Snapdragons: Where Scorpions Hide

Last week's announcement of the Snapdragon platform by Qualcomm went barely noticed. Looks like the majority of QCOM audience is focused on the royalty tension with Nokia. This is as expected. News is always driven by emotions, and emotions in turn, are driven by the FUD (Fear, Uncertainty, Doubt) factor. Often it is difficult to spot the real gems, and the Snapdragon seems to be the real gem.

We have talked about convergence here many times. You probably can feel it coming. Even the cheapest mobile phones offer games and music playback. The more advanced ones have email readers and cameras, while the top of the line models can be used as personal navigation devices, either with built in GPS receiver or with an external "Bluetooth GPS Antenna". But at the same time when phones become converged teleputers, there are plenty of other devices screaming for connectivity. iPods, Zunes, PSPs, GameBoys, Nuvis, digital cameras... Just show me any electronic gizmo that is not built around a microprocessor and that wouldn't benefit from some kind of digital radio communications, be it cellular (HSDPA/HSUPA included), WiFi, or Bluetooth. That is what the Snapdragon is about. It is a chip with a powerful, yet very energy - efficient microprocessor (the Scorpion) and a palette of all mobile connectivity options. It can be programmed to decode MP3's (to be the heart of a next generation wireless iPod) or can be programmed to run games (being the heart of the next generation GameBoy) or can be programmed to operate a digital camera (with wireless photo sharing), thanks to on board CMOS sensor interface (up to 12 Megapixels). Suddenly the market for Qualcomm chipsets is multiplied several times.

Along with the announcement of the Snapdragon, Qualcomm announced Samsung as the first customer for the platform. I wonder if this means we are finally going to see a usable Origami - class device - a light web tablet with universal connectivity options, and 24hr battery life?

Skype over 3G: official

Hutchinson, one of the leading mobile network providers, known as "3" across Europe, and mostly in the United Kingdom announced officially unlimited Skype calls for a flat fee. Skype is just a piece of a broader package named X-Series. This is kind of a breakthrough. Hutchinson probably sees Skype invading mobile phones inevitable, so they decided to jump in as soon as possible to be the leader of that trend. Initially there will be two top of the line handsets supporting X-Series services, Nokia N73 and SonyEricsson W950i.

It will be very interesting to watch it work (or fail). People (myself included) have already been talking for years about Skype on mobiles. But it has not happened so far. Why? There are several very fundamental technical issues giving Skype very hard time. Let us go through them.

• Latency. This is the killer number one. People have been dreaming of Skype over GPRS or EDGE. I even happened to use such setup once. I mean I unintentionally left Skype running on my phone after doing some tests over WiFi, went out, and Skype still maintained my Online status over GPRS when somebody tried to Skype me. I saw him calling I could even pick up the conversation, but we just could not talk.
  Typical real life throughput of 20-40kbps over GPRS/EDGE combined with packet latency in the range of 1000ms or more is just too much for any VoIP to be useful. UMTS is so-so, I mean you can talk and hear the other party, but the voice quality is less than stellar, far from being comfortable, so for typical calls it is too inconvenient to use. It may be just right with HSDPA in low congestion areas, but hey, just right means the level of quality close to 100-year plain old phone conversation...
• Processing power. Skype codecs are extremely processor intensive. Probably this was on Niklas Zennstrom's mind when he said putting Skype on Symbian was challenging. Skype shines on notebooks and desktop PCs, but these machines have at least ten times the power of a mobile device. When you go down to something like a 200MHz ARM-based system, Skype just chokes. It could even bog down a 500MHz Xscale system, so while on a Skype call you could hardly do anything else like checking the calendar or taking notes. I wonder how it really performs on the N73 or W950i. Is it just "I can connect and I think I know what you say" quality or is it something more?
• Handset integration. I hope they have finally cleaned the issue with headphone / speakerphone modes. All the versions I have tried, used only the speakerphone. Kind of strange, especially when you run this on the handset and hold close to your ear. Then when I was on a Skype conversation, traditional mobile calls were coming in as I was free. This situation is very difficult to handle and its proper solution requires a presence-aware mobile network. As far as I know there is not a single one...

There are other winds blowing into faces of mobile Skype users as well. So operators should not panic yet. But with more advanced 3,5G networks coming and more powerful handsets on the horizon (Scorpion based), Skype may become the carrier of choice, turning the current operators into dumb pipe providers. I say "may", as there is a lot of work for them to do, especially in fine-tuning the handset application. Hutchinson believes it will come, so they decided to start gaining experience. Smart move. They will not loose a penny with this initiative and the mobile VoIP experience they will get is priceless.

Disconnected UMTS

UMTS is probably the most common form of a so called 3G (third generation) mobile network. The 3G buzz has been aloud for almost six years now. The outrageous money spent by the operators on the spectrum licenses in early 2000's still weighs on the operator's balance sheets. The promise of lightning fast mobile networks is still around. But how is reality? I mean the average mobile phone user. Does she/he care?

There are three advantages promised to be delivered with the ascent of UMTS. First and the most spoken about has been the fast packet data transmission. Has it happened? It surely has. Last week I was stuck in a traffic jam in rainy Warsaw (btw Warsaw just stands still when it rains... I don't know why, but this is 100% reproducible). So sitting in a car I was checking mail on my Teleputer. One of my important customers asked me to forward him a product presentation. So I turned on my notebook, found the 4MB PowerPoint file, hooked the notebook to the UMTS Teleputer over Bluetooth and started sending the reply email with the 4MB presentation attached. It took just a few minutes and the file was at the customer's desk. Later I did a quick calculation and found out the actual uplink speed must have been in the region of 250-300kbps. At least ten times as fast as with GPRS or EDGE. I have been in similar situations a number of times and have to say UMTS delivers on its promise here (as long as you are within the coverage). But then I wonder how many people really want to upload a 4MB PowerPoint from a taxi cab...

Second advantage of 3G/UMTS has been for the operators themselves. The technology behind (W-CDMA or Wideband Code Division Multiple Access) simply allows for more people talking over the network. 2G/GSM networks in crowded cities are running at the edge of their capacity. They simply cannot accommodate more subscribers. With UMTS there is room. So operators do not have to stress us with high peak-time rates. They have enough capacity, so they let us talk. Even encourage us to talk more, by lowering the rates. Do we care? Yes, as long as the rates are lower indeed and as long as we do not get into network congestion that manifests itself by not being able to make a call.

Third advantage has been the ability to have video calls. Almost every UMTS handset has a forward facing video camera and we can actually see each other when talking. Personally I do not find this feature very exciting for person to person calls. This is a great technology for person to machine calls (imagine video calling your bank and just see the menus, account balances on the screen instead of having to listen to the long and boring voice prompts).

But what about voice? Has something changed here with the advent of all the latest technology and after mountains of money have been spent? I do not want to generalize here, but just from my experience - the voice part of all this glamour has been just terrible. Or even worse. I have been using several 3G/UMTS handsets (Nokia 6630, Nokia 6680, iMate Jasjar, HTC TyTN to name just a few). And they all share the same, terrible voice experience. All the handsets are multi-mode, they log on to the UMTS base stations when signal is available and they fall back to GSM otherwise. And whenever I am within the range of a 3G base station, there is at least 50% chance the callers will not reach me. Why? I have no idea. This must be some really big problem with the network itself. But this phenomenon is very predictable. Why do I know that? Because I keep on getting MCNs (Missed Call Notifications) - SMSes telling me somebody tried to reach me but failed. So I am on line (as I get those MCNs) but when you call me, there is at least 50% chance you will hear an announcement telling you I am out of the network service area or my phone is off. I get those MCNs, call back and the call is completed. Well fine, thanks God we (or the operators) have the MCN systems in place (those who do not -my company has a very good MCN solution, so come and ask), as in the end we can talk and operators recover the lost revenue (a lot!).

Why is that? Well, to be honest I don't know, I can just speculate. What I have noticed, is the missed calls happen most of the time when I am at the edge of 3G/UMTS coverage. If you think this is not common, you are wrong. UMTS (unfortunately) works at 2100MHz, and this frequency has very poor ability to penetrate buildings. So whenever you are indoors, you are probably at the edge. And when you are at the edge, your phone tries to switch back to GSM. Then you walk close to the window and the 3G/UMTS reception is better and the handset switches back to UMTS. And it is somewhere within that back and forth switching process that the incoming calls are lost. Not nice. Not nice at all. I suppose the transition to UMTS would generate a huge drop in call completion ratio, and probably missed call notification systems rescue the operator's voice revenue. I have noticed another UMTS phenomenon that proves the 3G to 2G (UMTS to GSM) handoff does not work as expected. I live outside the city and on my way home there is no 3G coverage yet. Very often I drive home having a phone conversation. The conversation starts on a 3G (UMTS) network and as I leave the city it is handed off to the 2G (GSM) network. I even know the point where it happens. Why? Because most of the time the conversation breaks down at that point. And if it does not, all I can hear is a modem-like sound instead of my conversation party. The modem-like sounds is a 3G-encoded voice (AMR codec) being directly transmitted to the 2G network and down to my phone working in 2G (GSM mode). GSM requires voice to be encoded with GSM-compatible codec, and AMR is not one of them. The network obviously knows I am on the 2G part, yet it still feeds me with 3G-encoded voice.

All this proves 3G/UMTS is still in its infancy. Errors like 3G to 2G handoffs should not simply happen. Everybody seems to be so fascinated with data transmission capabilities, they simply forget what people use their phones for and what service (hint: voice!) still pays the bill. In the end this all will be sorted out, I believe, but in the meantime, it is good we have MCN systems to save us.

Presence: Mobile Killer App

Like a thousand years ago, we are still in search of The Holy Grail. Regardless of the industry sector, we keep on dreaming about the killer application that will make all the customers happy and all the suppliers rich. Of course in mobile communications sector everybody is doing the same, since the conception of the first mobile network. We are not talking here about client applications running on smartphones, focusing rather on the applications running on the network, available to all mobile subscribers, regardless of the type of the handset they use.

I think the first one has been the voice mail. Before mobile, we used to have answering machines at home. Especially in the US, everybody used to have one. But then you could not attach it to the mobile phone, so it had to be hosted on the network, and this way started to be available to all subscribes in form of a service (VAS or Value Added Service). Up to this day voice mail has been the most profitable VAS hosted by mobile operators. The second one has been SMS (Short Message Service). Initially not very popular in America, it is still hitting the new highs across Asia and especially Europe. While price per single SMS message keeps on dropping, SMS revenues are going up and up due to constant growth in volume (even US is waking up...). Then entertainment services emerged, placing personalized ringbacktones on the podium within the most profitable value added services.

So what is next? Are there any potential applications on the horizon to rival the voice mail? Or to displace SMSes or make personalized ringbacktones obsolete? Hah! This is a million dollar question and I think I may have an answer...

I wonder how many of you actually use traditional voicemail? How many voice messages a day do you retrieve from your mailbox? Do you have a personalized greeting? Or it is just there because you don't know how to turn the thing off...? My observation is people are giving up their voicemail accounts and turn to network notification services. For me it is far more convenient to get an SMS telling me somebody tried to reach me instead of my voicemail calling to tell me I have one message (that happens to be a hang-up tone most of the times). Or get an SMS telling me the person I tried to reach is now online and within the reach of the network.

And here is the hint. I call you. You have your phone turned off (for whatever reason). You turn it on and I get the SMS that you are back. So I can call you. Very nice service (BTW my company has implemented this recently in Poland). But isn't it a little "too automatic" (may be turning your phone on you does not exactly mean "call me now")? So here we come to the final: let users manage their presence status from their phones. Presence status you say? Yeah, we know that. Everybody uses an Internet communicator, be it Google Talk or Messenger or Skype... And presence management is the very basic feature on the Internet. Busy, at the meeting, away, available... we know all that. So my question here is why isn't it as "basic" on the mobile phone networks, as it is on the Internet? Do you know why? I do not... But I do know I would use a service like that. To be able to set a status with a simple press of a button on any phone. And have it announced automatically to anybody trying to reach me, when I cannot take a call (be it a meeting or vacation or out of the network coverage). And to be able to query statuses of my buddies - Internet style.

Presence management is extremely underutilized nowadays. And it is completely not converged. When I change my Google Talk status to busy, so should change my Skype and my mobile phone. It can be done. In fact we have already done that. BTW is this an opportunity for operators? You bet! Wouldn't you pay 50 cents a month to have your presence and availability synchronized across all the communications channels you use, mobile phone included? I would.