MCCI Blog

Business

News and comment related to the business side of connectivity.

IPextreme's "Take 5" Interview

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Friday, 10 October 2014
in Business

Warren Savage of IPextreme recently interviewed me for an segment of his "Take 5 with Warren" series.

I think I made three important points.

First, once it's possible to buy an IP block, or system software, from an external vendor, that techology almost by definition is no longer core technology, but is mission critical. 

Aside: I hear someone asking, "Why?"  My answer: if a competitor can buy that technology, so possession of the technology (by itself) no longer differentiates one product from a competitor's. Anything that is not differentiating is therefore (almost certainly) not core technology. 

Another aside: I hear someone else asking "core? mission-critical?"  Gordon Moore popularized this terminology in Crossing the Chasm. My example comes from digital cameras:  lens and imager are core technology (and are a basis of competition between vendors); USB is mission critical (everyone has it, but nobody competes on that basis).

Yet another aside: "what's an IP block?" An IP block is essentially a hardware component that performs a specific function, and is packaged to be bought by a chip maker and incorporated into a chip. Warren's company, IPextreme, specializes in IP blocks, and also in design for reuse. MCCI does a very similar thing, except that our blocks are pure software, and instead of being incorporated into the chip, are incorporated into the software that accompanies the chip.

Second, modern IP blocks (USB hardware blocks, for example) must be accompanied by specialized system software in order to actually perform the desired function. In order to reduce hardware complexity and increase flexibility, higher-level functions are delegated to system software. Much of this system software (like a TCP stack or a USB stack) only performs housekeeping operations; but without the housekeeping, nothing works. T

Third, and a consequence of my previous point: IP providers, embedded system software makers, and SoC IP users, and the software team that's supporting the SoC deployments need to cooperate closely. Our timescales differ a lot -- IP providers typically develop their code a year before the chip design starts. Then there's a need for system software for verification of the SoC design. Later, when the chip is ready, there's a need for cooperation between the various groups for problem resolution. At that point, if the IP providers and the embedded software company (like MCCI) have already been working together, we can do a much better job of helping resolve problems.

Many thanks to Warren for the interview and the interest!

Tags: Untagged
Hits: 1955 0 Comments

WinHEC is back

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Friday, 26 September 2014
in Business

Just heard from my friends at Microsoft: WinHEC is back.  https://www.winhec.net/en-Us.

I think the new team at Microsoft is moving in the right direction: not only cloud first, etc., but engineering as a foundation.

Tags: Microsoft
Hits: 874 0 Comments

More on windup of Renesas Mobile (correction)

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Friday, 28 June 2013
in Business

Well, I should have looked more deeply into this before posting https://forums.mcci.com/making-connections/entry/renesas-to-wind-up-renesas-mobile.

As I was thinking this over, I was puzzled by the lack of links to primary sources for the Renesas Mobile wind-up in the WSJ article. So I did some searching.

Here's the link to the official announcement, not cited in any of the articles I found: http://www.renesas.com/press/news/2013/news20130627.jsp

And here are the relevant quotes:

[Renesas] decided today to stop developing activities and sales expansion of the LTE Modem, and to work towards the closure of the RME Group.

"RME Group" refers to Renesas Mobile Europe, which is the parent company for the Finnish, Chinese and Indian operations.

The Group [meaning Renesas Electronics and Renesas Mobile Corporation] will continue to supply modem products older than 4th generation, or already in mass production and supplied to its customers. In addition, the Group is committed to continue its operation related to car information systems and industrial equipment business, and further strengthen them as its core business.

Apparently the WSJ (and most of the other reports so far, including mine) are not quite right. Renesas Mobile will still provide modems, and will still serve existing customers. But the European, Indian and China groups will be shut down, and there will be no new development for LTE.

So while this is a major change of direction, it doesn't spell the end of RMC by any means.

[Update: 2013-06-28 00:00 EDT: fix formatting and add a link to the WSJ article.]

Tags: Untagged
Hits: 2469 0 Comments
0 votes

Renesas to wind up Renesas Mobile LTE efforts

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Thursday, 27 June 2013
in Business

[Corrected 2013-06-27 23:45 EDT -- see https://forums.mcci.com/making-connections/entry/more-on-windup-of-renesas-mobile-correction]

Renesas Mobile Corporation (RMC), like ST-Ericsson, will soon be no more. According to the Wall Street Journal and other online sources, Renesas Electronics will wind up operations of Renesas Mobile by the end of 2013.

Not true: they're just closing the overseas divisions and stopping LTE development.

I wrote last year about Renesas' attempts to divest RMC ("Renesas restructuring; to sell Renesas Mobile?" and "More sale rumors on Renesas Mobile"). Apparently those efforts were unsuccessful, and Renesas has decided to cut its losses.

The situation for RMC customers and employees is actually worse than that for ST-Ericsson. ST-E is being split up and absorbed by the parent companies (ST and Ericsson); most of the products will continue to be manufactured, and many employees have been offered positions at the parent companies. RMC is just being wound up. Because of life-time employment, people at RMC Japan may be offered positions at Renesas Electronics, but people in the other locations are likely to be laid off.

Not quite correct: the overseas employees are losing their positions, but RMC Japan will continue to exist.

I know a number of people at RMC Japan, India, and Finland; they're a very talented and professional group. As in any large business shutdown, a lot of hard work will simply be discarded. I'm very sad to hear this news. Technical texts on economics talk about financial investments being "validated" by profit. But they don't talk about the time investments made by employees -- for engineers, validation comes from the use of ones work in a successful product. When a company is wound up, all possibility of validation is cut off. Most of us who have been in the industry for a while know that kind of pain. While it's no fun being laid off, it's much better if the company goes on to use your contributions. 

Most of the IP and know-how owned by RMC (radio designs, software, etc.) only has value in the context of an operating business that is producing chips. However, RMC has a big (and ostensibly valuable) patent portfolio. Look for Renesas will try to monetize that portfolio by selling or auctioning the patents.

Given an on-going business, any discussion about a patent auction is premature.

More at https://forums.mcci.com/making-connections/entry/more-on-windup-of-renesas-mobile-correction

Tags: Untagged
Hits: 2532 0 Comments
0 votes

USB Development Survey Results Available From Embedded Market Forecasters

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Friday, 19 April 2013
in Business

MCCI and Embedded Market Forecasters (EMF) are happy to announce that the results of EMF’s recent survey of USB developers are now available. With input from roughly 500 developers, this survey has the most accurate snapshot of USB development practices. This 19-page report can be downloaded gratis by accessing the following link: http://www.embeddedforecast.com/EMF_freewhitepapers38.php. Other whitepapers and industry data are available from the EMF website, http://www.embeddedforecast.com/.

Thanks to the members of the MCCI community who contributed to the results.

In case you’re wondering, the vast majority of respondents were not MCCI users; the respondents cover the entire embedded community, not just the high-volume telecoms and high-volume consumer electronics customers that have been MCCI’s traditional base. The biggest surprise of the survey was that 40% of respondents are doing safety-related designs (as indicated by the standards with which they comply); and yet only 10% of respondents are doing USB compliance testing, which I’d consider the minimum level for testing any USB design.

Tags: Untagged
Hits: 3054 0 Comments
0 votes

TI, OMAP, and the "long tail" of embedded systems

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Thursday, 27 September 2012
in Business

Interesting article today in EE Times:  TI Steering OMAP to Embedded.

Related to my previous posts:

In the embedded processing space, TI's OMAP and connectivity solutions generate about $400 million in revenue annually from a base of about 4,000 customers, [Greg Delagi, TI's senior vice president for embedded processing] said. In smartphones and tablets, OMAP and connectivity chips generate about $900 million from roughly 10 customers, he said.

Elsewhere, TI claims 12% market share for embedded systems, probably by revenue; they effectively claim a market size of $3.3 billion.

Microchip, as noted in Microchip and SMSC: funding technology development, has a lot more customers: 70,000.  Of course, OMAP chips are a lot more expensive than Microchips' products.  Furthermore, TI has historically been a lot more like SMSC in its approach to OMAP customers than Microchip, focusing on a smaller number of customers with higher volume.

Anyway, it's pretty clear that Microchip is not serving the "embedded system" market alone; their sales are roughly a billion dollars a year; so there must be a lot of "microcontrollers" in that mix.

As 32-bit processors replace 8- and 16-bit processors (see Freescale to Finish Development of 8 and 16bit Controllers), the embedded system market will be harder and harder to distinguish from the microcontroller market.

Tags: Untagged
Hits: 1809 0 Comments
0 votes

More consolidation: MediaTek tenders for Mstar

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Sunday, 24 June 2012
in Business

While I was distracted with other things on Friday, I missed the big news: MediaTek is tendering for Mstar.

This Reuters article is the orginal source of the news:

http://www.reuters.com/article/2012/06/22/us-mediatek-mstar-idUSBRE85L0G320120622

EEtimes had two articles, one covering the acquisition, and an analysis by Junko Yoshida, "Dial M for MediaTek". (As usual, her analysis is worth reading.)

Although both articles point out the strength of MediaTek, I think they underestimate the strength of MStar. In the volume-sensitive, low-margin TV business, there really is no profit in being number 2 (which is MediaTek's position). MStar has the lead OEMs -- the OEMs who make the profit. They also have a strong lead in the monitor business, by virtue of their relationships.

There haven't been a lot of mergers in the Taiwan silicon business in the past. MediaTek, however, has proven itself able to handle M&A of foreign companies. They bought AnalogDevices' mobile business a few years ago, and successfully integrated it -- or at least, managed not to be hurt overall by the integration. (Compare this to BenQ's fatal misstep with the acquisition of Siemens' mobile phone business.)

Still, given the relationship-driven nature of sales in the TV and monitor business, one wonders whether there really will be synergies to exploit. After a merger, it's common to simplify the sales channel, because there's usually some redundancy. But in relationship driven sales, there's less redundancy. I'd expect therefore that MediaTek's TV and LCD monitor chip business would be merged into MStar's, and (if there is redundancy) that it would fall more on the MediaTek side. It's more difficult to predict what will happen with the handset chip business. It's harder to consolidate because of ongoing business relationships. Look at how ST-Ericsson has struggled with rationalizing their multitude of platforms.

As an aside: I hadn't considered MediaTek as a possible acquirer of Rensas Mobile (see earlier posts Renesas Restructuring: to sell Renesas Mobile? and More Sale Rumors on Renesas Mobile). I've not heard that MediaTek has a really competitive global LTE platform; Renesas Mobile, despite the uncertainly on the corporate side, does.

Tags: Untagged
Hits: 2310 0 Comments
0 votes

More sale rumors on Renesas Mobile

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Tuesday, 19 June 2012
in Business

On May 28, in "Renesas restructuring; to sell Renesas Mobile?", I wrote

However, nobody in the English press has picked up on another rumor, which seems to be mentioned in this article from Jiji in Japan. (I don't speak Japanese, and so I'm relying on friends here in Tokyo who do.) The article seems to say that, in addition to a factory being for sale, Renesas Mobile is for sale.

It seems as if I actually interpreted the article correctly (and my Japanese friends guided me aright: they provided the feel, and I parsed the Japanese using my knowledge of Chinese characters and careful use of machine translation programs). EETimes today quoted the Asahi Shinbun as announcing that a sale is being considered.

Renesas Mobile Corp., the mobile chip company subsidiary of struggling Renesas Electronics Corp., could be put up for sale as part of a previously reported re-organization. Other details that emerged in an Asahi Shimbun report that referenced unnamed sources included that Renesas is considering closing or selling off at least 10 of its 19 domestic manufacturing plants over the next three years.

It's sort of cool to have been first to report this in English, by a couple of weeks. (Not that cool -- this and a couple of bucks will get you a cup of coffee in NYC. But it's amusing, anyway. Probably a fluke. Keep reading and find out!)

The question still is: who would buy? My list from late May still seems reasonable: Apple, Samsung, Amazon, TI. Perhaps I'd add Microsoft, given their move into producing tablets.

Any other ideas?

Tags: Untagged
Hits: 3155 0 Comments
0 votes

Renesas restructuring; to sell Renesas Mobile?

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Monday, 28 May 2012
in Business

Today's announcements about Renesas restructuring got a lot of press (for example, see articles at Reuters and FT.com).  

However, nobody in the English press has picked up on another rumor, which seems to be mentioned in this article from Jiji in Japan. (I don't speak Japanese, and so I'm relying on friends here in Tokyo who do.) The article seems to say that, in addition to a factory being for sale, Renesas Mobile is for sale. (A later article from Jiji in English mentioned job cuts at Renesas Mobile, but not sale; and the later articles focused on the announcment that Renesas will use TSMC.)

It's interesting to speculate about who might buy Renesas Mobile, who have both LTE products and a decent patent pool (inherited from Nokia). Qualcomm has no need. Google has Motorola. NVidia bought Insera to get similar technology. Intel bought Infineon's mobile group. Apple? Samsung? Amazon? TI?

Any other suggestions?

Tags: Untagged
Hits: 5049 0 Comments
0 votes

Microchip and SMSC: funding technology development

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Thursday, 10 May 2012
in Business

A reader contacted me to ask me a question about a recent post, "Consolidation continues: Microchip to buy SMSC".

The following was not clear to me:

"Funding that kind of technology development effort really requires a target customer base that is more similar to SMSC's than to Microchip's"

Could you please elaborate?

I'll try. The answer is fairly wonkish, as economists might say.

Here's the background.  Microchip is (as the original EETimes article so aptly put it) "a “long-tail” company with 70,000 customers". SMSC is a traditional high-volume semiconductor company, focusing on a smaller number of higher-volume customers.

From this, I immediately surmise a couple of things:

  1. Microchip has higher costs for servicing these customers. There are many more design wins, and therefore many more points where training, or support, will be needed. Almost certainly, Microchip has to work through distributors; this implies giving up some of the margin to distributors.
  2. SMSC has lower costs for servicing the customers. There are fewer design wins, and therefore not as much support is needed.  Further, most SMSC sales will be directly to their customers, without involving distributors

As a result, the operating expenses are likely to be divided up differently: Microchip will allocate more to sales and support, and less to R&D.  SMSC will allocate more to R&D, and less to sales and support.

Looking at their most recent financial statements, I seem to have surmised correctly. 

  • Microchip allocates only 43% of their operating expense to R&D; they allocate 57% to G&A and sales expenses.
  • SMSC allocates 49% of their operating expense to R&D; they allocate 57% to G&A and sales.

This difference persists if you look over the last 5 years:  Microchip has allocated only 42% of op-ex to R&D, whereas SMSC has allocated 48%.

This implies something to me about the management emphasis of the two companies. SMSC is more focused on R&D; Microchip is more focused on sales. 

Furthermore, SMSC, because of their customer base (the big customers) is more likley to develop and adopt new technologies faster than Microchip. There's backup in the EEtimes article and commentary; there's also empirical evidence. SMSC was active in the development of USB 3.0, active in the development of standards such as HSIC USB, and has already shipped USB 3.0 products. Microchip has not participated actively in USB standard development.

I believe that the relative allocation is due to the targeted customer base. Serving the Microchip customers requires a company that is biased more towards sales. Serving the SMSC customers requires a company that is biased more towards R&D.

Coming back to my original statement, I see that it was very telescopic (as in what happens to a railroad car in a collision: it folds up like a telescope). By "funding that kind of technology development effort really requires a target customer base that is more similar to SMSC's than to Microchip's", I really meant the (much longer) step-by-step analysis that I've presented here.  Customer base implies company organization and demand for new technologies; company organization implies relative funding for R&D, and management interest in advanced R&D; therefore if you want to fund R&D, you probably want a customer base that is more like SMSC's than Microchip's.

A long explaination for a very short sentence!  Read on if you want to see the data I used.

Tags: Untagged
Hits: 2935 0 Comments
0 votes

Consolidation continues: Microchip to buy SMSC

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Thursday, 03 May 2012
in Business

Yesterday it was IDT buying PLX.  Today it's Microchip and SMSC. 

According to EEtimes, Microchip is buying SMSC for $939 million, roughly 2x revenue.  As the article points out, Microchip mostly sells into low-volume applications, whereas SMSC has traditionally been higher volume.

I'm interested because SMSC has been an important developer and vendor of USB technology. For example, HSIC USB (used in MCCI's Catena 1910) was developed by SMSC, and has been licensed by many companies producing mobile products. SMSC is one of the important vendors of discrete USB 2.0 PHY chips (the specialized electronics that convert from normal digital logic to the specific analog levels needed for USB operation). They also were important vendors of USB 2.0 and USB 3.0 silicon, including the first USB 3.0 to VGA/HDMI adapter chip. (DisplayLink led in USB 2.0, but they were a little behind SMSC getting their USB 3.0 product to market.) SMSC has been very active in USB standards development.

Microchip is strong in USB also. MCCI uses one of the Microchip controllers in our USB 3.0 Connection Exerciser as well as our Catena 1640 fully programmable low-speed device.  However, they've not participated in standards and technology development nearly as actively as SMSC has. Funding that kind of technology development effort really requires a target customer base that is more similar to SMSC's than to Microchip's. It will be interesting to see how the merged company approaches the problem.

Updated 2012-05-10: fixed typo. I also posted more on this topic: Microchip and SMSC: funding technology development

Tags: Untagged
Hits: 2991 0 Comments
0 votes

IDT to buy PLX

Posted by Terry Moore
Terry Moore
 
User is currently offline
on Tuesday, 01 May 2012
in Business

According to EEtimes, IDT is buying PLX for $330 million.

PLX makes a family of USB 3.0 device chips (the USB 3380 and USB 3382) that run over PCIe. MCCI's USB DataPump supports this family.  It's a nice kit part for adding USB 3.0 support to PCIe-capable systems.

MCCI also has used the PLX bridge chips in a variety of ExpressCard products, including the Catena 1820 (bridging to the Renesas R8A66597) and the Catena 1910 (bridging to the Synopsys DesignWare High-Speed USB 2.0 Core).

Tags: USB3
Hits: 2962 0 Comments
0 votes
Legal and Copyright Information