z*******n
发帖数: 1034 | 1 By Simon Hill July 24, 2014
You could be forgiven for not knowing the name ARM, but you probably should.
It was founded in 1990, the same year that the World Wide Web was born, and
it’s the world’s leading supplier of semiconductor IP, licensing
processor technology and software tools to companies big and small.
You can find ARM technology in everything from digital cameras to car
braking systems to digital set top boxes, not to mention 95% of the
smartphones in the world today. This British multinational has a seemingly
never-ending list of partners and they’ve shipped more than 50 billion
chips between them since 1991.
The growth of ARM in the last few years has been pretty spectacular and it’
s all based on a clever business model that informs an open and progressive
attitude at the heart of the company.
ARM does things differently
“A business model of sharing the risk, always with a can-do attitude,”
explains Rod Crawford, ARM’s Director of Software Technologies, means that
“your success is our success.”
This approach has led ARM to fruitful partnerships with some of the biggest
names in tech. ARM has worked with an impressive list of companies over the
years including Samsung, Apple, and Microsoft.
It has also contributed significantly to a number of open source projects
and it continues to try and engage and collaborate, rather than always
focusing on the bottom line. There’s a sense that progress will benefit us
all. By solving problems and taking on the R&D burden for partners, ARM has
built some strong relationships and pulled off the impressive feat of
working with many major players that are in direct competition.
apple-newton br1dotcom (via Flickr)
The ARM approach
We’ve discussed ARM’s rise before, but not the ethos that enabled it. Rod
Crawford attributes the business model to CEO Robin Saxby, who was appointed
in 1991. The company was freshly spun out of Acorn and was working with
seed capital from Apple on a processor which would be used in the Apple
Newton. There was plenty of ambition, but Saxby provided the plan.
ARM's business model informs an open and progressive attitude at the
heart of the company
The company would run lean and mean for the next five years and instead of
doing big upfront license deals and selling its wares, Saxby envisioned a
partnership model. ARM would assist with the design of new technology and
help with the transition and the partner would pay a license fee, but it
wouldn’t be a straight business transaction. If ARM enabled the partner’s
success then they would pay a royalty fee on every chip shipped.
It was a clever way to build mutual success and trust, and it would ensure
that partners pulled in the same direction. As new partnerships were
established with the likes of Sharp and Texas Instruments, ARM extended the
idea to include companies further down the chain, companies providing tools
and operating systems. Rod created a Software Development Tookit (SDT) to
enable developers to design software for the ARM processors.
Left: Former CEO and former Chairman Robin Saxby
Investors.com Left: Former CEO and former Chairman Robin Saxby
The SDT could be split into different components with free open interfaces,
the compiler, the debugger, JTAG runcontrol unit, and ARM would model and
then license them out to allow tool vendors to combine components with their
technology and rapidly get onto the platform and find a route to success on
the ARM architecture.
“It was obvious to us that if we tried to create a war and tried to say you
can only use tools by ARM for ARM then the industry would have shunned us,
” says Rod, going on to explain, “it was much more important to understand
that you’re going to actually be more successful if you enable your
competition and compete.”
By licensing out components of the toolkit, ARM was building a kind of
ecosystem. It was acting as a bridge between companies that enabled them to
play nicely together. Instead of pointlessly competing and innovating in all
the wrong places they could build from a solid, standardized platform that
was constantly improving and focus their own efforts on other aspects of
their end products.
nokiaDreamsrain.com
Moving into mobile
The year before Rod joined the company in 1994, ARM was already in talks
with Nokia about bringing a 32-bit processor to their mobile phones. In the
end they re-encoded a subset of the 32-bit ARM architecture into 16 bits
instead (the THUMB instruction set) and Nokia shipped the first phone with
ARM technology in it in 1996.
Nokia shipped the first phone with ARM technology in it in 1996
Rod had already moved to California by 1995 to help set up a team to port
real-time operating systems to ARM. They would then hand the ports back to
the OS companies so that they could start delivering on the ARM platform. It
was a proactive approach and it worked.
In the late 90’s ARM architecture evolved. It got an MMU, a separate
instruction and data cache, and a partnership with Microsoft to run Windows
CE soon followed. ARM was the first company to port Sun Microsystems’ Java
OS onto the ARM architecture and they optimized it, creating a very fast
Java interpreter. They were also engaged by Metroworks (later acquired by
Motorola) to do a just-in-time compiler for Java. As the mobile games
industry began to take off, ARM pushed its limitation further by re-encoding
152 of the Java byte codes in hardware to create Jazelle technology,
delivering higher performance without high power demands or high costs.
By the year 2000 it was time for some changes. ARM had been looking at disk
drives and anti-lock brakes and figured there was a need for better real-
time responses.
The new millennium
“We split the ARM architecture into three categories, A for applications, R
for real-time, and M for microcontroller,” explains Rod.
The Cortex-A chips would power our smartphones, not to mention set top boxes
, enterprise networking, and a lot more. The Cortex-R chips were for areas
like automotive and disk drives. The Cortex-M series was for little things,
the sorts of sensors you might find in white goods and fitness trackers, a
low energy line that could enable ideas like the Internet of Things to
become a reality.
By the year 2000 it was time for ARM to make some changes
It’s a strategy that seems to have ARM in pole position to power the next
wave of devices, and they do so by empowering their partners, not dictating
to them.
ubuntu
Working on Linux
You can see an echo of the same attitude in ARM’s other pursuits. The
company established good relations with early embedded Linux companies. By
2008 the processors were reaching a level of maturity and performance that
there was an opportunity for full scale Linux availability on the ARM
architecture. A partnership with Canonical led to Ubuntu being ported and
ARM engaged with the Linux community.
“In 2010 all the companies with processors purporting to be capable of
running Linux had developed their own Linux kernel,” Rod explains, “that
led to many, many different kernels and it was difficult to maintain them
all in the open source upstream Linux kernel project.”
The Linaro open source company was born in 2011. Founded by ARM, IBM,
Freescale, Samsung, ST-Ericsson, and Texas Instruments, it has grown into a
broad coalition. Broadcom, LG, Qualcomm, MediaTek, HiSilicon and others have
joined in the last couple of years.
A partnership with Canonical led to Ubuntu being ported and ARM engaging
with the Linux community
This not-for-profit organization operates on a partnership model where they
all put in money and engineering expertise. A decentralized team works on a
common goal to share the load and base it all on one kernel. This enables
standardization of the platform, it reduces fragmentation, and it eliminates
redundant effort.
ARM’s drive to improve the hardware and software development landscape
extends beyond big name OEM partners. It even goes beyond the small software
companies. There’s a refreshing sense that ARM believes improving the
entire environment and knocking down barriers to entry will ultimately be
good for everyone.
mbed
Don’t forget your roots
When we ask Rod what he’s most excited about next he tells us about mbed.
It’s a development platform for creating products for ARM microcontrollers
and it grew out of an internal R&D project to build an embedded platform
that people could really innovate on.
“We want to encourage grass roots hobbyists,” Rod enthuses, “they can
plug in an mbed board (for around $10) via USB, write code in the browser,
and it’s complied in the cloud.”
This is an open source project, all the software and hardware design is open
source, and only the cloud compiler remains proprietary to ARM. He also
mentions ARM’s recent work with Arduino as a point of pride.
Mbed is about bringing the barriers down, making it easier to contribute
to the collective
For Rod this is all part of “bringing the barriers down, making it easier
to contribute to the collective,” and he envisages the potential for a “
different kind of business model, with common cloud contributions”.
“It’s never been easier to build an embedded system for a broad developer
community,” says Rod, “the ability to share knowledge via the World Wide
Web has enabled the open source community to be so successful and now it can
take advantage of the cloud. It’s a very interesting time ahead in the
next few years.”
Watch and learn
In a tech world fraught with rivalry, destructive lawsuits, and dirty
marketing campaigns it’s refreshing to find ARM espousing a “let’s work
together” attitude and acting to foster innovation. The fact that it is a $
20 billion company should be proof enough that a collaborative approach and
some commitment to open source ideals is not mutually exclusive with big
profits. Others could definitely take note. |
|
