WTRS Executive Interview

Interview with Eric Janson, Senior Vice President,Worldwide Sales at CSR.

June 8, 2007

George: Can you tell us about yourself and how you came to join CSR? 

Eric: Sure. I started in the semiconductor industry in 1974; 33 years ago. I was a wafer-slinger at my first company. I actually built a wafer fab with two other guys. You could do that with three people back then. I was there for three years and then went on to Analog Devices where I was in a variety of different marketing and PR positions for 11 years. When I left I was the product line manager for linear products. I joined AT&T Microelectronics to help take the Bell Labs treasure trove into the open market, as they were transitioning from captive to open market supplier. I had a very interesting 12 years there, doing everything from gallium arsenide and digital bipolar to complementary bipolar. Everything but CMOS circuits. That included starting up their wireless RF effort for cell phones, which was great fun and included four years in Munich, Germany. And following that it was early 2000 and I heard about this company that was going to do single chip Bluetooth in CMOS. My first reaction was, "They are crazy, they will never do it. There will be gnashing of teeth and wailing." But once I saw the guys at the stand at the 3GSM show, back when it was the GSM World Congress, I recognized them as the core team from a consulting company in the UK that I had known before who were quite clever at doing digits and RF together on CMOS as a single chip. And I thought, "Oh, it's those guys. They can do it." And I was with the company three months later to set up the US operation. I had a few different remits within CSR since setting up the US operation and, at the moment, I live and work in the UK and run worldwide sales and application engineering. So that's the quickie bio. 

George: That does feed right into this next question. You take yourself from 2000, where you were saying they can't possibly do this, to a few years ago where Bluetooth has achieved pretty good market uptake and certainly ubiquity in things like cell phone handsets. Have you been surprised by that growth and the market adoption over that short period of time? 

Eric: I'm not surprised by the market adoption of Bluetooth, or the growth. It was designed from the very beginning to be low cost, and the radio was designed to be very robust, even in noisy environments, and the profile was designed from the start to allow a lot of different devices to interact and exchange voice and data. I think that by many, many standards it's been an extraordinary success. And not long ago, though, if you looked at a lot of the comments in the press saying Bluetooth is dead, and so on and so forth. There was a huge trough of disappointment there for several years that we had to weather before things turned the corner. Of course, we in the industry who were working away at all the multitudes of designs of various companies, in various markets, knew from those activities that it was going to be a smashing success. In all my years of working with a lot of semiconductors in various markets, and a lot of volume ones, I have never seen anything quite like this. But it did fulfill my expectations, actually. I wasn't surprised because I saw the potential in low cost and robust. And as they say, the rest is history, sort of. Bluetooth is still growing significantly and there is still plenty of legs left in Bluetooth. CSR has now shipped more than half a billion devices and it won't be that much longer before we ship our billionth. That's pretty heady stuff when you reset back to the beginning, but we had the belief from day one that it had huge potential, and once we started working on hundreds and hundreds of different designs with customers, we knew it was going to go. 

George: It's interesting. I think that at the point where there was a lot of noise in the press about Bluetooth being dead was the point where we started to see a lot of products being proposed and hitting the market. At the time it was hard to reconcile what we were seeing in the marketplace with what the press was saying about Bluetooth.

Eric: Sure. That's right. And I think every successful technology has gone through the same sort of thing: way over-hyped at the beginning, then a huge trough of disappointment when it took a lot longer than everyone said it would, and then "Oh my gosh, look at what a great market this turned into". That was true of Ethernet and even Internet too. Remember all the Internet seminars there were? And how it was going to be this great thing. Then for a while people were saying that it was dead, not quite as much as they did with Bluetooth, but still people in those early days were saying it was really just a specialty and it will replace library archivists, but nothing more. That all proved not to be correct. And WiFi, when was 802.11 first adopted? 1991? And that really took a long time. But the fact is that the whole industry lies to each other. The bankers and the venture capitalists and such would never fund any of these things if they knew for sure it was going to take as long as it actually does. So they all pretend it's going to happen much, much quicker and I suppose that they get a fair ranking since everyone lies consistently. They get a fair ranking of what the better investment opportunities are and they invest in those portfolios. It's kind of funny.

George: I was thinking it's a systematic error in the portfolio that everybody's able to calibrate out. 

Eric: Exactly, yeah. So  everyone sees this signal which is a big DC value of hype and on top of it a small little AC wiggle of actual truth.

George: WiFi is always the one that I think of in terms of comparison to Bluetooth adoption. It appears that Bluetooth had a much quicker ramp in terms of adoption than WiFi did from the point where we started seeing chips to the point where we started seeing a lot of adoption for it. I think the original perception from Bluetooth was that it was going to be earphones or headsets. Maybe the personal area network got a little bit overwrought, but there has been success now going into so many other applications and it sounds from your earlier comments that some of this was considered almost from the start of developing the technology. What do you actually attribute the success to moving into some of the other segments like consumer electronics and PC Computing applications?

Eric: Bluetooth really was designed as a specification from the start to be low power, low cost, and robust. And thus it was suitable for a lot of handheld and portable things that more power-hungry wireless technologies were not. And because of the profiles that gave the makers of the end units tremendous flexibility in things that Bluetooth could do, that gave the standard the flexibility to grow and adopt. One example is stereo streaming. Bluetooth was designed from the start to have large data packet sizes and small data packet sizes, as well as synchronous connection-oriented channels to carry real-time traffic like voice and audio. And it's actually functioned very, very well at that. The technology has been criticized at times for being much slower than, say, WiFi for example. There is not enough data rate. But actually for 95% of all the applications out there the data rate is just fine. What is really important is the actual throughput the end user sees. So because Bluetooth is very robust, and because there are a variety of packet sizes and types, and differences in the amount of error correction you can turn on. In Bluetooth you can do a channel quality-driven data rate, so in other words you can dynamically change your power and your packets, you actually get a good throughput even in a noisy environment. Whereas with some of the other technologies that are forced into big packet sizes, if you have a collision you lose a lot of data and have to retransmit and your data rate really drops precipitously. And they have big, big packet overheads too. Bluetooth's packet overhead is about 20% or so and some of the packet types are even much lower than that. Whereas for other technologies the packet overhead can be 55, 60% of the entire packet. So the payload percentage is very small. You see when your computer pops up 54 Mbps, well, first of all half of that is gone in terms of payload delivery, and secondly it never ever operates at that anyway. So I think people are fooled into thinking that speed is as important as they think it is. The fact is that two times faster on a box as a check box item at Best Buy works, so that's why they do it. But actually it doesn't make that much difference in the end. So if you are looking at a wireless LAN technology, let's look at 802.11n. Actually, what people have found so much more useful about 'n' is not the speed, it's the range. Range is a big problem. Speed is not. So Bluetooth is really at a sweet spot in terms of size and performance and range and everything else, so it had good attributes on the radio and on the software stack side, it was very versatile which allowed it to be used in many different applications. So Sony Ericsson, then Ericsson, kicked off the whole thing by putting Bluetooth in a lot of cell phones and headsets and right away a lot of people said "oh ok, there's that cell phone. I can synch my PDA data with it. That kind of gets the ball rolling. And I remember the first device that was declared to come out with ZigBee. It was some cell phone from Taiwan or something. That was a very brave move, cause what's it going to connect to? At first there might be nothing to connect to. So you've got to really wait for someone to go out there and make useful devices for the thing to connect to. And hopefully it will turn out to be something that people will want to use your device to communicate with somehow. So OK, great. Now I can use my cell phone to switch on and off my lights, actually I'd find that useful. But for Bluetooth it really got kick-started by Ericsson and a lot of people started jumping on the bandwagon, which was great. So there were a lot of things initially in a short period that you could connect to. And that attracted more people to make more things to connect to.

George: So the Ericsson thing provided that initial market of significant size that then drove the chip prices down to where other applications became more attractive.

Eric: Right. So now we have seen the $5 Bluetooth barrier long since passed. And if you look at the ASP for Bluetooth in a cell phone, in a ROM-coded device, it's now down below $2. And it's become attractive enough price wise that now you see Bluetooth in the Sony Playstation 3, it's in the Nintendo Wii, and all the other wireless gaming stuff is going to come along. I really believe that Bluetooth for mouse and keyboard is going to become a very big market. People don't like wires, they're messy. And we are going to see a lot of music players adopt Bluetooth for stereo streaming headphones. There will be a large potential market for that, and for the stereo headphones themselves. And then of course people do like to use their PC at work to stream internet radio or stream MP3s or CDs or whatever their music library is. And so I think the PC guys will start to say "OK, well there are all those music players out there. Geez, maybe it's time we put stereo streaming into our Bluetooth." These things do snowball at some point.

George: You led very well into the next question. We have noticed that CSR has organized itself to focus on these different, what I call vertical applications. Does this suggest that the idea of convergence of entertainment devices and PC Computing devices is farther out than we think? Or I am almost hearing that you are thinking it may be closer than we think.

Eric: If you are talking about the end device, I am not smart enough to answer that. I wish I were because I would place some stock picks. A lot of consumer companies want to own the central device, or at least soak up parts of other devices. If you look at cell phones, for example, they are rapidly adopting FM radios as standard. So I think that by 2010, certainly more than half of all cell phones in the world will have FM radio as a standard feature. Also, a large percentage will have NFC (near field communication) for payment systems. And they will come along with Bluetooth and some will add WiFi, so that you can do on-campus VOIP at your workplace, then use the thing as a cell phone while driving home, and do VOIP telephony from home through your WiFi access point. So that's one form of converged device. And then there are the home entertainment and infotainment people who have all kinds of ideas for combining everything. The variety of what people call converged devices for home entertainment has been startling, actually. So I am sure that Sony has their own ideas and Dell have their own ideas and the set top box players have their own ideas and who knows how it's going to end up, but the people who make the most progress on that are going to be the ones who offer the simplest user interface. Because that is really a market of grandmothers and non-technical savvy people. There are not enough early adopters to pay the high ticket prices that you will need to completely drive that market. So we will see.

And there are also converged devices on the chip end too. Of course we are sure that wireless connectivity is going to be built into many more and many more things all along, so we are not quite as interested in who wins that war as we are in what the architecture will be of the system and how do we make sure we are well positioned to be chosen for future systems.

George: Our analysis is that Bluetooth has the potential to overlay as a protocol across a wide range of other protocols. At some level the radio becomes almost unimportant, and you have a lot more flexibility in terms of what radios you have down at the physical layer. Do you see this as an important attribute of the protocol itself?

Eric: Oh yes, absolutely. If you look at Bluetooth today, for example, there are really two radios in an EDR (enhanced device radio) device, there is the FSK simple radio at one Mbps cheap-and-cheerful one, and then there is the phase-modulated 3 Mbps one. So they are already using two radios today and when Bluetooth 3.0 comes out, we will add Ultrawideband as a third radio-on-chip. Basically, the Bluetooth chip will look at other Bluetooth chips and by exchanging the 'what are you capable of doing' packet, with the other one (service discovery in other words), it will know the other one has Ultrawideband as well. And when it comes to an audio file like a headset, if you are doing a voice con, it would be silly to use Ultrawideband because actually it's fairly high powered compared to Bluetooth, and there is no advantage to the data rate for that application. But if the user says ok I have a cell phone with a 40 gig microdrive in it because it also has a 6 megapixel camera and an MP3 player and I want to synch this with my home server because I haven't been home in two weeks. There is a Bluetooth device on my home server and my cell phone will say "oh, big file transfer. Open up the big Ultrawideband fat pipe and shoot this data across very, very quickly". So the flexibility of the protocol stack in choosing which radio to use for what operation is essential, obviously. Once you have built in that flexibility and adaptability you can adopt other physical radio schemes like WiMedia, which we will handle with our Bluetooth chips. So the question was a very good one and very insightful. The radios per se are necessary but not sufficient to make a good wireless standard. They have to go hand-in-hand with a good protocol capability.  We are covering most short- range wireless standards and you will see varying levels of functional combination based on what our customers need in their end markets.

George: Do you have any brief thoughts or anything you can tell us about CSR the company or how you see the company moving forward?

Eric: We were founded with a remark from our CTO, James Collier, who stated that the aim of the company is to democratize wireless. And when you look at that slightly deeper, in order to democratize something, what he means by that is to make it so cheap that everyone can afford it. And make it so easy to use, that there is no reason not to use it. So I think we are doing pretty well on the cheap side, and we are getting closer and closer all the time on the ease-of-use side. Of course the actual end device is really what determines the user interface, but we are doing a lot inside the chip to enable better UIs (user interfaces), easier pairing, trying to make things more natural and intuitive to use and apply so even then, there is a lot you do in your protocol stack to enable that. And so we don't see any quick end to wireless connectivity at all. We see more and more standards being combined into a single chip. For example, we have shipped hundreds of millions of Bluetooth chips with on-chip DSP, which is used for crunching audio codecs and all kinds of stuff.  We are shipping Bluetooth chips with FM radios. We are going to be shipping combo chips with WiFi, GPS, FM, et cetera, et cetera, et cetera. So we will keep cramming more and more functionality into single-chip solutions, and help make the end products more and more available to various wireless technologies, and help them adapt better to whatever environment they happen to find themselves in. I think that there is a very, very long-term growth path for wireless and so for the long battery life variant of Bluetooth that Nokia is promoting called WiBree, and I think it's likely to be adopted by the SIG and offered as part of its technology portfolio. So that will get away from the stigma of it being a Nokia standard. People are saying that if they can make the battery last as long as you say with WiBree, then I can put WiBree in my digital watch and when you get a phone call, and you are wearing your Bluetooth headset, you just look at your watch and the caller ID will pop up there. So you don't have to dig your phone out of your pocket and look at the phone's display to see whether or not you want to answer the call. You will just look at your wrist, which you are used to doing hundreds of times a day without thinking about it. So people are thinking of all kinds of creative ways to apply the new technology variants that will fit the circumstances of their end product. So the tradeoffs will always be with us on speed, size, power, cost, et cetera, et cetera, et cetera. But there will be a lot more flexibility and adaptability built into the combo chips for devices that can afford that flexibility and connectivity. And there will be huge potential markets for tiny little stand-alone specialist chips built into watches, and people are talking about building them into sneakers and things. Although I rue the day when my sneaker starts telling me that "oh, actually you should be walking one more mile today Eric". The amount of ideas in the marketplace for applying wireless technologies is very, very rich and I think it will remain so for many years to come.

George: I was just thinking it's OK for your sneaker to be telling you that, but it's not OK for your sneaker to be telling your doctor, "Oh he should have been walking an extra mile a day".

Eric: And debiting my bank account wirelessly at the same time.

George: Just to finish up, were there any questions that we should have asked that we didn't?

Eric: Gosh, it all depends on what sort of report you wanted to write. I didn't want to presume to do that. But you didn't ask some of the classic ones about Bluetooth versus this technology and Bluetooth versus that technology, but you guys are pretty savvy in terms of understanding that really there is more cooperation than competition anyway.

George: That's what I was sort of getting at with the question about Bluetooth providing a platform to put these other technologies into devices. 

Eric: Yeah, and they are all going to merge anyway, so it kind of doesn't matter.

George: Yeah.

Eric:

First you have dumb integration, and everyone is doing this. A good example is WiFi plus Bluetooth chips. They don't share anything. They don't share a single circuit. They don't share a single pad. So its just sticking the two together on a piece of silicon, and doesn't really buy you very much, but it is an integration step and it will force the chip makers to solve the issues of getting those two radios working in close proximity, so there will be value in that for future developments that aren't so dumb. But in terms of the wireless standards, we think there is plenty of room for all and they all have their own set of tradeoffs and their own set of end products and applications that they are good at. So I think the more, the merrier.

More information about CSR here...

This interview ran in our June 8, 2007 newsletter issue.