CIE Components in Electronics
Nigel Forrester, European head of development for embedded computing at Emerson Network Power, talks to Steve Rogerson in our series of interviews for CIEonline. In his current role with the embedded computing business of Embedded Network Power, Forrester is responsible for market development of embedded computer products across market sectors.
Since joining Emerson with its acquisition of Motorola”s Embedded Computing Group in January 2008, he has held various product and vertical marketing roles. He is based in the UK but reports into the embedded computing headquarters in Phoenix, Arizona, USA.
Forrester”s entire 30-plus year career has been within the electronics industry sector. A published author and speaker, he is particularly interested in the application and uses of embedded computing building blocks. He has a degree in computer science and statistics from Reading University.
His main interest outside work centres around vehicles, not just driving them but he is in the process of building his own kit car. He is also teaching his sons to drive although they”re not old enough to go on the public road. They’ve been members of the Under 17 Car Club for a couple of years, which he says has been a lot of fun as well as educational.
1. Intel says there will be15 billion connected devices by 2015. Is that achievable or desirable, and how will it happen?
I don’t know how we will ever connect them, but we are heading in that direction. I have even seen bigger numbers predicted. My children are always connected to various gadgets. It is a different world now. When I was at school, my dad was a bit worried about my interest in a career in computers because he thought mainframes was where it would end. But you look at it now and every few months you are connected to something else, and I am not a gadget person. I don’t buy gadgets for the sake of it. I don’t have a connected house, though I see that a lot of that will have uses. I don’t control my lighting and heating remotely, and I don’t really want to.
We do have a weather station that connects wirelessly and all the bits and pieces that my lads use. So, in a way, it is a connected environment that we live in. And this speed of change is going to get faster.
Whether all this is desirable is a more interesting question. There are aspects that are desirable, but from a Big Brother perspective I am not a fan of the government following my every move, which they can do with all the CCTV. That can be good from a law and order view but it can also be less desirable.
A lot of the M2M applications can be good. For example, remotely monitoring water levels and activating sluice gates can be good. The electricity supply can be tailored to demand. If they remotely change your heating or air conditioning by a degree, you wouldn’t notice and that could stop the supply going off altogether.
There are a lot of issues with security when we get these 15 billion connected devices. How do you make sure your devices are secure when they are connected? That is the big issue in embedded. When your train or gas pipeline is connected to the cloud, how do you make it secure?
2. How different is life under Emerson to what it was under Motorola?
From a personal point of view, not terribly. I do a similar role and we do similar things. The embedded industry is a long term, long life, stable industry. We have seen some changes in that Emerson’s strategy is to be a diverse company looking at different markets whereas Motorola was more focussed on selling into the telecommunications space. We still sell to telcos but part of my role is to search for new markets for the products. It is a different strategy and it is a more interesting approach because you get to see more environments. It is more challenging because you have to be cognisant with all the acronyms and so on. They all have different standards and buzzwords that you need to look at. But it is always interesting to find how people can use these products in different ways.
3. It has been a decade since Advanced TCA was ratified (December 2002). What has been its impact in the past ten years?
The promise of ATCA was that it would be an open standard for carrier grade telecoms applications. It would be big enough and have enough cooling for the telcos. It has delivered on that promise.
Before that, we were selling Compact PCI for that market and that has been largely superseded by ATCA. And ATCA has moved into other markets that Compact PCI didn’t because you couldn’t get enough electronics on there. There was also a number of proprietary designs and that has changed with ATCA. It has become the de facto standard. There is still proprietary stuff out there but very much less because of the cost of developing not just the hardware but the software and middleware. The big manufacturers had 10 to 15 platforms but now they just use ATCA, which is also making progress in the adjacent spaces, such as for ruggedised servers in the military. I am thinking of ship-based servers where they want something a bit more robust than a standard server. These will survive the US barge test where they get one working on a boat and then drop an explosive charge into the water nearby. ATCA comes through that quite favourably. There are markets where you would never envisage ATCA that are now using it.
And like all good standards it has evolved. It is keeping pace and I think there is quite a lot more to come. ATCA is used all around the world; it is very much a global standard.
4. Why are you building a car?
I’ve had a life-long passion for vehicles. I always thought it would be cool to build my own. I am building a kit car. Some kits are very easy and you can put it together in a weekend. But this one is more tricky. I bought the body panels as a kit. I then bought a scrapped car and stripped down the engine and gearbox and rebuilt them. But now I’ve got to the difficult bits such as the interior and that involves tricky thing like sewing.
I’ve been building this car for the past 15 years; I have too many other distractions. I will finish. I’ve not very far to go. I just need to finish the interior. I have some seats but they need to be re-covered. My wife says she’ll help but she’ll need an industrial sewing machine. I’ll probably hire one.
I really like mechanical and electrical engineering type things. I like to find out how things work. I like to take things apart and put them back together. I like looking at how other people have designed and made things.
5. What will be the main changes in the embedded boards industry in the next 20 to 30 years?
We have seen the industry go through different phases. The embedded board industry started off as a niche market. It was quite specialised engineering. You had to be an engineer or computer scientist to use them. They were quite difficult. You were using Assembler rather than a high-level language. Even getting an LED to light up required quite a bit of skill.
The technology has become more complex but they are easier to use. You have high-level languages and they are x86 compatible. You can use the same development tools and debugging tools as in the PC world. This means there is a massive pool of engineers.
If we now look at the current growth in tablets and smart phones, these are not x86 based for the most part. These tend to be more specialised. The whole thing has gone full circle. But getting this technology into the embedded market and making them long life is tricky. The consumer devices are short life but the companies in the embedded space still want seven to 20 years of active life. There is a big gap.
So I think the embedded board market will become more niche again. Look at medical. It takes a year to get a medical device certified. A commercial airline is expected to fly for 20 to 40 years and doesn’t expect to upgrade equipment in the first ten years. The railway industry expects equipment to be available for 20 years.
We have to divorce the two sectors. It is a compromise. If you want a piece of equipment to last five to ten years you have to accept that it won’t be state of the art. Yet they always have to leverage the cheaper elements from the consumer industry, such as memory.
Arm has dramatically improved in performance and has the advantage of low power consumption. But you can run a piece of old x86 software on the latest silicon and it will still run.
