Archive for September, 2011

The needle in the haystack

While the long-established oscilloscope companies manufacture multiple ranges of multi-purpose instruments as well as application specific ones and high-end systems, new players need to focus in order to create a niche. This is even true for experienced companies with expertise in other test and measurement areas.

Lighting: Powerful with less power

The controversial end of the incandescent light bulb holds no terror – at least for anyone who forgets about the ‘energy-saving lamp’ based on fluorescent tubes. These are obviously not satisfactory, but LED-based light sources are as everyone who tried them happily confirms. A research initiative now intends to minimize the already low power consumption of the semiconductor lamps.

Biceps training

With its newest acquisition, IAR Systems certainly strengthens its tool portfolio, especially for ARM processors. However, since there is a huge technological overlap between IAR and Signum, one can assume that this is not the only reason for that take-over.

Taking the reins

According to a study of 32/64-Bit Microcontrollers, Embedded Microprocessors and DSPs conducted by Semicast, ARM became the leading architecture for 32/64-bit MCUs/eMPUs in 2010, ahead of x64/x86 and Power Architecture. For the period 2010 to 2016 the analysts even predict the highest growth for ARM cores. Now Intel tries to take back the reins.

Kontron in ARMs

The (not really surprising) news of Kontron ‘evaluating’ ARM processor technology has created some hubbub in the blogosphere. A lot of it is based on misconceptions and half-knowledge.

The most obvious is about ‘Kontron’s first ARM’. Nope. Kontron has been working with ARM technology for years – am I the only one to remember XScale before Intel sold its PXA business to Marvell? The news is that processors don’t come from Intel exclusively anymore. Well, that’s business.

Not so obvious but nevertheless a strange notion by anyone who’s in the electronics business for more than a month is the inability to differentiate between ‘embedded’, ‘deeply embedded’ and ‘industrial PC’. Dear colleagues out there: these are NOT the same thing. The emphasis in ‘industrial PC’ lies on ‘PC’. While ARM technology came from the ‘embedded’ world, then moved into deeply embedded space and is now seeking its way upwards into the (industrial) PC area, Intel x86 architecture is the top dog in (industrial) PCs. This is certainly not going to change anytime soon, even if industrial PC majors like Kontron (or, by the way, MSC since this years’ embedded world show) offer ARM-based products, too. Since IPCs essentially are PCs, x86 may not be the best architecture (well, was it ever?) but the most compatible and widespread. The Kontron experts make it clear themselves: “Based on industry feedback, Kontron is evaluating the ARM processor technology and is currently reviewing a range of embedded SOC ARM processors. Use case examples may include, but not limited to, board-level products, fanless systems, enclosed HMIs, etc, applicable for example in the industrial, medical and signage market.”

This may be the rare case of a statement both bold and cautious but it definitely doesn’t mean “Goodbye Intel, don’t call me, I’ll call you” as some self-proclaimed experts like to think. The fact that (in Kontron’s words) “application developers are looking for reduced development costs, long-term availability, improved power consumption and lower TCO” is neither new nor surprising. The historically unusual thing is that anyone is convinced that those benefits, apart from the power issue, can be realized with ARM processor cores.

This is a consequence of ARM’s rather clever product and third-party strategy over the last few years. The Keil acquisition is a fairly good example, but also the Cortex products that opened up whole new markets.

Kontron’s statement that “the software ecosystem for ARM also provides a simplified migration path for application developers, through fully deployed Windows CE6, CE7 and several different Linux distributions” shows the direction of the company’s ARM plans. No full-fledged industrial PCs but specialized solutions. This notion is supported by the following comment on RTOSes: “In addition, vertical markets will benefit through fully supported real-time operating systems such as QNX, VxWorks and Green Hills to name just a few well known options out of the x86 environment.”

So, Kontron’s opening-up towards non-x86 processor architectures is not very surprising. Actually, it’s good sense of business. However, we’re certainly looking forward to the evalution results and the first products. There’s a lot more to come. Read it here!

Embedded goes Mainstream: Apple, ARM and Atom

Until a few years ago, the processor business was – no, not simple, but clearly structured. Intel was for PC’s and Servers, ARM for Embedded stuff and AMD wasn’t really anywhere. Today, AMD earns its money with graphics processors (remember ATI?), still trying to re-gain a foothold in CPUs. The PC world and the embedded space are becoming more and more entangled – in the words of Léo Apotheker: “The tablet effect is real”, leaving Intel and ARM as actual competitors.

It really did sound slightly exaggerated when Steve Jobs talked about the “Post-PC world” when he announced the first iPad in January 2010. Arguably, one could already have seen that coming with the rise of the netbooks during the first decade of the new century, but hardly anyone did. They were just cheaper notebooks, weren’t they? Netbooks changed some other things: This device category’s success caused Microsoft to rethink its XP licensing strategy and schedule. On the hardware side, Intel pushed its x86-based embedded roadmap with the Atom. While Microsoft didn’t want Linux to gain foot in a mainstream PC-like architecture, Intel couldn’t let companies like Via enter the PC market. Both couldn’t see the real changes and the competition emerging. ARM has never been a competitor in the PC area, since there is no good compatibility path from x86 to ARM’s RISC architecture; Apple used to do their proprietary and expensive fun stuff for graphics designers and Google was that dubious web search thingy.

The year 2007 brought the first iPhone and shortly after legions of ‘similar’ (i.e. copied) smartphones from practically every consumer electronics company all over the world. Most of them, like the orignal, with some kind of ARM processor inside. In due course, this development became the beginning of the end of mobile phone giants like Nokia (for the moment kept alive by Microsoft) and Motorola (taken over by Google). The search engine guys at Google were able to establish their Linux distribution ‘Android‘  as the only existing alternative to Apple’s iOS (nobody knows yet what’s going to become of Windows Phone). While this was slighty annoying for Microsoft, Intel didn’t care much, since the mobile phone business has never shown up on any of Andy Grove’s roadmaps.

The iPad – closer to an embedded computer than every device before, but better-looking – changed everything again, Léo Apotheker has got that one right even though one could argue whether giving up the world’s largest PC business completely may be some kind of overreaction. Hardware compatibility is no longer an issue, so the ARM architecture is actually entering Intel’s area of interest (even Microsoft has plans for Windows 8 on ARM) . Linux in its Android cloak is on innumerable iPad-lookalikes, a special browser-based Google-OS (Chromium) is even driving a new kind of notebook while PC and netbook sales are dropping worldwide. The giants of the olden days (yesterday, actually) are struggling. Intel and Microsoft are certainly not going bust anytime soon, but the markets they used to dominate easily are changing rapidly.