Ethernet.

The Signaling System Number 7 (SS-7) used in the circuit switched world is still used in this architecture but has been changed to some degree. The protocols of this family are used for the following tasks:

  • For the establishment of voice calls
  • For the interaction between different network nodes (e.g. between the switching center and the subscriber database node)
  • For communication between the switching center, the radio network and the mobile device

The main difference with SS-7 over IP is that the lower layer MTP protocols have been replaced by IP, SCTP and M3UA, so the messages can be transported over IP instead of a circuit switched timeslot. The figure on the left shows the MTP based SS-7 protocol stack in comparison to the IP based SS-7 protocol stack.

Above the MTP layers, the ISUP protocol that is used for establishing voice circuits has been replaced by BICC (Bearer Independent Call Control). BICC is very similar to ISUP. Message names are the same and only some parameters have been changed as the messages are now used to control media streams instead of circuit connections.

Protocols for the interaction between different network nodes (e.g. between the voice switch and the subscriber database) such as SCCP, TCAP and MAP have not been altered at all. DTAP (Direct Transfer Application Part), the protocol used for interaction between a mobile device and the switching center for purposes such as authentication, location updates, etc., has also remained unaltered. In other words, applications that use these protocols are not aware if the messages are transported over signaling timeslots of a circuit switched network or over an IP link.

To enable IP based SS-7 nodes to communicate with MTP based nodes in the network, Signaling Gateways are used to translate MTP into IP / SCTP / M3UA. This way, a traditional circuit switched MSC is able to communicate with a subscriber database node that is connected to the network over an IP connection.

And finally, from a mobile point of view, the air interface between the base station and the mobile device also remains unchanged. This means that GSM and UMTS mobiles have no visibility what kind of access or core network technology is used.

Today, both traditional circuit switching and BICN can be found in live networks so knowing only one of them won’t do, at least for the moment. So I’ve decided to coin two terms:

  • Traditional circuit switching”, i.e. the origins of circuit switching with voice calls transported over physical circuits and SS-7 messages being transported over circuit timeslots.
  • Virtual circuit switching over IP”, i.e. a voice channel is transported over an IP stream and the SS-7 protocol is used in a modified form in the IP world.


Traditional circuit switching vs. virtual circuit switching over IP. Do the terms make sense to you?

GPRS Detached, Attached and a PDP Context

Over at Boy Genius Report there's been a post yesterday about how to keep a data connection alive on S60 devices over some time of inactivity. Unfortunately, the conclusions drawn there are technically not quite correct. So, as there is probably some interest in the community of how always-on connections work over GPRS and UMTS from a lower layer point of view, I've put together a summary to show what is really going on.

GPRS Detached

When you power up an S60 mobile for the first time, it is usually configured to only establish what is referred to as a "packet data connection" when needed. The term "packet data connection" however, is grossly misleading. When set this way, the mobile will only attach to the voice part of the network but will not make itself known to the GPRS packet switched side of the network. An S60 device shows this state, referred to as GPRS detached, by showing a little antenna symbol below the signal strength bars at the top left of the screen.

GPRS Attached

Screenshot0055 When the "packet data connection" option is set to "when available", the mobile performs a GPRS attach procedure as soon as it is switched on. This means that the mobile in addition to becoming known to the voice part of the network also becomes known to the GPRS part of the network. An S60 device shows this state to the user by showing two little dashed horizontal arrows below the signal strength bars. If the network supports EDGE, a little E is shown in addition above the arrows. This is shown in the picture on the left. If UMTS is supported, 3G is displayed above the arrows. In this state, and this is the crucial point here, the mobile device is NOT connected to the Internet, i.e. it has no IP address. Hence, the user is not charged and it has nothing to do with always-on connectivity.

So what's the use of this state? Not a lot actually. In the past, it was envisaged that applications from the Internet could trigger an IP connection being established to the mobile device when packets for it are received. However, to the best of my knowledge, no operator has ever made use of this functionality as it is pretty much obsolete due to the real always-on connectivity described just below. It's only real benefits are that the user sees on the display if EDGE is available or not, and that Internet connections are established faster in a subsequent step because the mobile is already registered.

In terms of power consumption, being GPRS attached should not require any more power than if one only attaches to GPRS when establishing an Internet connection if the network operator has configured the network accordingly. In practice, many network operators these days use an interface between the circuit and packet switched part of the network (the Gs interface between the MSC and the SGSN) so the mobile registers only once to the network when it is powered on and not to each part of the network individually. Also, periodic location update timers are set to the same value for the circuit and packet switched part so there is no additional signaling involved for being GPRS attached.

PDP Context Activated

Once an application wants to access the Internet for the first time, an IP address is requested from the network. This procedure is referred to as "PDP (Packet Data Protocol) context activation". Once an IP address has been acquired, the horizontal arrows at the top left of the screen changes from dashed to solid. The IP connection stays in place until one of two things happen. When the user closes an application that has requested an IP connection and no other applications are still running that use the connection as well, the IP address is released and the two arrows go back to their dashed state (GPRS Attached). But until this happens, the IP address remains assigned even if the application doesn't transfer any data. There is no technical limit of how long an IP address can stay assigned without exchanging data, it can persist for days, weeks or years. In practice, however, some (but not all) operators reset all IP connections once every 24h or after some time of inactivity. There are also operators who don't have their network properly configured so it can also happen that the IP connection is dropped when moving through the network and neighbor relationships are not properly set or if a core network component (SGSN or GGSN) malfunctions and is rebooted. However, those are exceptions.

A little Mystery

In the picture on the left there is a second option in that GPRS Attached/Detached configuration menu which is called "Access Point". I am a little bit puzzled what this option could be good for as it doesn't change the behavior of my N95 no matter whether it is set to "none" per default, to an APN or to a profile name.

Summary

The summary of this somewhat long text is that the description in the Boy Genius post is not accurate. The post suggest that by changing these options, the mobile will be switched to an always-on state and that the data connection will be kept. The only thing these options do is toggle between GPRS Detached and GPRS Attached state. The assignment of an IP address (PDP context activation) is a separate procedure, completely independent of these settings.

As a consequence, the "Packet Data Connection" option does not make anything better or worse if one has an unlimited data plan, or if one is billed per kilobyte or by the minute. Since it only speeds up getting an IP address when the first application that uses a connection requests data for the very first time, the setting should be the same for all kinds of data plans: "when available".

For more details, great places to look are the GPRS service description stage 2 (3GPP TS 23.060) or the GPRS introduction Chapter 2 in my book on mobile networks.

The Current LTE Spectrum Situation

I was recently asked by a friend how I see the current LTE spectrum assignments. I would have liked to give a simple answer but it is actually not quite straight forward. This is how I see it at the moment:

Europe:

  • 2.1 GHz band, used for 3G today but still a lot of unused capacity: Most likely the first band where 5 MHz LTE carriers will be deployed. No limitations from regulators, LTE can be deployed straight away.
  • 2.5 GHz spectrum auctions still outstanding in most countries. Good for going beyond 5 MHz carriers
  • 900 MHz band: Maybe some deployments of 5 MHz carriers or less. Good for in-house coverage but the band is heavily used for GSM today so it's difficult to clean up enough space for a meaningful LTE deployment without running into congestion issues. It might get better as more people get 3G phones and some of voice and data traffic currently running over GSM in the 900 MHz band will start flowing over 3G in the 2.1 GHz band. That reduces the load, hence, it might be possible for carriers to clear some spectrum for LTE, if they haven't opted for UMTS 900 deployment, which is available today.
  • 800 MHz band (Digital Dividend): Strong push in Europe at the moment to free the same bandwidth in all member countries. First trials have started to bring high speed Internet to rural areas with HSPA and LTE.

North America:

  • Verizon will deploy LTE in the 700 MHz band and has a single 10 MHz carrier available. That's not much. The spectrum has been assigned, so it can be deployed straight away.

Japan:

  • According to the presentation of NTT DoCoMo at the Mobile World Congress in Barcelona this year, they will also deploy LTE in the 2.1 GHz band, replacing one of their UMTS carriers with LTE at the beginning.

China:

  • Not sure

Unfortunately, it does not end here, 3GPP has lots of additional frequency bands defined for LTE. So fracturization is likely to increase. As always, comments, additions, etc. are very welcome!

Operator QoS for Skype & Co.

Recently, Nokia has announced that they will integrate Skype into the Nokia N97. Reactions, obviously, have been mixed. But I think the trend is difficult to stop, if not on this device it will be on another or in another way entirely. Some network operators have responded by announcing that they are thinking about introducing special tariffs which would include VoIP. But there is one thing over the top VoIP (i.e. non-operator circuit switched voice) doesn't have today, and that is the possibility to ensure the quality of service (i.e. latency, delay and jitter) especially over the air interface.

However, with a bit of imagination it wouldn't be too difficult to set this up. Here's one example of how it could work: In tariffs that take VoIP into account, the network could establish a secondary PDP context (UMTS) or a dedicated bearer (LTE) when it detects IP traffic of VoIP applications. This prioritizes the voice IP packets over other IP packets in the data stream of the user and also over IP packets of other users. Most mobile network operators already have deep packet inspection devices in their networks for all sorts of things and these could easily do the job.

I think it's an interesting technical possibility, let's see if somebody picks it up and puts it into commercial reality.

LTE – The First Global Cellular Standard – But Does it Matter?

Indeed on first thought, LTE will be the first global cellular standard in the future to which GSM, UMTS/HSPA, CDMA and potentially other cellular wireless technologies are likely to converge on. But does it really change anything?

Being a global standard does not necessarily mean all LTE capable devices can communicate with all networks around the globe. There are two main issues:

1) FDD and TDD Mode

While in most parts of the world, FDD (Frequency Division Duplex) will be the dominant air interface technology, TDD (Time Division Duplex) is pushed especially by China as an upgrade path for TD-SCDMA. So an FDD LTE device will not be able to use a TDD network and vice versa. With some luck, we might see devices that can do both FDD and TDD but nobody's really commenting on how feasible this really is. Only time will tell.

2) Two Dozen Different Frequency Bands

What's worse is the number of frequency bands are foreseen for LTE. In practice, this will mean that devices will be built for some but not all of those frequency bands. So it's nice to have a global standard but it's unlikely the mobile devices themselves will be usable on a global scale. The single 4G device working everywhere will remain a nice dream.

A Little Light At The End Of The Tunnel For Vendors

LTE being a global standard is a good thing for network equipment vendors. Most of the equipment will be the same including the base stations where only a few parts or modules are different to work on a different frequency band or operating mode (TDD/FDD).

Benefits For Network Operators

An economy of scale is created for networks operating on the
main LTE frequency bands (e.g. 900, 1800, 2100 and 2600 MHz). Most other frequency bands are only used by a few network operators so it's unlikely these will get the same prices from network vendors as their colleagues who use the mainstream bands. Also, the number of devices working outside the standard LTE bands are likely to be as limited as for UMTS/HSPA today. Just have a look of how many 3G devices are available for HSPA network in the U.S. compared to Europe.

Benefits For Users

For the users, I don't see a big change from the situation with HSPA today. Where the mainstream frequencies are used, there is a big choice of devices and this is likely to be the same with LTE. And network operators using less used frequency ranges will probably receive as few devices as those operating 3G network in such bands today.

What We Really Need

So what we really need is not only a global standard but also global frequency bands so everyone benefits the same. But, unfortunately, that's a dream that is very unlikely to come true anytime soon.

Voice – Bearer Aware, Bearer Adaptive or Bearer Agnostic?

It seems I am not the only one thinking quite positivity about Voice over LTE via Generic Access Network (VOLGA). Recently, Ajit Jaokar posted an interesting article in which he mentions that with VOLGA, the traditional circuit switched voice service becomes a bearer aware application, as it can choose between a 2G circuit bearer, a 3G circuit bearer and an IP based bearer over LTE. All seamlessly with handovers during the call with all bells and whistles attached!

An interesting way to look at it even more so as the bearer awareness does not come into play on the mobile device but actually in the network. This is because the controlling entity for the voice call, the mobile switching center (MSC), sits in the network and is informed by the network that a different bearer should be selected. It can then decide to go along, arrange for the network to prepare the handover and then instructs the device to make the jump.

So maybe VOLGA makes voice even more than bearer aware!? So far the term 'bearer aware' has mostly been used for applications being aware what kind of networks are available at a time and then make a choice as to which IP network to use or to stay put in case a network is available but the cost attached to it is too high to make the application feasible.

In the case of voice, however, the service can ensure continuity by jumping from one bearer to another. So terms like 'bearer adaptive' or maybe even 'bearer agnostic'  come to my mind, because that voice call will just work over any kind of network the device supports.

It could even work over the Wi-Fi you have at home if you extend the idea of VOLGA. Not for the moment, as the standard currently focuses on LTE, but in the future, who knows?

Satellite Internet on Thalys High Speed Trains – A Report

Thalys-Internet It's great when two high speed technologies come together: High speed trains running at over 300 km/h and high speed Internet access. Thalys, whose trains travel between Paris, Brussels, Amsterdam and Cologne has equipped all of their trains with satellite, Wi-Fi, UMTS and GPRS based high speed Internet access, accessible to passengers via standard Wi-Fi. When I recently traveled on one of those trains, I could hardly wait to get on board to test and use the system.

The picture on the left shows the satellite antenna installation on top of one of the coaches. It looks a bit odd on the otherwise very streamlined train but the round shape probably keeps the additional drag to a minimum. Nevertheless, I'd be interested in finding out how much extra energy is necessary to  push the train beyond 300 km/h due to that.

At 7 a.m. in the morning, throughput in both uplink and downlink between Paris and Brussels was tremendous. Speedtest.net reported downlink speeds of more than 10 MBit/s and more than 3 MBit/s in uplink direction! While the link dropped a number of times on the trip to Brussels it was only for a few seconds each so that was probably only apparent to an attentive observer like me running a data trickle in the background to detect just such occurrences. However, the outages were short enough that it didn't affect streaming applications once enough data was buffered. Watching a Youtube video, full screen and in HD quality worked just fine.

As all data is transferred via a satellite in a geostationary orbit, round trip delay times were in the region of 650 ms. While voice calls and even Skype video calls work well over the system the delay can be felt in the conversation. Loading a graphics intensive web page works quite well and fast but it feels a bit sluggish for a moment after clicking on a link or entering a web address before the download of the page starts. This is again due to the very high round trip delay time compared to other systems such as ADSL with a round trip delay time of 50 ms, or the 120 ms over a 3G connection. Having said all that, the experience is still great, especially taking into account that the countryside is passing by at 300+ km/h when looking out of the window while that HD video is streamed over the satellite.

The satellite connection has one real several imitation: In Europe, the geostationary satellite hangs close to the horizon, so it is not always possible during the trip to keep the connection. In such cases a ground based backup is used. In the Brussels main station, for example, Wi-Fi is (probably) used. Downlink speeds came close to 16 MBit/s and round trip delay times were lower than 50 ms. The tunnels around Brussels were covered as well, although I was not sure exactly what technology was used. In other places, especially in the hilly terrain between Belgium and Germany, the satellite connection doesn't work too well either, probably because the train winds its way through narrow valleys and many tunnels. GPRS and UMTS network in that region seem to be patchy at best so the experience on that part of the track wasn't too great.

In between I should also mention that I didn't find any services that were blocked. VoIP worked well, IM worked well and my IPSec based VPN also worked fine over the system.

In the evening I made the same trip in the other direction. It seems a lot more people were using the system in the evening as speeds were much slower than in the early morning. While I could still reach fantastic transmission peaks of 3 MBit/s in downlink and more than 1 MBit/s in uplink, I experienced continuous high packet loss and frequent connection outages in the range of minutes even on the flat terrain between Brussels and Paris. The bad weather and heavy rain might also have had something to do with it, it's difficult to tell from a single ride.

Summary:

Of course I had my expectations before trying the system. In most cases I found it to be much faster than I expected. Especially the main applications such as web browsing, e-mail and VPN tunneling to the company network worked fine. The system has its limitations in hilly areas and cities when there is no direct line of sight to the satellite. While the system automatically switches to GPRS or UMTS in such cases, it didn't work particularly well in many of those places areas, as they were probably not covered very well. It can work much better over 3G as I have experienced here. Overall, however, I was very impressed with the system and I think it's a great service!

Can One Deduce From Chipset Specs How Future Devices will Look Like?

…I was asked today. A clear opinion here: Yes and no.

Yes: When I reviewed some of the future chipsets for my recent book it was clear we are moving to processor speeds beyond 600 Mhz, built in camera hardware units of those chipsets supported resolutions of 10-12 megapixels and a touch panel interface was also part of the unit.

No: Such specs tell you nothing about the form factor of a future device. Examples: Just knowing that there is a touch interface tells you nothing of how usable the interface will be. A chip spec doesn't tell you the physical characteristics of the device, e.g. like will it have a hardware QWERTZ keypad, etc. Also, there are usually supporting chips around the chipset like GPS, motion sensors, compass, etc. How they are mixed and matched is not on the datasheet either.

One interesting domain I haven't yet too much looked into is the specs for the radio front end chips. This info would be very interesting to get an insight which technologies (GSM, UMTS, LTE, CDMA) can be supported in future devices, and, equally important, which frequency bands can be handled with a single front end chip. If you have some good references here, please consider leaving a comment.