Embedded Intel® Solutions - Subscribe
Embedded Intel® Solutions - Subscribe
Embedded Intel® Solutions - Subscribe
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>Embedded</strong><br />
Intel ®<br />
<strong>Solutions</strong><br />
Winter 2011<br />
<strong>Embedded</strong> Virtualization—the Key to<br />
Real-time Determinism<br />
Combining x86 Architecture with FPGAs<br />
Optimizing the Virtual Environment<br />
Intel ® Atom Processor E6xx Series<br />
Reaches the SUMIT<br />
VPX Apps for Wide-Area Airborne<br />
Gold Sponsors<br />
www.embeddedintel.com
Copyright © 2010 Kontron AG. All rights reserved. Kontron and the Kontron logo and all other trademarks or registered trademarks are the property of their respective owners and are recognized. Rev. # G109us02<br />
» Looking for the latest fl exible<br />
processor technology? «<br />
Kontron brings next generation <strong>Intel®</strong> processors, like the <strong>Intel®</strong> Atom processor E6x5C<br />
and E6xx series, to market on high performing, highly fl exible embedded platforms.<br />
» The next generation <strong>Intel®</strong> Atom processor is a cost effective building block for future designs<br />
» Open standard interfaces deliver unprecedented levels of I/O flexibility<br />
» For use in extended temperature ranges of -40°C to +85°C<br />
» Increased software services including consultation, design, porting and validation<br />
INTEL® ATOM PROCESSOR-BASED SCALABLE PLATFORMS<br />
nanoETXexpress-TT<br />
Computer-On-Module<br />
» <strong>Intel®</strong> Atom processor E6xx series<br />
» Ready for use in harsh environments<br />
» Optimized integrated resources<br />
» Custom I/O via carrier boards<br />
CRITICAL QUESTIONS ... ANSWERED<br />
Call, Email or Visit today.<br />
MICROSPACE® MSMST<br />
PCIe/104 SBC<br />
» <strong>Intel®</strong> Atom processor E6x5C series<br />
» Flexible I/O functionality with FPGA core<br />
» HSMC for custom interfaces<br />
» Ideal for small, low-power devices<br />
Call: 1-888-294-4558<br />
Email: info@us.kontron.com<br />
Visit: kontron.com/NextGen<br />
If it’s embedded, it’s Kontron.
IN THIS ISSUE<br />
WINTER 2011<br />
FROM THE EDITOR<br />
4 The Easy Stuff is Done<br />
By Ed Sperling, Guest Editor<br />
NEWS<br />
6 Product News<br />
By Jim Kobylecky<br />
DEPARTMENTS<br />
FOCUS ON INTEL<br />
16 Intel Targets Digital Signage<br />
By Cheryl Coupé, Contributing Editor<br />
FOCUS ON INTEL<br />
17 Wide-Area Airborne Surveillance (WAAS)<br />
application Ideal for VPX<br />
By Cheryl Coupé, Contributing Editor<br />
FOCUS ON INTEL<br />
20 Break Away with Intel Atom Processors<br />
A Guide to Architecture Migration<br />
By Cheryl Coupé, Contributing Editor<br />
STANDARDS WATCH<br />
22 Digital Audio, Standards, and Chips<br />
By Henk Muller, XMOS<br />
SPECIAL FEATURES<br />
26 Trouble In The Patent Of�ce<br />
By Ed Sperling, Contributing Editor<br />
Why Current Practices Have Green IP Proponents Seeing Red;<br />
How to Limit the Impact of Trolls<br />
28 <strong>Intel®</strong> Atom Processor E6xx Series Reaches<br />
the SUMIT<br />
By Robert A. Burckle, WinSystems, Inc.<br />
Advanced Connector Unlocks Small-Form Factors<br />
30 Optimizing the Virtual Environment<br />
By Bob Carlson, Criterion HPS<br />
2 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
TECHNOLOGY APPLICATIONS<br />
32 <strong>Embedded</strong> Virtualization—the Key to Realtime<br />
Determinism in Multi-OS Systems<br />
By Kim Hartman, TenAsys Corporation<br />
33 What Do You Get When You Combine x86<br />
Architecture with FPGAs?<br />
By Christine Van De Graaf, Kontron<br />
35 Con�guration Is Key to Success with <strong>Embedded</strong><br />
Virtualization<br />
By Chris Ault, Wind River<br />
LAST WORD<br />
48 PCIe vs. RapidIO for <strong>Embedded</strong> Applications<br />
By Dave Barker, Extreme Engineering <strong>Solutions</strong> (X-ES)<br />
PRODUCT SHOWCASE<br />
39 6WIND<br />
40 AXIOMTEK<br />
40 AXIOMTEK<br />
41 Commell<br />
41 Emerson Network Power<br />
42 Emerson Network Power<br />
42 Emerson Network Power<br />
43 Emerson Network Power<br />
43 ITOX<br />
44 MSI Computer<br />
44 Pinnacle Data Systems, Inc.<br />
45 Pinnacle Data Systems, Inc.<br />
45 Pinnacle Data Systems, Inc.<br />
46 Pinnacle Data Systems, Inc.<br />
46 Pinnacle Data Systems, Inc.<br />
47 Pinnacle Data Systems, Inc.<br />
47 Extreme Engineering <strong>Solutions</strong>,<br />
Inc.
X-ES 2 nd Generation Intel ® Core i7 Processor <strong>Solutions</strong>: Delivering Innovation<br />
In 2010, Extreme Engineering <strong>Solutions</strong>, Inc. (X-ES) developed more <strong>Intel®</strong> Core i7<br />
processor products based on VPX, CompactPCI, VME, CompactPCI Express, and XMC form<br />
factors than anyone in the industry. This year, X-ES has added solutions based on the 2nd<br />
generation Intel Core i7 processor. Providing products customers want, when they want<br />
them – that truly is innovation that performs.<br />
X-ES offers an extensive product portfolio that includes commercial and ruggedized single<br />
board computers, high-performance processor modules, multipurpose I/O modules,<br />
storage, backplanes, enclosures, and fully integrated systems.<br />
2nd generation Intel Core i7 processor solutions available in a variety of form factors.<br />
Call or visit our website today.
FROM THE EDITOR<br />
The Easy Stuff Is Done<br />
By Ed Sperling - Guest Editor<br />
�e oil industry makes a good comparison to what’s going on in the semiconductor<br />
world. While everyone talks about the end of the oil era, the fact is<br />
that there is still plenty of oil left in the ground. It just gets harder to extract.<br />
�e same is true of power savings on a piece of silicon. It’s still possible to<br />
save lots of power in devices. It’s just harder to do—and more expensive. It<br />
will require new materials, new techniques, and in some cases a new way of<br />
thinking.<br />
�e materials that are likely candidates for lowering power and minimizing the effects<br />
of shrinkage at each new node are well known and have been tested for years. SOI leads the<br />
pack—particularly fully depleted SOI—but there also are some more exotic combinations that<br />
could play an important role for certain industries.<br />
�e techniques for lowering power are also well known. At 45nm, almost everyone is working<br />
with power islands, power modeling and multiple voltages, and at 28nm and beyond there<br />
will need to be standards for how to bring all of this stuff together in a more consistent manner.<br />
�e problem gets even worse in 3D, because power equates to heat, and heat is much tougher<br />
to get rid of in a multi-die stack.<br />
“Power is global in a device …<br />
reaching far beyond an individual<br />
chip and across the PCB …”<br />
�e new approaches are something of a shift in approach. While the semiconductor industry<br />
has been marching to the beat of smaller, faster and cheaper (or at least less expensive) for<br />
nearly half a century, the answer may no longer be from the standpoint of a single chip. It may<br />
be multiple chips, in addition to multiple chips in a stack. If one chip can do a specific function<br />
with less power, then that can be a significant gain for conserving overall system power. �e<br />
problem is that most engineering teams aren’t organized to make this type of approach work.<br />
Power is global in a device, though, which means it will require a global approach among<br />
systems companies—reaching far beyond an individual chip and across the PCB, which in<br />
some cases may mean across multiple geographies and time zones. �is isn’t so simple, and it<br />
may be the hardest piece for chip developers to come to grips with. Internal organizations and<br />
politics are common tripwires for companies, and that becomes even harder when it’s built<br />
around a somewhat vague concept such as power.<br />
Nevertheless, power will define which companies are competitive over the next decade and<br />
which ones are not, and the companies that embrace it first will be the ones best positioned to<br />
deal with it when their customers begin asking for bids on a more effective approach to extending<br />
battery life or cutting the power in plug-in devices, which soon will be measured directly<br />
by consumers.<br />
Ed Sperling is Contributing Editor for <strong>Embedded</strong> <strong>Intel®</strong> <strong>Solutions</strong> and the Editor-in-Chief<br />
of the “System Level Design” portal. Ed has received numerous awards for technical journalism.<br />
4 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
www.embeddedintel.com<br />
V.P. and Associate Publisher<br />
Clair Bright<br />
Editorial<br />
Editorial Director<br />
John Blyler<br />
jblyler@extensionmedia.com<br />
(503) 614-1082<br />
Managing Editor - NA<br />
Jim Kobylecky<br />
Contributing Editors<br />
Cheryl Ajluni, Geoffrey James,<br />
Ed Sperling, Craig Szydlowski, Nicole Freeman,<br />
Ann Steffora Mutschler, Cheryl Coupé<br />
Creative/Production<br />
Graphic Designers<br />
Brandon Solem<br />
Keith Kelly<br />
Production Coordinator<br />
Spryte Heithecker<br />
Online Production Director<br />
Jeff Cheney<br />
Advertising / Reprint Sales<br />
Vice President, Sales<br />
<strong>Embedded</strong> Electronics Media Group<br />
Clair Bright<br />
cbright@extensionmedia.com<br />
(415) 255-0390 ext. 15<br />
Marketing/Circulation<br />
Jenna Johnson<br />
To <strong>Subscribe</strong><br />
www.embeddedintel.com<br />
Extension Media, LLC<br />
Corporate Office<br />
President and Publisher<br />
Vince Ridley<br />
vridley@extensionmedia.com<br />
(415) 255-0390 ext. 18<br />
V.P. and Associate Publisher<br />
<strong>Embedded</strong> Electronics Media Group<br />
Clair Bright<br />
cbright@extensionmedia.com<br />
(415) 255-0390 ext. 15<br />
Vice President, Marketing and Product Development<br />
Karen Murray<br />
kmurray@extensionmedia.com<br />
Vice President, Business Development<br />
Melissa Sterling<br />
msterling@extensionmedia.com<br />
Human Resources / Administration<br />
Rachael Evans<br />
Special Thanks to Our Sponsors<br />
<strong>Embedded</strong> Intel ® <strong>Solutions</strong> is sent free to engineers and embedded<br />
developers in the U.S. and Canada who design with <strong>Embedded</strong> Intel ®<br />
Processors.<br />
<strong>Embedded</strong> Intel ® <strong>Solutions</strong> is published by Extension Media LLC, 1786<br />
18th Street, San Francisco, CA 94107. Copyright © 2010 by Extension<br />
Media LLC. All rights reserved. Printed in the U.S.
Fanless 1.1GHz SBC<br />
with SUMIT Expansion<br />
Runs Fast, Stays Cool<br />
Our fanless EBC-Z510-G is a powerful, nextgeneration<br />
EBX-size computing platform.<br />
Its outstanding complement of onboard I/O<br />
is augmented with MiniPCIe, SUMIT,<br />
and PC/104 connectors for highbandwidth<br />
PCIe and USB 2.0<br />
interfaces plus standard legacy<br />
bus expansion. This extended<br />
temperature SBC is compact and<br />
flexible enough to meet a broad<br />
range of application requirements.<br />
� ��������������������������������<br />
1.1GHz or 1.6GHz onboard<br />
�����������������������������������<br />
displays resolutions up to 1920 x 1080<br />
� ����������������������������������<br />
simultaneously<br />
� �������������������������������<br />
� ���������������������������������������<br />
� ��������������������������������������<br />
� ���������������������������<br />
� ���������������������������������������������<br />
��������������������������������������������<br />
���� ��������<br />
� �������������������������������������<br />
� ����������������� ������<br />
� ������������������������� �������� �������������<br />
����������������������������������������������<br />
I/O expansion capability<br />
� �������������������������������������������<br />
� ����������������������������������������������<br />
� ����������������������������������������<br />
Contact our factory application engineers for additional<br />
product information, custom configurations, and pricing.<br />
Ask about our 30-day product evaluation.<br />
TM<br />
WinSystems also offers...<br />
SUMIT-ISM CompactFlash<br />
High-Speed I/O<br />
PC/104 Modules<br />
���������������������������� ������������������<br />
Call 817-274-7553 or Visit<br />
www.winsystems.com /EBC-Z510<br />
��������������������������������������������<br />
��������������������������������������������<br />
���������������������������
NEWS<br />
NEWS<br />
Intel Unveils All New 2010 <strong>Intel®</strong> Core<br />
Processor Family<br />
Intel Corporation introduced its all new 2010 <strong>Intel®</strong> Core<br />
family of processors featuring unprecedented integration,<br />
smart performance, and including <strong>Intel®</strong> Turbo Boost<br />
Technology1 for laptops, desktops and embedded devices.<br />
The introduction of new <strong>Intel®</strong> Core i7, i5 and i3 chips<br />
coincides with Intel’s new 32 nanometer (nm) manufacturing<br />
process which is key to delivering processors and features<br />
at a variety of price points, and integrating high-definition<br />
graphics inside the processor. More than 400 laptop and<br />
desktop PC platform designs are expected from computer<br />
makers based on these products, with another 200 expected<br />
for embedded devices.<br />
CriticalBlue Optimizes 2nd Generation Processor<br />
Support<br />
CriticalBlue has enhanced its Prism product for software<br />
developers who are migrating to the 2nd generation <strong>Intel®</strong><br />
Core processor family. Software developers can now<br />
analyze their existing software applications, evaluate<br />
benefits, and select the appropriate <strong>Intel®</strong> processor. The 2nd<br />
generation Intel Core processor family supports the <strong>Intel®</strong><br />
Advanced Vector Extensions (<strong>Intel®</strong> AVX) which provide<br />
a new 256-bit SIMD floating point vector extension to the<br />
<strong>Intel®</strong> architecture. Exploiting features such as Intel AVX and<br />
<strong>Intel®</strong> Hyper-Threading Technology increases performance<br />
and lowers application power consumption.<br />
GE 2nd Generation <strong>Embedded</strong> Designs Deliver<br />
Greater Performance Per Watt<br />
GE Intelligent Platforms is utilizing 2nd Generation <strong>Intel®</strong><br />
Core i7 processors – formerly codenamed ‘Sandy Bridge’ -<br />
in a series of new products to be announced in the coming<br />
months. The first product will be the SBC624, a 6U OpenVPX<br />
single board computer featuring a range of processor options<br />
including a quad core Intel Core i7 processor running at up<br />
to 2.1GHz, ideal for applications like signal processing in<br />
intelligence/surveillance/reconnaissance and radar/sonar.<br />
The SBC624 is available in five build versions from benign<br />
to fully rugged.<br />
6 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
Product News<br />
By Jim Kobylecky, Managing Editor<br />
Mercury Computer Launches 2nd Generation<br />
Products for Defense Applications<br />
Mercury Computer Systems, Inc. announced the<br />
Ensemble Series 6U OpenVPX LDS6521 and the 3U<br />
OpenVPX SBC3510 modules based on the 2nd generation<br />
<strong>Intel®</strong> Core processor family. Mercury’s new modules<br />
enable best-of-breed subsystem application performance<br />
for extremely demanding ISR, defense, and aerospace<br />
applications. The 6U OpenVPX LDS6521 module utilizes the<br />
<strong>Intel®</strong> Core i7-2715QE processor to handle overwhelming<br />
volumes of high-resolution digital data. The Intel Core i7-<br />
2715QE processor doubles the number of cores available on a<br />
single device over the previous generation and now supports<br />
four cores.<br />
RadiSys Quad-Core COM Express 2.0 Module<br />
Based on 2nd Generation <strong>Intel®</strong> Core i7<br />
Processor<br />
RadiSys’ Procelerant CEQM67 module delivers extreme<br />
processing power and graphics within a basic 95mm x<br />
125mm form factor for medical, enterprise telecom and mil/<br />
aero applications Based on the 2nd generation <strong>Intel®</strong> Core<br />
i7 processor at 2.1GHz and <strong>Intel®</strong> QM67 Express chipset<br />
the module is designed for industries including medical,<br />
enterprise telecom and mil/aero as they migrate to smaller,<br />
more powerful, scalable devices that demand graphics and<br />
processor performance for video-intensive applications.<br />
Procelerant CEQM67 is RadiSys’ first Type 6 COM Express<br />
2.0 compliant product that provides quad-core processing in<br />
a basic 95mm x 125mm form factor COM Express revision<br />
2.0 module.<br />
X-ES Introduces Product Line Featuring 2nd<br />
Generation <strong>Intel®</strong> Core i7 Processor<br />
Extreme Engineering <strong>Solutions</strong>, Inc. (X-ES) introduces<br />
the XPedite7470, a conduction- or air-cooled 3U VPX Single<br />
Board Computer (SBC). X-ES’s first of six standard form<br />
factor products based on the 2nd generation <strong>Intel®</strong> Core<br />
i7 processor, the XPedite7470 utilizes the processor’s quadcore<br />
technology operating at 2.1 GHz to deliver enhanced<br />
performance and efficiency, making it an excellent COTS<br />
product for deployed military applications. X-ES has teamed<br />
with RunTime Computing <strong>Solutions</strong>® to support applications<br />
that can take advantage of the SIMD architecture of Intel<br />
AVX. VSI/Pro®, the premier math and signal processing
library available from RunTime Computing, will be<br />
supported on the XPedite7470 and all X-ES products based<br />
on the 2nd generation Intel Core i7 processor.<br />
Kontron Announces Over 10 <strong>Embedded</strong> Platforms<br />
Based on 2nd Generation <strong>Intel®</strong> Processors<br />
Kontron announced that it will be supporting this<br />
advanced new microarchitecture with CPU, processor<br />
graphics, memory controller (some with ECC support) and<br />
PCIe controller on a single die on over 10 embedded platforms.<br />
The 2nd generation <strong>Intel®</strong> Core i3, Core i5 and Core i7<br />
processors have a completely new microarchitecture, enabling<br />
small-sized designs to be created with unmatched computing<br />
power, graphics performance and energy efficiency on a small<br />
footprint*. In addition,<br />
with Kontron’s valueadded<br />
middleware, and<br />
strength in offering<br />
market-specific I/Os<br />
via FPGA, even more<br />
specific applications<br />
can be customtailored.<br />
LynuxWorks<br />
Announces<br />
LynxSecure<br />
Support for 2nd<br />
Generation <strong>Intel®</strong><br />
Processor Family<br />
L y n u x W o r k s ,<br />
Inc., announced<br />
LynxSecure 4.0<br />
support of the new<br />
<strong>Intel®</strong> Core i7 and<br />
Core i5 processor<br />
families. LynxSecure is<br />
a separation kernel and<br />
embedded hypervisor<br />
that provides an<br />
environment in<br />
which multiple guest<br />
operating systems and<br />
their applications can<br />
execute at the same<br />
time, in their own<br />
virtual partitions,<br />
without compromising<br />
security, reliability<br />
or data integrity.<br />
LynxSecure offers<br />
two virtualization<br />
schemes, paravirtualized<br />
guest OSes<br />
such as Linux and<br />
LynxOS-SE offering<br />
NEWS<br />
maximum performance, and fully virtualized guest OSes<br />
such as Windows, Solaris and Chromium OS requiring no<br />
changes to the guest OS.<br />
IEI Launches Dual-core <strong>Intel®</strong> Atom Processor<br />
D525 Fanless System<br />
IEI Technology Corp.<br />
released advanced fanless<br />
embedded systems featuring<br />
the dual-core <strong>Intel®</strong> Atom<br />
processor D525. The dual-core<br />
<strong>Intel®</strong> Atom processor D525<br />
-based embedded systems provide an increase of processor<br />
speed to 1.8 GHz with no bump in power consumption<br />
Feel the <strong>Embedded</strong> Rhythm<br />
sales-us@congatec.com<br />
Tel. +1 858 - 457- 2600<br />
congatec sets the beat for innovative embedded<br />
form factors. The Qseven standard offers the<br />
perfect solution for ultra mobile applications that<br />
require extremely reduced power consumption.<br />
The new conga-QA6 embedded computer<br />
module features Intel ® Atom processor E6xx series<br />
Extended temperature -40 °C to +85 °C<br />
Dramatically improved 3D Graphics<br />
All of the essential interfaces (3x PCIe, CAN, SPI... )<br />
conga-QA6<br />
No more compromises.<br />
Experience the congatec rhythm at:<br />
www.congatec.us<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 7<br />
Intel and Intel Atom are registered trademarks of the Intel corp. in the U.S. and other countries.
NEWS<br />
and support 800 MHz DDR3 memory. The debut D525<br />
series embedded system is the ECW-281B/B2-D525. This<br />
advanced series includes wide range DC power input from 9<br />
V to 36 V and dual GbE LAN – ideal for high speed network<br />
applications.<br />
<strong>Intel®</strong> Atom Processor E600C Series Includes<br />
Altera FPGA in a Single Package<br />
Intel Corporation is making it easier for customers to<br />
go-to-market with differentiated, custom-made designs.<br />
The company announced the configurable <strong>Intel®</strong> Atom<br />
processor E600C series, which features an <strong>Intel®</strong> Atom<br />
processor E600 (formerly codenamed “Tunnel Creek”) paired<br />
with an Altera* Field Programmable Gate Array (FPGA) in a<br />
single package. The new processor offers board space savings<br />
and better inventory control due to the single package, as<br />
well as a simplified manufacturing flow and single vendor<br />
support through Intel.<br />
Kontron SBC with Configurable <strong>Intel®</strong> Atom<br />
Processor E600C Series<br />
Kontron introduced a PCIe/104<br />
embedded single board computer<br />
(SBC) which pairs an <strong>Intel®</strong> Atom<br />
processor E600 series with an Altera<br />
Field Programmable Gate Array<br />
(FPGA) in a single package based<br />
on the <strong>Intel®</strong> Atom processor E600C series with industrial<br />
temperature range. With flexible FPGA I/O options, the<br />
Kontron MICROSPACE® MSMST allows OEMs to efficiently<br />
develop designs with the exact I/O requirements needed to<br />
address applications in markets such as automation, medical,<br />
transportation, energy, military and communications.<br />
Validated IP cores are available for CAN-bus, serial interfaces<br />
(SPI Master / UART) and PCI-Express, I2C and GPIO.<br />
Collaboration to Drive OpenVPX Standard in<br />
Military and Aerospace Applications across<br />
Product Lines<br />
Emerson Network Power and Mercury Computer Systems,<br />
Inc. will collaborate to promote interoperability on open<br />
standards-based subsystems for military and aerospace<br />
applications. This alliance seeks to provide interoperability<br />
between the companies’ rich range of embedded computing<br />
solutions, in order to enable defense customers to<br />
migrate their performance away from proprietary closed<br />
architectures to flexible open solutions, reducing risk and<br />
lowering development and deployment costs as a result.<br />
Wind River Broadens <strong>Embedded</strong> Virtualization<br />
Support<br />
Wind River introduced a new version of Wind River<br />
Hypervisor, a real-time, embedded hypervisor for<br />
virtualization, which adds support for the <strong>Intel®</strong> Xeon®<br />
processor 5600 series. Wind River Hypervisor, together<br />
with the latest versions of Wind River’s operating systems<br />
8 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
and development tools, represents the breadth and depth of<br />
the company’s multi-core software portfolio that combines<br />
multi-core-aware operating systems, tools, and embedded<br />
virtualization that scales from unicore to high-core count<br />
multi-core processors. Wind River Hypervisor is highly<br />
optimized for and integrated with the latest versions of<br />
VxWorks, Wind River Linux, Wind River Workbench and<br />
Wind River Test Management.<br />
X-ES Announces 6U CompactPCI SBC with <strong>Intel®</strong><br />
Core i7 Processor<br />
Extreme Engineering <strong>Solutions</strong>, Inc. (X-ES) announces the<br />
immediate availability of XCalibur4301, a conduction- or aircooled<br />
6U CompactPCI Single Board Computer (SBC) based<br />
on the <strong>Intel®</strong> Core i7 processor. For upgrades to existing<br />
commercial, industrial and military 6U cPCI applications,<br />
the XCalibur4301 delivers significant performance and<br />
efficiency improvements over previous SBCs. To satisfy the<br />
widest range of applications, from telecommunications to<br />
military applications, the XCalibur4301 is engineered to<br />
scale from an air-cooled, commercial (0 to 55ºC) version<br />
to a rugged, conduction-cooled (-40 to +85ºC) version with<br />
appropriate environmental test methods.<br />
Kontron Brings <strong>Intel®</strong> Core i7 Processor<br />
Performance to Mission Critical 3U CompactPCI®<br />
Platforms<br />
Kontron introduced two<br />
conduction-cooled 3U CompactPCI®<br />
boards for mission critical, data<br />
intensive, and network-centric<br />
platforms. The conduction-cooled<br />
Kontron processor board CP3002-RC<br />
with integrated graphics based on the latest <strong>Intel®</strong> Core i7<br />
mobile processor technology and the conduction-cooled<br />
Kontron Gigabit Ethernet Switch CP3923-RC with Layer<br />
2/3 Gigabit Ethernet (GbE) switching and full IPv4/v6<br />
management capabilities. Both boards help OEMs and<br />
system integrators improve the Size, Weight and Power<br />
(SWaP) of their applications with easy to implement COTS<br />
components.<br />
Kontron Implements <strong>Intel®</strong> Atom processor E6xx<br />
series on Computer-on-Modules and Panel PCs<br />
Kontron announced that the newly introduced industrialgrade<br />
<strong>Intel®</strong> Atom processor E6xx series will be supported<br />
on Computer-on-Modules and Panel PCs. The first products<br />
to integrate the new <strong>Intel®</strong> Atom processor, specified for the<br />
industrial temperature range (E2) from -40 to +85 °C, include<br />
the Kontron Computer-on-Module nanoETXexpress-TT,<br />
an ultra sized COM compatible to the new release of the<br />
COM Express COM.0 Rev 2.0 pin-out Type 10, and the<br />
Kontron rugged display HMITR, a fanless and maintenancefree<br />
intelligent display. In addition to optimized power<br />
consumption and improved graphics performance, target<br />
applications benefit from the new processor’s high level of
NEWS<br />
integration including memory, PCI Express and video / audio<br />
accelerators.<br />
LynuxWorks Announces LynxSecure 4.0 support<br />
for New <strong>Embedded</strong> <strong>Intel®</strong> Platforms<br />
LynuxWorks, Inc., a leader in secure virtualization<br />
technology, announced availability of LynxSecure 4.0 on<br />
the latest hardware platforms from Intel. The <strong>Intel®</strong> Core<br />
i7 processor family has been widely used in the desktop<br />
and laptop market, and now the quad-core versions of the<br />
processor are available for embedded designs. LynxSecure<br />
4.0 provides the most flexible virtualization solution for use<br />
in embedded systems. Built from the ground up, and now<br />
in its 4th generation, LynxSecure offers the combination of<br />
security with functionality, allowing embedded designers to<br />
use the latest software and hardware technologies to build<br />
complex multi-operating systems (OSes).<br />
IEI Introduces Dual-core <strong>Intel®</strong> Atom Processor<br />
D525-based SBCs<br />
IEI Technology Corp. (IEI) released advanced Single<br />
Board Computers (SBC) featuring the dual-core <strong>Intel®</strong><br />
Atom processor D525. The most notable advantage over its<br />
predecessors is its support for 800 MHz DDR3 memory. The<br />
dual-core D525 also provides an increase of processor speed<br />
at 1.8 GHz with no bump in power consumption. The initial<br />
launch of <strong>Intel®</strong> Atom processor D525-based series SBCs<br />
includes the PCISA-PV-D5251, PICOe-PV-D5251, NANO-<br />
PV-D5251 and NANO-PV-D5252. The key features in<br />
common are gigabit LAN supporting ASF 2.0 and TPM V1.2<br />
hardware security functions and uEFI BIOS architecture to<br />
support over 2.2 TB of HDD storage with a 64-bit operating<br />
system.<br />
Green Hills Software Supports <strong>Intel®</strong> Atom<br />
Processor E6xx Series<br />
Green Hills Software, Inc. announced support for the<br />
embedded <strong>Intel®</strong> Atom processor E6xx series. The Green<br />
Hills Software offering includes the INTEGRITY® real-time<br />
operating system, INTEGRITY Secure Virtualization (ISV)<br />
technology, MULTI® integrated development environment<br />
(IDE), optimizing C/C++ compilers, and DoubleCheck<br />
static analyzer.<br />
10 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
ROHM / OKI Semiconductor Co-Develop a<br />
Dedicated Chipset for <strong>Intel®</strong> Atom Processor<br />
E6xx Series<br />
ROHM Semiconductor, in collaboration with its affiliated<br />
company, OKI Semiconductor Co., Ltd., has introducted a<br />
dedicated Large Scale Integrated (LSI) circuit family designed<br />
to support the <strong>Intel®</strong> Atom processor E6xx series (formerly<br />
code-named Tunnel Creek). This chipset consists of a power<br />
management IC (PMIC), a clock generator IC (CGIC) and an<br />
input/output hub (IOH) IC. A reference board that simplifies<br />
customer development is also available.<br />
Out-of-the-box Development Platform For<br />
Emerson Network Power MicroATX Motherboard<br />
Emerson Network Power launched a new <strong>Embedded</strong><br />
Development Kit to enable design engineers to begin<br />
developing their application in a fraction of the time<br />
traditionally required. Part of the Wind River On-Board<br />
Program, the <strong>Embedded</strong> Development Kit includes an<br />
Emerson Network Power MicroATX motherboard based<br />
on the <strong>Intel®</strong> Core i7 processor with optimized trial<br />
versions of Wind River’s operating systems, development<br />
tools, embedded hypervisor and graphics software to<br />
help equipment manufacturers save time and money on<br />
application integration for a faster, more efficient timeto-market.<br />
The LiveUSB format enables designers to boot<br />
directly from the included USB flash drive to evaluate a<br />
fully operational development environment, eliminating the<br />
installation process.<br />
Kontron COM Express Basic Form Factor<br />
Computer-on-Module ETXexpress®-AI<br />
Kontron presented an additional version of the Computeron-Module<br />
ETXexpress®-AI based on the new COM Express<br />
Type 6 pin-out definition. As compared to the Type 2 pinout<br />
definition for COM Express Basic form factor modules,<br />
the new Type 6 pin-out offers configurable Digital Display<br />
Interfaces (DDI) SDVO, DisplayPort and HDMI/DVI along<br />
with 23 PCI Express Gen 2 lanes. This provides more native<br />
display options and higher serial bandwidth than previously<br />
possible. Most noteworthy is that the native support for all<br />
the new display interfaces simplifies carrier board designs,<br />
reducing time-to-market and total cost of ownership for<br />
graphics-intensive applications. The extensive PCI Express<br />
support underscores the trend of moving away from legacy<br />
parallel interfaces towards pure serial embedded system<br />
designs for higher bandwidth and reduced latency. This<br />
represents a smooth transition path for application designers<br />
looking to enhance their designs with next generation<br />
technology such as faster drives and peripherals.
<strong>Embedded</strong> PCs COMplete!<br />
COM Module with <strong>Intel®</strong> Atom Processor E6xxT/EG20T<br />
CoreExpress-ECO2 Module<br />
* available Q1/2011<br />
� Smallest COM module standard, 65x58mm<br />
� <strong>Intel®</strong> Atom processors E620T, E640T, E660T<br />
and E680T<br />
� Up to 2 GB soldered DDR2 RAM<br />
� Processor independent standard (sff-sig.org)<br />
� Especially designed for battery-driven mobile systems<br />
� CPU + Chipset: max. 5 watts Thermal Design Power (TDP)<br />
� CAN controller<br />
� Shock and vibration resistant<br />
� Wide temperature range (-40°C ... +85°C)<br />
� Fail-safe BIOS for secure BIOS updates<br />
� Condition monitoring using LEMT<br />
(new: with power sense)<br />
� Module availability 10 years<br />
LiPPERT <strong>Embedded</strong> Computers Inc.<br />
5555 Glenridge Connector, Suite 200<br />
Atlanta, GA 30342<br />
Phone (404) 459 2870 · Fax (404) 459 2871<br />
ussales@lippertembedded.com · www.lippertembedded.com<br />
* available Q1/2011<br />
Toucan-TC<br />
COM Express-Compact Module<br />
� COM Express form factor, 95x95mm, Type 2 pinout<br />
� <strong>Intel®</strong> Atom processors E620T, E640T, E660T<br />
and E680T<br />
� Up to 2 GB soldered DDR2 RAM<br />
� SATA SSD, max. 64 GB, soldered<br />
� MicroSD card slot<br />
� CAN and 4 COM ports on option connector<br />
� Shock and vibration resistant<br />
� Wide temperature range (-40°C ... +85°C)<br />
� Fail-safe BIOS for secure BIOS updates<br />
� Condition monitoring using LEMT<br />
(new: with power sense)<br />
� Module availability 10 years
Flexible <strong>Embedded</strong> Systems Architectures for a Future of Change<br />
Two major trends will drive the software architecture of embedded systems in the future. One is the partitioning of<br />
applications to make the best use of multi-core processors, and the other is the growing use of networked computing<br />
to distribute intelligence among multiple computing platforms.<br />
In an interesting way, these trends almost<br />
oppose one another. With the advent<br />
of multi-core processors with 2, 4, 8,<br />
or more processing units comes the<br />
prospect of consolidating embedded<br />
systems that were once comprised of<br />
multiple discrete processor platforms<br />
onto a single platform with different tasks<br />
running on different cores.<br />
This consolidation has the promise of<br />
decreasing system costs by eliminating<br />
the need for redundant computing support<br />
hardware. Additionally, improvements<br />
in network architectures and speeds<br />
Advanced<br />
Human<br />
Interaction<br />
with Data<br />
Post-Processing<br />
HUMAN / DATA PROCESSING<br />
No/Little<br />
Human<br />
Interaction<br />
and Data<br />
Post-processing<br />
RTOS Legacy App<br />
eVM ® for Windows<br />
Virtualization Technologies<br />
+<br />
Interactive<br />
Machine<br />
Control<br />
RTOS Legacy<br />
Application<br />
System Control<br />
Deeply <strong>Embedded</strong> Control<br />
Multicore and <strong>Embedded</strong> Virtualization<br />
Multicore Processors<br />
allow for increased use of distributed<br />
computing in the future. For example,<br />
Real-time systems evolutionary matrix<br />
INtime ® -for -<br />
Windows 4.0<br />
INtime ®<br />
Distributed<br />
RTOS<br />
Single Function MACHINE PROCESSING COMPLEXITY<br />
putting real-time processing nodes close<br />
to where the system’s physical work is<br />
being done improves performance for<br />
specifi c tasks. Whether the integrated<br />
platform approach, or the distributed<br />
computing approach, or the combination<br />
12 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
of both approaches is best for a particular<br />
embedded system depends on the<br />
application. <strong>Embedded</strong> system software<br />
architectures need to be fl exible enough<br />
to accommodate all of these system<br />
implementation models.<br />
Multi-function<br />
By Chris Grujon, TenAsys Corp.<br />
One way of profi ling embedded system<br />
software requirements is to consider the<br />
evolution of embedded system applications:<br />
In the beginning, real-time operating<br />
systems (RTOS) were used for single-<br />
function/purpose applications and were<br />
run on dedicated processors. An example<br />
is iRMX, an RTOS with rudimentary<br />
interfaces like all computers in those days.<br />
As personal computers evolved, people<br />
saw the value of adopting its technology<br />
for advanced graphical user interfaces. By<br />
providing more meaningful information to<br />
the user about the devices and/or system<br />
that is being controlled, he/she may perform<br />
more sophisticated functions, thereby<br />
enhancing the value of the device/system<br />
that is being controlled. Two approaches<br />
were taken:<br />
Some RTOS vendors added functionality to<br />
enhance the interface. These often didn’t<br />
meet the expectation and users eventually<br />
evolved to a two-box solution, one running<br />
a Windows-based user interface and the<br />
other running the controlling software on<br />
the RTOS.<br />
Others, for example TenAsys with INtime for<br />
Windows, took the approach of coupling<br />
their RTOS to Windows, providing the<br />
advanced user interface and the controlling<br />
software onto one platform.
Multi-OS multi-core systems are enabled<br />
���� ��� ������ �� ���������� ���������� ���������� ���������<br />
���� ��������� ���� �� ������� ��� ������� ��� ��� ��<br />
��� ���� �������� ���� ���� ��������� ����� ��������� ��<br />
������� ��� ������ ��������� �� ��� ����� �����������<br />
���������� �������� �� ���� ������ ������� ���� ���� �������<br />
���� ������������� ������� ��������� ���� ������� ������������<br />
�������� �������� ������� ���� ������ ����������� �� ���<br />
������� ��� ��� ������ �� ������� �������� ��������� ������������<br />
�� ���� �������� �������� �� ��� ���� �������� �� ��������<br />
��������������� �� ������� ����� ��������� �������� �������<br />
���� �� ������ ���� ��������� ����������������� ��������������<br />
������� �� ����� ��������� ����������� ��� ������ �������<br />
�������� ���� ��������� ��� �� �������� ���� ��� �������� ��<br />
� ������������� ���� ���������� ��������� ������� ���� ����<br />
����� � ������� ��������� �������� ��� �� �������� ���� ���<br />
��������� ��� �������� �������� ���� ��� ������� ��������<br />
�������������� ������� ������ ����� ��������� ������� ���� ��<br />
������ ������ �� ��� ��������� ������� ������� ��������� ���<br />
������������ �� ��� ����� �� �� ��� ����������� ��������� �� ��<br />
������� ��� ��� ������������ ������� �� ���<br />
OPC UA<br />
INtime RTOS Kernel<br />
ACPI<br />
INtime Distributed RTOS<br />
System Management Environment Machine Control Environment<br />
INtime API<br />
Real I/Os<br />
Memory<br />
Memory<br />
Shared Memory<br />
ATOM ® D510<br />
Core<br />
0<br />
Core<br />
1<br />
Control<br />
Application<br />
INtime API<br />
Real I/Os<br />
EtherCAT<br />
INtime RTOS Kernel<br />
OPC UA Traffic EtherCAT Traffic<br />
Networked embedded systems growing in popularity<br />
��� �������� �� ���������� ����� �� ���� ��� ��������� ��� ���<br />
��������� �� �������� �������� ��� ����� ����� �� ��������� ��� ��<br />
���������� ��� ����������� ���������� �������� ������ ���������<br />
���� �� ����� �������� ������������ ���� ��� ���������� �� �������� ������<br />
���� ���������� �� ������������� ������������ ���� ��� ���������<br />
���� ��������� �� �� ������������ ������������ ������ � ������� ��<br />
����� �� �� ������ �� ��� ������� ����� ����� �� �����������������<br />
������� ���� ���������<br />
�������� ������ ����������� ���� �� ��� ���� ���� �������� ��<br />
�������� ���������������� ��������� ���������� ��� ��������� ������<br />
� ������� ���� � �������� ����������� ���������� ������ �������<br />
������� ������� ��������� ���� ������ �������� ��������� �� ���<br />
������ ���� ��� ������������ ������� �� ���� �� ����������������<br />
��� ���������� ��� ���� ����� �� � ������������� ���� ���������<br />
������������ �� ������������� ������ ������������� �������� ��� ����<br />
������������ �� �� ��� �� �������� ����� �� � ��������� ��������� ��<br />
� ������������ ���� �� ������� ��������� ���� �� ������� ��������<br />
������ ����������� ���� �� ����� �� � ����� ��� ������ ���� ������<br />
���� ��� ���� ���� ������ �������� �� ������� �������� �������������<br />
Features<br />
� Distributed AMP Architecture<br />
enables predictable,<br />
deterministic inter-core<br />
communications across<br />
multiple platforms<br />
without special hardware<br />
for large real-time<br />
applications using<br />
multi-core processors.<br />
Enables segmentation to<br />
functional blocks, easier<br />
to control and debug.<br />
� Visual Studio Development<br />
Environment<br />
� INtime for Windows<br />
Compatibility for easy<br />
addition of an advanced<br />
HMI with the integration<br />
of Windows. Allows<br />
current INtime for<br />
Windows applications to<br />
run without modifi cation.<br />
30<br />
years<br />
������� �����������<br />
���� �� ������� ������ ����<br />
���������� �� ����� ���<br />
�� ����� ��������<br />
��� �� ����� ��������<br />
����������������<br />
���������������<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 13
conga-QA6<br />
Extended Temperature<br />
Range -40° ... +85°C<br />
3D Intel ® Graphics<br />
CAN Bus<br />
Formfactor Qseven Form Factor, 70x70 mm<br />
CPU <strong>Intel®</strong> Atom processor E6xx series with 1.3 GHz, 1.0 GHz and 600 MHz<br />
Integrated dual channel memory controller, up to 17.1 GByte/sec. memory bandwidth<br />
Integrated Intel ® HD Graphics with dynamic frequency up to 667 MH<br />
DRAM Up to 2 GByte onboard DDR2 memory with 667/800 MT/s<br />
Chipset Intel ® Platform Controller Hub EG20<br />
Ethernet Micrel ® Gbit Ethernet Phy KSZ9021RN<br />
I/O Interfaces 6x USB 2.0, 2x SATA, 1x SDIO, 3x PCIe, I²C Bus, 1x USB client, LPC bus, CAN Bus<br />
Mass Storage Onboard SATA Solid State Drive up to 32 GByte (optional)<br />
Sound Intel ® High Definition Audio (Intel ® HD Audio)<br />
Graphics Intel ® Graphics Core with 2D and 3D hardware Accelerator and dual independet display support.<br />
Video Decode Acceleration Full hardware acceleration for MPEG2, H.264, WMV9 and VC1<br />
Motion Video Support Single channel 112MHz LVDS transmitter, support for flat panels with 1x18 and<br />
1x24 bit data mapping up to resolutions of 1280x768 pixel<br />
Single channel SDVO interface, supports resolutions up to 1280x1024 pixel<br />
Dual independent display support<br />
congatec Board Controller Multi Stage Watchdog, non-volatile User Data Storage, Manufacturing and Board information, Board Statistics,<br />
I²C bus (fast mode, 400 kHz, multi-master), Power Loss Control<br />
<strong>Embedded</strong> BIOS Features OEM Logo, OEM CMOS Defaults, LCD Control, Display Auto Detection, Backlight Control, Flash Update, based on AMI Aptio UEFI<br />
Power Management ACPI 3.0 compliant, Smart Battery Management<br />
Operating Systems Windows ® XP, Windows ® XP embedded, Windows ® CE 6.0, LINUX, QNX<br />
Power Consumption Typ. application ~5 Watt @ 5V<br />
Temperature Range Operating: -40 to +85°C<br />
Storage: -40 to +85°C<br />
Humidity Operating: 10 to 90% r. H. non cond.<br />
Storage: 5 to 95% r. H. non cond.<br />
Size 70 x 70 mm (2¾" x 2¾")<br />
14 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
© 2010 congatec AG. All rights reserved. All data is for information purposes only. Although all the information contained within this document is carefully checked, no<br />
guarantee of correctness is implied or expressed. Intel and Intel Atom are trademarks of Intel Corporation in the US and other countries.
Watchdog<br />
Board Controller<br />
STM32F100R8T6b<br />
Hardware Monitoring<br />
and<br />
Fan Control Circuitry<br />
Engineering Tools / Accessories for Qseven<br />
I2C<br />
BC SPI<br />
GPIOs<br />
HDA I/F<br />
1x Channel LCD I/F (LVDS default XDP optional)<br />
6x USB Host 1x Client<br />
RTC<br />
<strong>Intel®</strong> Atom<br />
processor<br />
E6xx Series<br />
<strong>Intel®</strong> Platform<br />
Controller Hub<br />
EG20<br />
Signals<br />
Micrel® Control<br />
Gbit Ethernet<br />
Phy KSZ9021RN and<br />
SSD NandDrive<br />
OR SATA1 (optional)<br />
Uses SATA 1<br />
Management<br />
Fan<br />
Control Power<br />
SM Bus<br />
1x SDVO I/F<br />
conga-QEVAL 007001 Evaluation platform for Qseven<br />
1x SDIO/MMC<br />
PCIe Port 0<br />
CAN Bus<br />
BIOS<br />
(Flash)<br />
conga-QKIT 077455 Starterkit for Qseven including conga-QEVAL, conga-LDVI/EPI, conga-FPA1, Dual DVI-D ADD2 Card, ATX power supply...<br />
Qseven Mobility Kit 077460 Complete ready-to-use starter-kit for mobile applications<br />
conga-Mini Carrier Board 020710 Qseven mini carrier board for mobile applications.<br />
2x SATA (Port 0, 1)<br />
1x SATA0<br />
Memory Bus<br />
(533MHz or800MHz)<br />
Onboard DDR2<br />
Maximum 2GB<br />
conga-Mini Carrier cable kit 400015 Cable Kit for Qseven mini carrier board. Consisting of: 4x USB adapter cable, 1x LVDS display cable, 1x Feature connector cable, 1x HDMI to DVI adapter cable.<br />
Qseven MXM carrier board connector 400012 Socket for Qseven Carrier Boards, Foxconn PN:AS0B321-S78N-7F, standard type, SMT, lead free, 230 positions, 0.50mm pitch, 7.8mm height. 1pcs<br />
conga-LDVI/EPI 011115 LVDS to DVI converter board for digital flat panels with onboard EEPROM<br />
Single DVI-D ADD2 Card 058500 ADD2 display adapter card with single DVI-D digital output. Suitable for all Intel based platforms that support Serial Digital Display Outputs (SDVO) on the standard x16 PCI Express Graphics (PEG) port.<br />
Article PN Description<br />
conga-QA6-HSP-B 042745 Heatspreader, Standoff with 2.7mm bore hole<br />
conga-QA6-HSP-T 042740 Heatspreader, Standoff with 2.5mm threat<br />
conga-QA6-CSP-B 042735 Standard passive cooling solution, surface orange anodized with 1.8mm fins,<br />
bore hole bracket, threaded standoffs.<br />
CONTACT INFORMATION<br />
Gbit Ethernet<br />
Gbe MAC<br />
SPI<br />
3x x1 PCIe Links (Ports 1,2,3)<br />
LPC Bus<br />
congatec, Inc.<br />
6262 Ferris Square<br />
San Diego,<br />
CA 92121 USA<br />
+1 858-457-2600 Phone<br />
+1 858-457-2602 Fax<br />
www.congatec.com<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 15
FOCUS ON INTEL<br />
Intel Targets Digital Signage<br />
Next-generation digital signs will be interactive, connected and energy-efficient<br />
According to Jose Avalos, Intel’s director of retail and digital signage,<br />
there are more than 700 companies creating digital signage<br />
solutions worldwide. �at includes many large multinationals, but the<br />
majority are small players with fewer than 40 employees and under a<br />
million dollars in revenue. And while a fragmented market provides opportunities<br />
for innovation from many small players, it can be difficult<br />
for those small companies to scale production for broad distribution.<br />
Intel’s solution? A design platform based on standard PC hardware and<br />
software technologies to support both innovation and scalability.<br />
A recently announced collaboration between Intel, Microsoft and<br />
NEC Corporation intends to meet that need, with NEC integrating<br />
solutions and services (including content management and media<br />
distribution) with a digital signage platform based on <strong>Intel®</strong> Core<br />
i7 and Core i5 processors running Microsoft Windows <strong>Embedded</strong><br />
Standard 7. Intel’s digital signage model is a familiar one: �e client<br />
side consists of a display driven by an <strong>Intel®</strong> platform-based media<br />
player running content management software. On the network operating<br />
side, software manages the aggregation, delivery and scheduling<br />
of content to clients over a network such as WiFi, 3G or 4G.<br />
In October, Intel announced the Open Pluggable Specification<br />
(OPS) for digital signage, supported by Microsoft, NEC Display <strong>Solutions</strong><br />
and the Taiwan Digital Signage Special Interest Group. �e OPS<br />
articulates a specification for media players based on a small form factor<br />
and standardized electrical interfaces that would ultimately allow<br />
developers to create a family of digital signage products from entry level<br />
to high-end platforms with swappable media players to meet specific<br />
application needs. In a release, Avalos stated, “�e Open Pluggable<br />
Specification was created by Intel to address fragmentation in the digital<br />
signage market and simplify device installation, use, maintenance<br />
and upgrades. With the specification, digital signage manufacturers<br />
will be able to deploy interchangeable systems faster and in higher volumes,<br />
while lowering costs for development and implementation.”<br />
Avalos describes a vision he says Intel shares with its partners, in<br />
which next-generation digital signs are interactive, connected and<br />
energy efficient. In addition, these signs will be able to support anonymous<br />
video analytics that can report demographic information back to<br />
the advertiser on everything from the number of people who pass by or<br />
stop in front of the sign, and – using technologies such as facial recognition<br />
– report information back to the advertiser on audience gender,<br />
age group, and even facial expression, such as smiles. �is capability<br />
combines with remote content management technology to provide<br />
16 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
By Cheryl Coupé, Contributing Editor<br />
proof-of-play and proof-of-impression information used by the advertiser<br />
to measure return on investment (ROI). It can also be used to test<br />
creative and to adapt content to more closely match the sign’s audience.<br />
Intel is watching several technology trends in digital signage, including<br />
the use of gesture technology, which is deployed today mainly in<br />
gaming applications, but which will drive added interactivity with customers.<br />
Avalos says, “We see gesture being a key ingredient of digital<br />
signage over time. In fact we’re developing, as part of our commitment<br />
to the industry, proof of concepts that are intended to articulate advanced<br />
usage models to the industry and to motivate the industry.”<br />
Other trends to watch include energy efficiency at both the chip<br />
and platform level, as well as using remote management technologies<br />
to schedule maintenance and content updates and turn systems off<br />
to improve the life of the product and save energy. From a business<br />
model perspective, digital signage will become even more intelligent<br />
in the future. Traditional models will continue to be one-to-many, but<br />
advances in cell phone connectivity will be able to establish a oneto-one<br />
interaction with consumers. New retail applications even tie<br />
in social networking through in-store kiosks that allow consumers to<br />
consult their friends and then share purchasing decisions.<br />
�e biggest market trend Avalos sees is that “the multinationals<br />
have started to come in.” He believes this will create a nice balance between<br />
innovation from smaller players and increased investment from<br />
multinationals that will allow solutions to scale in volume and across<br />
geographies worldwide. Beyond this, he expects to see service providers<br />
such as telecoms entering the U.S. market over the next couple of years<br />
– a trend that is already occurring in Europe and China.<br />
Cheryl Berglund Coupé is Editor of EECatalog.com.<br />
Her articles have appeared in EE Times, Electronic<br />
Business, Microsoft <strong>Embedded</strong> Review and Windows<br />
Developer’s Journal and she has developed presentations<br />
for the <strong>Embedded</strong> Systems Conference and<br />
ICSPAT. She has held a variety of production, technical<br />
marketing and writing positions within technology companies<br />
and agencies in the Northwest.
FOCUS ON INTEL<br />
Wide-Area Airborne Surveillance<br />
(WAAS) Application Ideal for VPX<br />
Richard Lourette, chief scientist and principal investigator<br />
for ITT, worked closely with Dynatem to develop a<br />
VPX board for a payload sensor processing application that<br />
he describes as similar to the US Air Force’s “Gorgon Stare”<br />
program deployed last year. That program is a wide-area<br />
airborne surveillance (WAAS) system that supports multiple<br />
cameras and gives ground operators access to imagery<br />
from up to 12 different angles at the same time, at rates of<br />
16 MegaPixels per second for each camera. For the new system,<br />
demands were high and the project was underway at<br />
the same time that the draft standard for VPX was being<br />
developed. Together, Dynatem and ITT started the design,<br />
tracked changes as the standard developed, and spun the final<br />
board to meet the ratified standard.<br />
The resulting VPX single board computer (SBC) is based<br />
on a quad-core <strong>Intel®</strong> Xeon® processor for high performance.<br />
The board (which Dynatem dubbed “VPQ”) includes a 24port,<br />
10 Gigabit Ethernet switch that allows full-mesh<br />
backplane data-layer interconnectivity and up to eight VPQ<br />
SBCs to be integrated into a single chassis without the use of<br />
an additional switch board. A PCI Express switch provides<br />
connectivity to two fully capable PMC/XMC sites with extensive<br />
user I/O.<br />
Despite the challenges of designing the board in parallel<br />
with the standard, Lourette explains why VPX was the way<br />
to go for this demanding sensor application:<br />
� <strong>Intel®</strong> processor-friendly. With support for PCI Express<br />
and 10 Gigabit Ethernet, VPX is inherently <strong>Intel®</strong> CPUfriendly<br />
and Lourette needed the processing power of the<br />
high-performance quad-core Intel Xeon processor. When<br />
Apple bought P.A. Semi (in 2008), Lourette abandoned all<br />
efforts around Power Architecture. “I looked at that and said<br />
I’m not going to wait around for someone else to come up<br />
with something more powerful,” he stated. “At that point, all<br />
my designs were based on the <strong>Intel®</strong> platform and I’ve never<br />
looked back.”<br />
� Signaling in backplane. At the time, VPX was the only<br />
mil/aero backplane that could accommodate up to 6 Gigahertz<br />
of digital signaling in the backplane. ITT used XAUI,<br />
By Cheryl Coupé, Contributing Editor<br />
following the Open VPX standard for the VPX P1 connector<br />
for 10 Gigabit Ethernet. The additional slot width in the design<br />
also allows room for more cooling fins and higher signal<br />
integrity for the XMC sites.<br />
� GPGPU friendly. The VPX platform lends itself well to<br />
utilizing general-purpose computing on graphics processing<br />
units (GPGPU) technology. ITT and Dynatem repurposed<br />
NVIDIA and ATI graphics controllers for very high-performance<br />
number-crunching.<br />
The Dynatem VPQ is a high-performance single board<br />
computer (SBC) based on the 6U VPX (VITA 46) form factor.<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 17
FOCUS ON INTEL<br />
� Cluster-capable. The 10 Gigabit Ethernet switch built<br />
into the board supported flexible data flow topologies. Lourette<br />
was able to gang multiple CPUs in a system to create a<br />
supercomputing cluster that can process massive amounts<br />
of sensor data. The predecessor system to this design used a<br />
CompactPCI board that was limited to Gen 1 SATA IO signaling<br />
and had reached capacity at 65 watts per slot. The<br />
new VPX board can go up to 150 watts per slot for plenty of<br />
payload processing.<br />
Developing the VPX board<br />
was not without its<br />
challenges:<br />
� Cooling. Thermal design<br />
is somewhat less of a challenge<br />
now that the standard<br />
is ratified, but ITT and Dynatem<br />
worked closely with<br />
backplane and chassis vendors<br />
to develop a reference design<br />
that software developers could use that was electrically and<br />
software equivalent to the final system. Plenty of advanced<br />
thermal analysis ensured that ITT could power a system<br />
with seven boards and provide both air- and conductioncooled<br />
designs depending on the final application’s mission.<br />
� Mechanical. One of biggest – and unexpected – challenges<br />
was seating the cards in the backplane. Lourette<br />
explains, “It takes 250 pounds of force to insert a VPX card<br />
into the backplane. As thick as a 20-layer backplane is, it<br />
would bow like a credit card.” Reinforcements had to be applied<br />
to the backplane to make sure all the contacts were<br />
made. ITT and Dynatem also had to find metal front panel<br />
ejectors that were robust enough to handle the cards and<br />
BLOG<br />
Intel announces plans for a leading edge 22nm chip<br />
development facility in Oregon, while upgrading other manufacturing<br />
plants in the US.<br />
�e large, white tent near the D1D development fab<br />
parking lot belied the importance of coming announcement.<br />
Oregon’s leading congressmen and governor had assembled<br />
under this tent to share their support for Intel’s staggering<br />
investment in the US.<br />
�e event began with Bill Holt, Senior VP and GM at Intel,<br />
announced that Intel would be investing up to $8 billion in<br />
US facilities. Of particular interest to the Oregon high-tech<br />
community were plans for D1X, a new 22nm semiconductor<br />
18 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
“There’s technology available for us<br />
to go faster but the limit is 20 gigabits<br />
and then you have to go to optical.”<br />
New Semiconductor Fab Comes to Oregon<br />
Blog By John Blyler, Editorial Director<br />
adapted conduction-cooled levers to air-cooled cards so they<br />
could get them in and out without special tooling.<br />
Looking forward, Lourette is watching new VME technologies,<br />
including small form factor and optical developments.<br />
ITT has worked with Dynatem on 3U VPX and will be interested<br />
to watch developments in smaller form factors. About<br />
new small form factor developments he says, “It needs to be<br />
a little bit more mature before I can stand up a project on it,<br />
but I would certainly entertain that.” As for optical developments,<br />
he says, “Everything I work on is always pushing the<br />
envelope on data processing.”<br />
He continues, “I’m already<br />
saturating the VPX bus with<br />
data. VPX is only good to 20<br />
gigabits per channel and then<br />
it runs out of steam. Right<br />
now I’m at 10 gigabits per<br />
channel. There’s technology<br />
available for us to go faster but<br />
the limit is 20 gigabits and then<br />
you have to go to optical.”<br />
Cheryl Berglund Coupé is Editor of EECatalog.com.<br />
Her articles have appeared in EE<br />
Times, Electronic Business, Microsoft <strong>Embedded</strong><br />
Review and Windows Developer’s<br />
Journal and she has developed presentations<br />
for the <strong>Embedded</strong> Systems Conference and<br />
ICSPAT. She has held a variety of production, technical<br />
marketing and writing positions within technology companies<br />
and agencies in the Northwest.<br />
fab. �is fab would provide processor and related chip for the<br />
PC, mobile and embedded markets. Holt also said that two<br />
existing Oregon fabs – D1D and D1C – would be updated.<br />
�is is welcome news for Oregon, which is struggling with<br />
one of the highest unemployment rates in the nation. Holt<br />
said that the new 22nm fab would result in 6,000 to 8,000<br />
construction jobs over the next several years. Perhaps more<br />
important was that these projects would require 800 to 1,000<br />
new permanent high-tech jobs at the company.<br />
To read more, please visit: www.chipdesignmag.com/blyler
FOCUS ON INTEL<br />
Break Away with Intel ® Atom<br />
Processors<br />
A Guide to Architecture Migration<br />
Break Away with <strong>Intel®</strong><br />
Atom Processors: A<br />
Guide to Architecture Migration<br />
provides insight<br />
into architecture migration<br />
discussing real world software<br />
migration issues and<br />
highlighting them with case<br />
studies. �e book covers pertinent<br />
topics that are at the<br />
heart of the software migration,<br />
such as techniques to<br />
port code originally written<br />
for other processor architectures,<br />
as well as capturing the benefits of <strong>Intel®</strong> Atom platform<br />
technologies. �e reality is that architecture migration is not a<br />
one-size-fits-all activity and developers must understand all of<br />
the decisions that comprise a successful migration. �e authors<br />
deliver this information as a handbook to your software migration<br />
plan and project activities.<br />
Customer Comments<br />
“�is is an essential guide to working with the <strong>Intel®</strong> Atom<br />
processors that should be read by anyone who wants to understand<br />
how to use the platform effectively. �e book covers key<br />
concepts including power management, parallel applications,<br />
and machine virtualization, while providing a clear explanation<br />
of the system architecture and how to use it to its fullest advantage.”<br />
- Paul Krzyzanowski, CTO, OpenPeak Inc.<br />
“�is isn’t just a book about Intel extending its reach into the<br />
embedded market. In a way that is fresh and exciting, the authors<br />
have really understood that it is a powerful software methodology<br />
and tools eco-system that is needed for developers to make<br />
best use of the multicore and multithreaded world we now live in.<br />
�e style of this Atom book will be the standard against which all<br />
processor architecture books are measured from now on.”<br />
- David Stewart, CEO, CriticalBlue<br />
20 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
By Lori Matassa and Max Domeika<br />
“�is book is an excellent resource for someone trying to<br />
quickly come up to speed on the benefits, features, and differences<br />
of the <strong>Intel®</strong> Atom processors. As an added benefit, many<br />
of the concepts discussed throughout this book will provide the<br />
reader with generic implementation details, thereby serving as a<br />
good read for anyone new to the practical aspects of processor<br />
programming. �e author’s elaborate discussion on multicore<br />
programming techniques and tools is especially valuable, providing<br />
accurate details that are difficult to come by.”<br />
- Markus Levy, President, �e Multicore Association<br />
To learn more, visit: http://www.intel.com/intelpress/sum_<br />
ms2a.htm<br />
Lori Matassa is a Staff Technical Marketing<br />
Engineer for the <strong>Intel®</strong> <strong>Embedded</strong> and Communications<br />
Division and holds a BS in Information<br />
Technology. She has over 20 years of engineering<br />
experience developing software for embedded<br />
systems. In recent years at Intel she has contributed<br />
to Carrier Grade Linux, as well as the software enablement<br />
of multicore adoption and architecture migration for embedded<br />
and communication application.<br />
Max Domeika is a senior staff software engineer<br />
in the Developer Products Division at Intel, creating<br />
tools targeting the <strong>Intel®</strong> Architecture market.<br />
Max earned a BS in Computer Science from the<br />
University of Puget Sound, an MS in Computer<br />
Science from Clemson University, and a MS in<br />
Management in Science & Technology from Oregon<br />
Graduate Institute. Max recently authored, “Software Development<br />
for <strong>Embedded</strong> Multi-core Systems” from Elsevier Inc.
Start Your Adventure and Save!<br />
Secure your place at RSA® Conference 2011, and<br />
stay one step ahead of the challenges that face<br />
the information security industry.<br />
������������������������� with 200+ key<br />
industry-focused sessions, including network<br />
security, identity management, cloud computing,<br />
Web 2.0 security and more<br />
����������������������������� and constantly<br />
evolving compliance rules a� ect both the<br />
industry and your job<br />
������������������������������ and discover<br />
what’s new and upcoming from over 350<br />
exhibiting security companies<br />
�������������������������������������������<br />
to network with thousands of peers, industry<br />
leaders and security luminaries<br />
������������������������������<br />
SAVE<br />
$ 400<br />
Register by<br />
January 14 to secure<br />
your discount!<br />
Follow the Adventures of Alice & Bob at<br />
�������������������������������������<br />
©2010 EMC Corporation. All rights reserved. EMC, RSA, RSA Security, the RSA logo and the RSA Conference logo are registered trademarks of EMC Corporation in the United States and/or other countries. All other marks are trademarks of their respective companies.
STANDARDS WATCH<br />
Digital Audio, Standards, and Chips<br />
STANDARDS WATCH<br />
Digital audio provides a means for both professionals<br />
and consumers to record, modify, and play high-quality<br />
audio. USB-Audio-1.0, FireWire, and CobraNet are popular<br />
choices. But they’re showing shortcomings in terms of available<br />
bandwidth, limited synchronization, and lack of support<br />
on desktops and laptops.<br />
Two new standards are overtaking existing standards:<br />
USB Audio-2.0 and the AVB audio/video bridging standard<br />
based on IEEE 802. USB Audio-2.0 allows many more channels<br />
than Audio-1.0. It also has<br />
robust time-synchronization<br />
mechanisms. AVB is the IEEE<br />
audio-over-Ethernet standard.<br />
Unlike CobraNet, AVB<br />
isn’t licensed and is an open<br />
standard. It’s integrated with<br />
open time-synchronization<br />
and bandwidth-allocation<br />
protocols.<br />
USB Audio-2.0 and AVB<br />
both aim to cover all segments<br />
of the audio markets. Given that<br />
every PC has several USB-2.0 ports, USB Audio-2.0 is a likely<br />
candidate for the consumer and prosumer markets. For example,<br />
USB speakers and microphones could all be built around<br />
the USB Audio-2.0 standard. For a more complex system, such<br />
as a USB 7.1 surround-sound system, Blu-ray players and/or<br />
televisions would need to be equipped with a USB-2.0 port.<br />
But USB cables are typically restricted in length. In addition,<br />
USB Audio-2.0 is less likely to deliver audio over long distances.<br />
Because AVB is based on Ethernet rather than USB, it’s appropriate<br />
for longer-distance, networked distribution of audio<br />
in both pro-audio applications and multi-room consumer installation.<br />
It also is suitable for applications that don’t have a<br />
typical host-device interaction model, such as the automotive<br />
market. The Ethernet cable can simultaneously carry data (Internet,<br />
GPS) and audio between a variety of devices.<br />
22 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
“New standards are taking over<br />
(in digital audio) and both device<br />
vendors and OS designers are working<br />
hard to make systems compliant<br />
and hence interoperable.”<br />
By Henk Muller, Principal Technologist, XMOS<br />
All modern protocols are based around clock-recovery<br />
algorithms. The audio clock is carried with either the audio<br />
stream or using a separate time-exchange protocol. One clock<br />
in the network is denoted the master clock. Other devices will<br />
follow this clock by locally recovering the master clock. Depending<br />
on the protocol, just the audio clock is recovered or<br />
both an audio clock and a global common reference time.<br />
For the new protocols to be adopted, they must satisfy two<br />
criteria: low latency and bit perfect operation. Low latency<br />
can be achieved if the digital device<br />
can avoid buffering data<br />
by processing each sample of<br />
data in a predictable and deterministic<br />
manner. The only<br />
buffer that’s required is the<br />
one that smoothes out bursts<br />
of traffic (due to, for example,<br />
network switches). It also<br />
must buffer a small amount of<br />
data to cope with short- and<br />
medium-term fluctuations<br />
between the master and locally<br />
recovered clock. The minimal fundamental latency on USB-<br />
Audio-2.0 and AVB endpoints is around 250 {LC MU?}s (to<br />
store two frames of data). Typically, operating-system (OS)<br />
drivers and application programs add several milliseconds<br />
of delay.<br />
Bit-perfect operation would seem an easy target. Just<br />
make sure your protocol is implemented correctly! Audio<br />
protocols have many corner cases, however, and need to be<br />
timed to perfection in order to guarantee zero sample loss.<br />
In particular, there’s no retransmission or other form of<br />
backup in case a bit is corrupted or data is missing from a<br />
data stream. In addition, bit errors on the physical network<br />
(caused by electrical interference) will translate directly into<br />
an audible bit error in the signal. In this respect, the uptake<br />
of AVB could be seriously harmed if vendors push for Gigabit
Ethernet while customers still have old Category-5 wiring<br />
that isn’t up to Gigabit standards.<br />
Both interfaces support a variety of optional extras that<br />
can be implemented in the digital domain, such as sample-rate<br />
conversion, mixing, and equalization. In order to<br />
customize their products, manufacturers of digital-audio devices<br />
will look at spare processing capacity (DSP capability)<br />
to implement those features. The addition of these features<br />
cannot interfere with the timing of other parts, as that may<br />
compromise bit-perfect operation.<br />
Chip vendors have chosen a variety of methods for implementing<br />
those new audio standards: application-specific<br />
standard products (ASSPs), field-programmable gate arrays<br />
(FPGAs) configured using a hardware description language<br />
(HDL), or processors programmed in a high-level language.<br />
Typically, ASSPs implement a specific device and hence only<br />
a subset of the protocol. If this subset matches the requirements,<br />
an ASSP provides an out-of-the-box solution. However,<br />
it’s difficult to add extensions to the subset of the protocol,<br />
follow changes and improvements to standards, or differentiate<br />
a design. An example ASSP-based solution is offered by<br />
C-media. A single chip offers an out-of-the-box, two-channel<br />
synchronous USB-Audio-2.0-to-I2S terminal.<br />
Programmable solutions, such as FPGAs or processors,<br />
can be upgraded when designing future products or even<br />
field-upgraded to add functionality. For example, Xilinx and<br />
BLOG<br />
Being a Green Programmer<br />
Blog By Gary Stringham, <strong>Embedded</strong> Systems Expert<br />
Much of what I read about Green Design involves<br />
building the product with recycled or renewable materials<br />
and consumables, and making the product recyclable<br />
at the end of its life. Hardware engineers can contribute<br />
to the green effort by influencing how the product is<br />
made and what it is made of. But what about software<br />
and firmware engineers? Their end product is a bunch<br />
of intangible bits flowing through hardware. Those bits<br />
don’t pollute the environment or end up in landfills. So<br />
how can they be green in their designs?<br />
I came across the answer while reading about industrial<br />
designer Gadi Amit’s unusual approach to Green<br />
Design: make products so desirable that people just<br />
won’t want to throw them away. (Sacks, Danielle. “What’s<br />
Wrong With Green Design.” Fast Company Oct. 2010:<br />
166-69. FastCompany.com. 1 Oct. 2010. Amit gave an<br />
example that the 8-year-old Palm Zire is still used today<br />
because it works well and people have an emotional connection<br />
with it. He then observed that in contrast, PCs at<br />
first work fast, but after a while get bogged down with old<br />
STANDARDS WATCH<br />
Broadcom offer an FPGA implementation of AVB that can<br />
be modified once the IP has been purchased. XMOS offers<br />
a processor that can implement the I/O, signal processing,<br />
and USB or Ethernet stacks—all using a high-level language.<br />
As such, reference designs and software can be downloaded<br />
(at no cost). The software can be modified at will in order to<br />
customize the design or implement extra functionality.<br />
These are exciting times for digital audio. New standards<br />
are taking over and both device vendors and OS designers<br />
are working hard to make systems compliant and hence<br />
interoperable. (Apple, for example, natively supports USB-<br />
Audio-2.0 since MacOSX 10.6.) The new standards will<br />
support bit-perfect transmission of many 24-bit channels at<br />
96 or even 32-bit channels at 192 kHz. All of these protocols<br />
provide clock-recovery schemes and are designed to provide<br />
a low-latency solution.<br />
Henk Muller is currently the principal technologist<br />
at XMOS Ltd. In that role, he has been<br />
involved in the design and implementation of<br />
audio and other real-time protocols. Previously,<br />
Muller worked in academia for 20 years in computer<br />
architecture, compilers, and ubiquitous<br />
computing. He holds a doctorate from the University of Amsterdam.<br />
programs, temporary files, and bloated registries. Users<br />
often get so frustrated with their slow machines that they<br />
simply buy new ones and throw away the old.<br />
In other words, Amit is encouraging us to make the<br />
product function so well that users will not want to throw<br />
it away. What a novel and green concept! Let me cite two<br />
examples from Hewlett-Packard.<br />
The Hewlett-Packard 12C financial calculator was<br />
green before green was mainstreamed. It was introduced<br />
in 1981 and is still being sold in stores today, 29 years<br />
later. It is a well-designed product that well suits the<br />
financial community. HP did not set out to design a green<br />
financial calculator; they set out to design a top-notch<br />
calculator that exceeds the needs of their customers.<br />
To read more, please visit: http://eecatalog.com/stringham/<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 23
Delivering in-depth<br />
product, technology and<br />
design information<br />
www.embeddedintel.com<br />
<strong>Embedded</strong><br />
<strong>Solutions</strong><br />
Intel SUMMER 2010<br />
®<br />
Processors Merge Form Factors<br />
on Wireless<br />
Roundtable: Form Factor<br />
Versus Features<br />
Multicore Migration<br />
Guide: Part II<br />
Enhancements for Medical<br />
and Retail<br />
Computational Powerhouse<br />
Hidden in Island Jungle<br />
www.embeddedintel.com<br />
Gold Sponsors<br />
<strong>Embedded</strong><br />
<strong>Solutions</strong><br />
Intel Winter 2011<br />
®<br />
<strong>Embedded</strong> Virtualization—the Key to<br />
Real-time Determinism<br />
Combining x86 Architecture with FPGAs<br />
Optimizing the Virtual Environment<br />
Intel ® Atom Processor E6xx Series<br />
Reaches the SUMIT<br />
VPX Apps for Wide-Area Airborne<br />
Gold Sponsors<br />
www.embeddedintel.com<br />
<strong>Embedded</strong><br />
<strong>Solutions</strong><br />
Intel SPRING 2010<br />
®<br />
<strong>Embedded</strong> Processors<br />
Go Dual Core<br />
Intel ® Atom Processor<br />
Goes Lower Power For Mobile<br />
Energy Saving Systems Enabled<br />
by Intel ® Atom Processor<br />
Best Practices for Hardware/Firmware<br />
Design<br />
www.embeddedintel.com<br />
Gold Sponsors<br />
<strong>Embedded</strong><br />
<strong>Solutions</strong><br />
Intel Fall 2010<br />
®<br />
Special Resource Catalog Issue<br />
Power in Power-Limited Environments<br />
Chips See <strong>Embedded</strong> Diversity<br />
Handshakes & Acquisitions<br />
Ensuring Power Substation Reliability<br />
at the Highest Level<br />
www.embeddedintel.com<br />
Gold Sponsors<br />
Intel WINTER 2009<br />
®<br />
Gold<br />
Sponsors<br />
Featured Products<br />
Emerson Network Power: MATXM-CORE-<br />
411-B MicroATATX Motherboard<br />
Congatec: conga-QA6<br />
Viking Modular <strong>Solutions</strong>’ SATADIMM<br />
<strong>Embedded</strong><br />
<strong>Solutions</strong><br />
Industrial Power Savings<br />
Architecture Migration to Multicore<br />
Invites Software Decisions<br />
<strong>Embedded</strong> Instrumentation has Intel ® Xeon ® Processor Covered<br />
802.3 PoE and Intel ® Atom Processor<br />
Advance Devices<br />
Solution Providers Forum<br />
Articles from companies providing important solutions for engineers<br />
and embedded developers utilizing <strong>Embedded</strong> Intel ® Processors<br />
www.embeddedintel.com
<strong>Embedded</strong><br />
<strong>Solutions</strong><br />
Intel Fall 2010<br />
®<br />
Special Resource Catalog Issue<br />
Power in Power-Limited Environments<br />
Chips See <strong>Embedded</strong> Diversity<br />
Handshakes & Acquisitions<br />
Ensuring Power Substation Reliability<br />
at the Highest Level<br />
www.embeddedintel.com<br />
Gold Sponsors<br />
Featured Products<br />
Emerson Network Power: MATXM-CORE-<br />
411-B MicroATATX Motherboard<br />
Congatec: conga-QA6<br />
Viking Modular <strong>Solutions</strong>’ SATADIMM<br />
<strong>Embedded</strong><br />
<strong>Solutions</strong><br />
Intel Winter 2011<br />
®<br />
<strong>Embedded</strong> Virtualization—the Key to<br />
Real-time Determinism<br />
Combining x86 Architecture with FPGAs<br />
Optimizing the Virtual Environment<br />
Intel ® Atom Processor E6xx Series<br />
Reaches the SUMIT<br />
VPX Apps for Wide-Area Airborne<br />
Gold Sponsors<br />
www.embeddedintel.com<br />
Annual Resource Catalog Quarterly Magazine<br />
A network dedicated to the needs of the<br />
global <strong>Embedded</strong> Intel ® community<br />
www.embeddedintel.com<br />
For a FREE subscription visit:<br />
www.embeddedintel.com<br />
Monthly Email Newsletter<br />
Advertising —Print, Online,<br />
Email Newsletter, & Reprints:<br />
Clair Bright<br />
+1 415-255-0390, ext 15<br />
cbright@extensionmedia.com
SPECIAL FEATURE<br />
Intellectual Property<br />
Trouble In The Patent Office<br />
Why current practices have green IP proponents seeing red; how to limit<br />
the impact of trolls.<br />
The next wave of patents being filed in the United States involves<br />
efficiency and power rather than simply speeding up<br />
performance of chips.<br />
But the patent system is moving so slowly these days that<br />
startups are unable to talk about what exactly they do, for<br />
fear of having their IP stolen, extending the time it takes to<br />
come out of stealth mode and arrange deals with potential<br />
partners who could speed the delivery of new technologies to<br />
market. And in some cases, offshore<br />
companies aren’t even<br />
filing patents in the United<br />
States anymore.<br />
Low-Power Engineering<br />
sat down to discuss the U.S.<br />
patent system with Bijal Vakil,<br />
partner at White & Case<br />
and an expert in intellectual<br />
property. What follows are<br />
excerpts of that conversation.<br />
LPE: What’s wrong with<br />
the U.S. patent system?<br />
Vakil: We need patent applications to proceed to final<br />
determination quicker. Right now it takes about three years.<br />
That’s a long time. The U.S. Patent Office could definitely<br />
expedite the process. The Patent Office generates a surplus<br />
of revenue and Congress siphons that off. The Patent Office<br />
is supposed to encourage small entrepreneurs and to obtain<br />
a monopoly in timely manner. It’s like a tax for innovators.<br />
When you add in a delay that certainly doesn’t encourage<br />
innovation.<br />
LPE: But when there were fewer patents there was incredible<br />
innovation in the semiconductor industry. Now we’re<br />
seeing “patent trolls” everywhere. How do you fix that?<br />
Vakil: The patent system was important enough that<br />
it was included in the Constitution. I don’t believe the in-<br />
26 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
“There is the natural tension between<br />
the Silicon Valley companies, which<br />
want a more effective patent system,<br />
and the pharmaceutical companies,<br />
which have a different interpretation.”<br />
By Ed Sperling, Contributing Editor<br />
novation of large companies has been impacted by patent<br />
control. The non-practicing entities have added to litigation<br />
costs, which are an expense to companies. But companies<br />
are innovating more quickly than ever, both in software and<br />
hardware.<br />
LPE: So what do you do with these so-called patent trolls?<br />
Vakil: There should be some patent reform legislation<br />
that brings certainty to a number<br />
of areas with respect to<br />
non-practicing entities. For<br />
example, the patent damages<br />
laws are not clearly defined<br />
with respect to non-practicing<br />
entities. Some clarification<br />
of those rules would discourage<br />
the volume of these cases.<br />
You also could force changes<br />
in venue. A number of these<br />
cases are brought in the Eastern<br />
District of Texas. That<br />
increases the cost for a company<br />
in Silicon Valley, and it allows a non-practicing entity to pick<br />
a jurisdiction it perceives to be friendlier to patent holders. A<br />
lot of these companies don’t even have offices.<br />
LPE: Any other areas?<br />
Vakil: Yes, the Eastern District of Virginia and the Western<br />
District of Wisconsin.<br />
LPE: Is the volume of patents increasing?<br />
Vakil: No. In 2009 there was a trend downward for new<br />
patents. At the same time, there has been an increase in other<br />
jurisdictions, including China. There are several reasons for<br />
this. First, companies are trying to limit their legal expenses<br />
so they’re cutting back on patent applications. Second, it may<br />
be a result of a decrease in the workforces at innovative com-
panies. And third, patent holders are no longer viewing the<br />
United States as the center of the universe. You file patent<br />
applications in countries where you expect your business to<br />
grow. The role of the United States is diminishing, and patent<br />
holders see markets in China and India on the rise. With<br />
patents you have to anticipate what markets are going to be<br />
like in 15 years.<br />
LPE: There’s also a lot of concern about whether patents<br />
filed in China will be upheld, right?<br />
Vakil: Yes, but steps are being taken to make it easier to<br />
enforce patents. Countries such as India and China are big<br />
markets for U.S. companies.<br />
LPE: Let’s go back to the U.S.<br />
Patent System. What exactly<br />
needs to be fixed?<br />
Vakil: Improving the IT<br />
infrastructure is paramount.<br />
One of the problems is that<br />
the Patent Office grants<br />
many patents that should<br />
not be granted. That’s where<br />
the problem is with nonpracticing<br />
entities trying to<br />
cajole money out of larger<br />
defendants. By allowing the<br />
Patent Office to maintain its<br />
own revenue base it could<br />
invest in an IT infrastructure<br />
that would allow more accurate<br />
‘Prior Art’ searching before patents are granted. That would<br />
improve the integrity of the patents and ultimately benefit<br />
the economy. Second, an improved IT infrastructure could<br />
help the application. If the Patent Office’s database was made<br />
public then the person applying for a patent would have that<br />
same resource and know what the examiner would be looking<br />
at. With that information they could tailor their patent,<br />
either narrower or broader, making the whole system more<br />
efficient and improving the time it takes to approve a patent.<br />
LPE: If it is improved, will more patents be filed or will<br />
patents simply be quicker?<br />
Vakil: Both. Over the long term, increased confidence in<br />
the patent system will lead to companies wanting to innovate<br />
“The patent system was<br />
important enough that it was<br />
included in the Constitution. I<br />
don’t believe the innovation<br />
of large companies has been<br />
impacted by patent control.”<br />
in the United States. That would create jobs and improve the<br />
United States’ relevancy. If we don’t, some other country will.<br />
LPE: In the United States, who is most affected by the<br />
patent system?<br />
Vakil: Silicon Valley. It’s estimated that 60% of the United<br />
States’ exports relate to intellectual property, and more than<br />
70% of the patents originate in Northern California. Fixing<br />
the patent system can encourage innovation and ultimately<br />
create jobs.<br />
LPE: Is there any progress?<br />
Vakil: There was a movement in Congress, led by a number<br />
of Silicon Valley heavyweights, to get patent reform<br />
legislation approved. It always<br />
gets pushed off. There is the<br />
natural tension between the<br />
Silicon Valley companies,<br />
which want a more effective<br />
patent system, and the pharmaceutical<br />
companies, which<br />
have a different interpretation.<br />
LPE: What are pharmaceutical<br />
companies looking for?<br />
Vakil: The public reports<br />
show the pharmaceutical<br />
industry is happy with the<br />
system as it is.<br />
LPE: Because it keeps out smaller companies?<br />
Vakil: That’s correct. But while patent reform isn’t high<br />
priority, the appellate courts are aware of public reaction<br />
to patent reform and have taken a much more active role in<br />
dealing with these issues that would have been addressed in<br />
patent reform regulation.<br />
Ed Sperling is Contributing Editor for <strong>Embedded</strong><br />
<strong>Intel®</strong> <strong>Solutions</strong> and the Editor-in-Chief of the<br />
“System Level Design” portal. Ed has received numerous<br />
awards for technical journalism.<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 27<br />
SPECIAL FEATURE Intellectual Property
SPECIAL FEATURES<br />
Connectors<br />
Intel ® Atom Processor E6xx<br />
Series Reaches the SUMIT<br />
Advanced connector unlocks small-form factors<br />
Actively enabling an embedded ecosystem, the <strong>Intel®</strong><br />
Atom processor E6xx series-based platform features<br />
flexibility, performance, low power, an extended operating<br />
temperature, long life, and widely available software support.<br />
Introduced at this year’s Intel Developer Forum (IDF),<br />
the <strong>Intel®</strong> Atom processor E6xx series® and <strong>Intel®</strong> Platform<br />
Controller Hub EG20T together comprise the Tunnel<br />
Creek platform. This platform is ideal for small-form-factor<br />
systems targeting industrial, military/aerospace, transportation,<br />
security, and pipeline/utility applications—whether<br />
portable or fixed. As with any embedded system, however,<br />
a variety of I/O support is essential for “real-world” interfacing.<br />
Complicating this issue, an engineer who selects or<br />
designs a single-board computer (SBC) can face the problem<br />
of last-minute specification changes from either the customer<br />
or his or her own marketing department. So what’s<br />
the best way to offer flexible I/O expansion options without<br />
adding cost and space “taxes” to the processor board?<br />
For deeply embedded systems, the Stackable Unified<br />
Modular Interface Technology (SUMIT) connector is the<br />
solution. It provides a way for modular I/O expansion to tailor<br />
systems to a wide range of requirements. The connector<br />
also serves as an insurance policy to protect against feature<br />
creep or changes. SUMIT provides an advanced, robust,<br />
and reliable multi-bus, board-to-board interface for stackable<br />
systems regardless of SBC form factor. This interface is<br />
optimized for Intel’s new generation of low-power, high-integration<br />
platforms. SUMIT can be implemented on various<br />
board sizes for a single board (as a mezzanine) or for multiple<br />
boards without the need for a card cage.<br />
The hallmark of SUMIT is how easily this expansion architecture<br />
can be used to support a blend of high-bandwidth PCI<br />
Express lanes and Universal Serial Bus (USB) ports as well as<br />
lower-speed multiplexed and serial buses. Using SUMIT reduces<br />
cost and bulk while increasing mounting and packaging<br />
options for small-form-factor embedded systems.<br />
28 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
By Robert A. Burckle, vice-president, WinSystems Inc.<br />
Figure 1: High-speed connectors allow single or multiple I/O<br />
cards to be stacked “piggy back” fashion.<br />
SUMIT, which is pronounced “sum it,” uses one or two<br />
52-pin, high-speed connectors that allow either a single or<br />
multiple I/O cards to be stacked “piggy back” on top of each<br />
other (see Figure 1). Signal-integrity test results demonstrate<br />
that a stack of modules will support data rates of 5 GT/s,<br />
which is required for PCI Express 2.0 and USB 2.0. In the “A”<br />
connector, SUMIT supports one x1 PCI Express lane and four<br />
high-speed USB 2.0 channels as well as the Low Pin Count<br />
(LPC) Bus, SPI/uWire, and SMBus/I²C signaling. SUMIT-A<br />
fits Tunnel Creek like a glove, considering that several of the<br />
four PCIe x1 lanes will be consumed on the host SBC itself.<br />
The “B” connector supports five additional x1 PCI Express<br />
lanes as well as additional power, ground, and control<br />
signals. SUMIT-B and SUMIT-AB target higher-end platforms<br />
like <strong>Intel®</strong> Atom processor N450/D510 and even <strong>Intel®</strong><br />
Core i7 “Calpella” platforms. Or they can be used with a<br />
PCIe switch to fan out the integrated PCIe lanes of the Intel<br />
Atom processor E6xx series-based platform. SUMIT-based<br />
I/O modules each use one or more of the bus signals and pass
unused resources further up the I/O stack for other cards to<br />
utilize.<br />
Currently, two I/O-expansion-board standards are defined<br />
and available in the market: SUMIT-ISM and Pico-I/O. Both<br />
are small, rugged, easy to use, and scalable. SUMIT-ISM is<br />
Figure 2: Pico-I/O is an even smaller, 60-x-72-mm module<br />
standard.<br />
BLOG<br />
How does one really decide between a hardware vs. a<br />
software design implementation? Early in a product’s life<br />
cycle, designers must consider high-level architectural<br />
issues like power, performance, area and cost. But a highlevel<br />
view is just an abstraction of the – as yet uncertain<br />
– lower-level details. A good designer understands both<br />
the high- and the low-level points of view.<br />
At the lowest level in a hardware-software trade-off<br />
decision, the designer must decide what system functionality<br />
will be realized in software blocks (via a processor)<br />
and what will be accomplished by hardware devices.<br />
Here’s an example common to most embedded engineers:<br />
for (i=0; i
SPECIAL FEATURE<br />
Virtual Optimization<br />
Optimizing the Virtual Environment<br />
New Tools that Help Industry Dance to Your Enterprise Tune<br />
According to a recent industry survey, more than 46% of<br />
CIOs reported that 51% to 85% of their data-center servers<br />
have been virtualized. By the end of 2012, Gartner is predicting<br />
that about 50% of the x86 architecture workloads—representing<br />
approximately 58 million deployed machines—will be running<br />
in virtual machines. IDC analyst Cindy Borovick believes that<br />
2010 was the first year in which the number of deployed virtual<br />
servers outnumbered deployed physical servers.<br />
Clearly, the virtualization megatrend is delivering real, tangible<br />
cost savings and operational improvement. Yet CIOs are left<br />
with a number of questions: How do the major technology trends<br />
impact my virtualization strategy? How do I get my conversation<br />
with industry aligned to my objectives? What is the technical<br />
solution that delivers the lowest possible cost of the virtualized<br />
operation for my applications?<br />
In the early stages of the server-virtualization journey, there<br />
were significant technology differences between the various<br />
hypervisor solutions. �ey ranged from pure performance differences<br />
to the available administrative and management tools.<br />
Over time, the marketplace has eliminated these material differences.<br />
As a result, CIOs can focus on other technology advances,<br />
which will significantly impact current architectures and reduce<br />
costs. �ree major trends and potential solutions are fundamentally<br />
impacting the virtualization discussion:<br />
�� ��������������������������������������������������<br />
interoperability<br />
�� ������������������������������������������������������<br />
families<br />
�� �������������������������������������������<br />
�� �����������������������������������������������������<br />
identify the best solution for their environments<br />
Hypervisor Interoperability<br />
In November 2009, the Open Virtualization Format (OVF)<br />
specification was published by IBM, VMware, Microsoft, Citrix,<br />
HP, Red Hat, and other industry leaders. �is specification defines<br />
a standards-based, portable format so that enterprises can<br />
deploy the virtual machine (VM) in any hypervisor that supports<br />
OVF. It therefore creates a platform-independent, efficient, extensible,<br />
and open packaging and distribution format for VMs.<br />
30 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
By Bob Carlson, EVP, Criterion HPS<br />
Although the support for the full scope of this specification<br />
hasn’t yet been fully implemented by the leading hypervisor<br />
vendors, I believe VM packaging will be independent of the<br />
hypervisor in the near future. CIOs will then be able to move<br />
application workloads frictionlessly to the lowest-cost processing<br />
environment.<br />
Dense Computing<br />
With the emergence of the new <strong>Intel®</strong> microarchitecture<br />
(codenamed Nehalem), the industry has begun to build and deliver<br />
very dense servers. Over this period, the number of vCores<br />
per motherboard has increased from four in early 2008 to 64<br />
vCores (and more) per motherboard. �is dramatic increase in<br />
processing density has coincided with a dramatic reduction in<br />
power requirements, which allows these very dense servers to be<br />
air-cooled. �is technology has the potential to achieve considerable<br />
infrastructure consolidation and cost reductions.<br />
64-bit Architectures<br />
�e adoption and migration to the 64-bit architecture has<br />
been one of the most significant migrations that have never been<br />
discussed. �is migration is being led by the tremendous success<br />
of Windows 7 and the availability of multicore processing<br />
chips. If you spend any time on the Internet searching on the<br />
benefits of 64-bit architectures, the preponderance of comments<br />
focuses on the ability to address and use more memory. While<br />
this is indeed an important benefit, I don’t believe it is the most<br />
important advantage for CIOs. Arguably, the most critical reason<br />
to move to 64-bit architectures is to ensure that the application<br />
portfolio takes full advantage of the processing and cost benefits<br />
of the new, dense-server solutions. �e result of this server<br />
density is significantly reduced costs due to server rack consolidation.<br />
More importantly, the physical architecture’s complexity<br />
is tremendously reduced. Eliminating this complexity has a multiplying<br />
effect on cost reductions, as every component of the<br />
system life cycle becomes easier to manage.<br />
Individualized Toolsets<br />
With all of the technology breakthroughs and new capabilities<br />
being delivered by industry, CIOs need new tools and methods to<br />
regain control of their decision-making processes. Like all new<br />
technologies, “it depends” what the specific benefits will be in
each individual environment. New, individualized toolsets are<br />
emerging to help guide CIOs as they converse with industry.<br />
�ese toolsets also complement the generic solutions resulting<br />
from hypervisor interoperability, dense computing, and 64-bit<br />
architectures.<br />
One example is V CO (pronounced V-COE), which helps the<br />
CIO communicate his or her requirements to industry. He or she<br />
also can evaluate and compare industry’s recommendations in a<br />
straightforward and meaningful manner.<br />
�is management tool allows alternative solutions to be evaluated<br />
in order to help the CIO select the solution or solutions<br />
that provide the lowest possible total cost of ownership future. It<br />
creates the following:<br />
�� ��������������������������������������������������������<br />
different virtualization stacks<br />
�� ������������������������������������������������������������<br />
VM for the specific application<br />
�� �������������������������������������������������������<br />
unique operational requirements are included<br />
�e V CO is based on the insight that the VM file contains<br />
all of the business and operational value. Yet the infrastructure<br />
required to run and operate the VM file has become a pure<br />
commodity. Once the VM is created, the historical dependence<br />
between the hardware, operating system, and application has<br />
been permanently severed. With the emergence of the OVF,<br />
VMs can simply be deployed to any available virtualized infrastructure.<br />
In essence, virtualization turns the underlying<br />
infrastructure into a pure commodity. It’s therefore reasonable<br />
Figure: The V CO framework is depicted here.<br />
for the CIO to demand that this infrastructure deliver the lowest<br />
possible total cost of ownership to the organization.<br />
�is model establishes the “cost per VM baseline,” which is<br />
used to compare and contrast new technologies that can deliver<br />
superior and ongoing cost savings. �e framework is open and<br />
extensible to meet future requirements. To evaluate alternative<br />
VM infrastructures, it begins by determining the V CO for the<br />
target application (see the Figure). In modern data-center operation,<br />
the inputs required for the tool should all be available. Once<br />
the application’s V CO is determined, the CIO can provide it to<br />
industry and benefit from the open competition.<br />
Now, the virtualization conversation between the CIO and<br />
industry can be aligned with the CIO’s objectives. �e newest<br />
and greatest innovations can be assessed in the context of how<br />
they reduce the total cost of operation for an application stack.<br />
By adopting these types of tools, the CIO will reward marketplace<br />
innovations that continually reduce the cost of running<br />
and operating a VM. For example, moving the emphasis to total<br />
system optimization and away from any individual-solution<br />
component will result in the lowest total cost of operation for the<br />
application.<br />
�e V CO drives the optimization of these tradeoffs, resulting<br />
in the lowest operational cost per VM for the organization.<br />
Understanding these tradeoffs and the impact on the cost of<br />
running and operating a single VM provides the CIO with new,<br />
critical insights into the operation. It also puts him or her back<br />
in control of understanding how industry innovation can reduce<br />
ongoing operational costs.<br />
Together, hypervisor interoperability, dense computing, and<br />
the migration to 64-bit architectures have defined the server-virtualization<br />
journey. �ey also have allowed CIOs to focus on how<br />
these advances are impacting current architectures and reducing<br />
costs. �e resulting solutions have propelled virtualization to the<br />
forefront of the industry and made the case for distinct benefits.<br />
As CIOs wade through the solutions advanced by these trends,<br />
however, the search continues for tools that help to identify the<br />
best virtualization strategy for their unique enterprise environments.<br />
�ese tools also should get industry to respond to their<br />
specific requirements.<br />
1Bernd Harzog, “Gartner Projects Server Virtualization to Grow from 16%<br />
to 50% of Workloads by 2012,” �e Virtualization Practice,October 21, 2009.<br />
(http://www.virtualizationpractice.com/blog/?p=2496)<br />
2Sean Michael Kerner, “Virtual Servers Top Physical Ones, WAN Opti-<br />
mization Soars: IDC,” Datamation, April 28, 2010. (http://itmanagement.<br />
earthweb.com/datbus/article.php/3879246/Virtual-Servers-Top-Physical-<br />
Ones-WAN-Optimization-Soars-IDC.htm)<br />
3Copyright © 2009 Distributed Management Task Force, Inc. (DMTF). All<br />
rights reserved. Document Number: DSP0243 Date: 2009-02-22<br />
Mr. Carlson’s 26 year career has paralleled the growth<br />
of the information technology industry from back office<br />
automation to strategic business process enabler.<br />
Both as an IBM executive and as the CEO of 2 start-up<br />
companies, Mr. Carlson has specialized in developing<br />
and implementing leading-edge business solutions that<br />
provide competitive advantage through the systematic exploitation of<br />
technology. He is now focused on the rapidly growing optimized solution<br />
server market and leading business development and product initiatives<br />
for Criterion HPS.<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 31<br />
SPECIAL FEATURE Virtual Optimization
TECHNOLOGY APPLICATION<br />
<strong>Embedded</strong> Virtualization<br />
<strong>Embedded</strong> Virtualization — the<br />
key to Real-time Determinism in<br />
Multi-OS Systems<br />
Virtualization is a commonly used term, but not all virtualization<br />
is the same. If not implemented correctly in embedded<br />
applications, performance and reliability will suffer and your design<br />
goals may not be met.<br />
�e key difference between embedded applications and office or<br />
enterprise computer systems is the need for determinism. Generalpurpose<br />
virtualization approaches, such as those that enable servers<br />
to run multiple copies of the same general purpose operating system<br />
(GPOS), frequently virtualize the entire machine environment with<br />
the goal of maximizing the utilization of the CPU. Unfortunately,<br />
this comes at the expense of responsiveness to external events, making<br />
this approach unusable for time-critical applications.<br />
In contrast, embedded virtualization is engineered to respond<br />
to hardware-generated signals deterministically. It does this by<br />
partitioning the platform for GPOS and real-time operating system<br />
(RTOS) use, only virtualizing what is required. �is enables<br />
real-time events to be serviced within the required timing, while<br />
the GPOS keeps running in its usual way.<br />
�e release of the <strong>Intel®</strong> 80386 processor twenty years ago introduced<br />
support in the hardware to make switching the context of tasks<br />
more reliable and easy to implement. A group of engineers, some of<br />
whom are still with TenAsys today, devised ways of using that feature<br />
to make DOS, the most popular GPOS at the time, run alongside the<br />
iRMX real time operating system. �at day the concept of real-time<br />
embedded virtualization was born. Even with the HW features, most<br />
of what was achieved was done through a technique called para-virtualization,<br />
an adaption of the OSes and drivers to allow them to work<br />
together.<br />
As GPOSes evolved and DOS was replaced by Windows as the<br />
desktop market leader, it became increasingly evident that GPOSes<br />
like Windows were never designed to service events with a fixed<br />
priority and hence could not respond deterministically to real-time<br />
control of machines. Even with faster processors, Windows can not<br />
be counted on to respond to external events with predictable timing.<br />
Some software vendors have attempted to deal with Windows’<br />
lack of determinism by developing special Windows drivers that<br />
do I/O processing along with responding to hardware interrupts,<br />
but these techniques have proved unreliable and typically must be<br />
re-engineered each time a new version of Windows comes out.<br />
To solve the problem of getting real-time tasks and Windows<br />
to co-reside on an embedded system, the para-virtualization tech-<br />
32 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
By Kim Hartman, TenAsys Corporation<br />
niques that had worked with DOS were evolved to run updated<br />
RTOSes, like TenAsys’ INtime, alongside Windows to maintain<br />
the integrity of the real-time environment and get the benefit of<br />
the highly evolved graphical interface that Windows provides.<br />
As the software evolved, so did the support for virtualization that<br />
is built into the underlying processor hardware. Intel has added a set<br />
of hardware assists called <strong>Intel®</strong> Virtualization Technology to their<br />
processors which are used to accelerate the performance of virtualization<br />
software.<br />
Using these techniques and hardware aids, embedded virtualization<br />
software partitions resources, such as CPU cores (in the case<br />
of today’s multi-core systems), RAM, interrupts, and I/O, between<br />
RTOSes and GPOSes. Only shared and emulated resources need<br />
to be virtualized. Each OS is guaranteed direct physical access to<br />
its dedicated I/O, interrupts, RAM and CPU cycles. �e RTOS I/O<br />
response cycles take absolute priority, without placing any limits on<br />
the functionality of the Windows side of the system.<br />
In multi-core processor systems, the virtualization manager<br />
can give the real-time software guaranteed ownership of a CPU<br />
core – delivering maximum performance and removing the requirement<br />
that the RTOS be designed specially to work with<br />
Windows. With the proper embedded virtualization software, virtually<br />
any RTOS can work alongside Windows without the need<br />
for making any modifications to the RTOS or legacy real-time application<br />
software that has been designed to run with it.<br />
Now, embedded system products that used to incorporate separate<br />
computers for real-time and general-purpose processing can<br />
be implemented using a single processor platform. �is will revolutionize<br />
the way that embedded systems are built, saving costs<br />
and improving system reliability and performance.<br />
Kim Hartman is VP of Sales & Marketing at TenAsys,<br />
serving the embedded market with HW<br />
analysis tool and RTOS products for 25 years.<br />
Kim has recently been a featured speaker for Intel<br />
and Microsoft on the topic of embedded virtualization.<br />
He is a Computer Engineering graduate<br />
of University of Illinois, Urbana-Champaign and<br />
degreed MBA professional of Northern Illinois University.
Q: What Do You Get when You<br />
Combine the x86 Architecture with<br />
FPGAs?<br />
A: A Common Platform that Is Open and Flexible for Applications that Need Dedicated I/Os<br />
Field-programmable-gate-array (FPGA) technology has<br />
been a useful design resource for quite some time. It continues<br />
to be a mainstay because it delivers many of the same<br />
benefits as x86 processor architectures. Among the FPGA<br />
technology’s many common advantages are multifunctionality,<br />
a healthy and broad-based ecosystem, and a proven<br />
installed base of supported applications. Giving embedded<br />
designers hardware platforms that combine x86 processor<br />
boards with FPGA-controlled I/Os expands these benefits<br />
even further. Quite simply, it allows them to design dedicated<br />
I/Os to support a wider range of application requirements.<br />
By employing next-generation x86 processors with FPGAs<br />
on a single hardware platform, engineers can opens up the<br />
chance to reduce the overall bill of materials by eliminating<br />
chipsets. Different areas of applications can then be built<br />
on the same platform without a full redesign, but rather just<br />
the the exchange of the IP cores. Further cementing this<br />
approach as an appealing, long-term design solution is Loring<br />
Wirbel of FPGA Gurus. He estimates that the compound<br />
annual growth rate (CAGR) for FPGAs will continue at a<br />
strong 8.6%, which will put the FPGA market at $7.5 billion<br />
worldwide by 2015.<br />
Advantages of Open, Dedicated Platforms<br />
The x86 processor architecture continues to evolve to<br />
enable new levels of platform openness, thereby supporting<br />
enhanced design flexibility. This is demonstrated by the vast<br />
number of x86-based developers and a staggering installed<br />
base of applications. However, the number of x86 designs<br />
may have contributed to too much of a good thing in the<br />
form of features. A plethora of pre-determined instruction<br />
sets has been spawned from those features, limiting many<br />
types of embedded applications. For this reason, embedded<br />
designers have been looking for a better way to meet the<br />
specific I/O requirements or have the ability to customize<br />
embedded solutions with proprietary I/O or acceleration.<br />
By combining the central-processing-unit (CPU) core<br />
with an FPGA, designers gain access to pure IP. They can<br />
use that IP to increase design flexibility and streamline the<br />
design process for new applications. Furthermore, only the<br />
IP cores need to be maintained in the design (compared to<br />
the daunting task of managing several different controller<br />
components). Integrating the latest x86 processors with<br />
By Christine Van De Graaf, Kontron<br />
FPGAs provides a high-performance open platform. It can<br />
utilize dedicated I/Os for proprietary interfaces and other<br />
functionality, which is configured to an application’s needs.<br />
Even more importantly, this solution ensures design longevity<br />
with interface support that’s available for as long as it’s<br />
required. The advantages gleaned from a combined CPU<br />
core and FPGA solution solve dedicated I/O requirements<br />
for a broad array of systems—especially systems with unique<br />
I/O that also must interface with a range of various other<br />
devices.<br />
Sustaining Legacy Designs, Enabling Migration<br />
FPGAs have long been valued as a design tool to support<br />
older interfaces, such as ISA, RS-232, and CAN. Such<br />
interfaces are no longer supported by chipsets or dedicated<br />
hardware-designed I/O add-on cards. Looking to the future,<br />
it’s foreseeable that PCI will become obsolete and not be supported<br />
by standard chipsets. Current processor generations<br />
provide only PCI Express support, leaving designers to find<br />
other solutions such as PCI switches. These switches have to<br />
steal resources that may be needed elsewhere for the application,<br />
thus hindering the overall design. While leading-edge<br />
technologies have their place, there’s still a 20-year installed<br />
base of PCI-based applications actively deployed today. The<br />
designers working with these systems will find that it’s overkill<br />
to migrate to next-generation PCI Express or Gigabit<br />
Ethernet for applications that only call for 32-bit/66-MHz<br />
performance. An optimal solution would be to use an FPGA<br />
versus a chipset to execute the PCI interface for the needed<br />
I/O.<br />
With FPGAs, software and IP cores can now assert a<br />
major role in embedded computing at the hardware level.<br />
Rather than using a chipset with pre-configured I/O support,<br />
designers can utilize IP to customize I/O with software<br />
in a single-board solution. This approach makes it easier<br />
to migrate from legacy designs. For system upgrades, this<br />
streamlined method facilitates the adding of devices that<br />
can interface with the application. The need to include additional<br />
hardware is taken out of the equation.<br />
Support for Proprietary Applications<br />
The reality is that many proprietary applications exist<br />
today. This is especially true in the industrial-automation<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 33<br />
TECHNOLOGY APPLICATION<br />
Multicore Platform
TECHNOLOGY APPLICATION<br />
Multicore Platform<br />
market, which has to take into consideration the number of<br />
industrial Ethernet technologies that are currently in use. It<br />
has become an overwhelming task to develop different fieldbus<br />
implementations for all of these proprietary protocols.<br />
Further adding to the proprietary industrial-automation-application<br />
dilemma is how panel PCs and HMIs are supported<br />
in these countless different installations. What if an industrial-automation<br />
designer could build on the same hardware<br />
platform as a CAN, PROFIBUS, or LonWorks terminal—or<br />
any other field-bus or industrial Ethernet that’s required in<br />
a specific application? By using a platform solution that only<br />
requires the IP cores to be exchanged, one can really see the<br />
value of a flexible approach that allows a wide variety of devices<br />
to easily speak with each other.<br />
The same is true in many medical applications. Each<br />
piece of diagnostic and patient-monitoring equipment has<br />
been designed with unique I/O that has special interface<br />
requirements. It has been a tedious endeavor to upgrade<br />
these systems and network them, due to the tremendous<br />
additional hardware, configuration, and programming that<br />
would be required. An advanced CPU core and FPGA-based<br />
single-board computer is a welcome solution that delivers<br />
the design flexibility to bring all of the pieces together.<br />
Real <strong>Solutions</strong><br />
Hardware solutions now exist based on this combined approach.<br />
Recently, the <strong>Intel®</strong> Atom processor E6x5C series<br />
paired with an integrated Altera FPGA in a single package<br />
was introduced. The energy-efficient processor core offers<br />
high performance with 3D graphics, display, memory, and<br />
a PCI Express controller. In addition, <strong>Intel®</strong> Hyper-Threading<br />
Technology delivers increased performance and system<br />
responsiveness to enable multitasking and faster web-page<br />
downloads. The integrated, hardware-assisted <strong>Intel®</strong><br />
Virtualization technology helps to consolidate multiple environments<br />
into a hardware platform (see Figure 1).<br />
Figure 1: With flexible FPGA I/O options, OEMs now have the<br />
design flexibility to develop applications that have specific or<br />
dedicated I/O requirements.<br />
34 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
Kontron, in turn, has announced its MICROSPACE®<br />
MSMST. This PCIe/104 embedded single-board computer<br />
employs this latest <strong>Intel®</strong> Atom processor and FPGA combination<br />
(see Figure 2). With the Kontron MICROSPACE®<br />
MSMST, embedded-systems designers can react quickly to<br />
changing application requirements. They can easily configure<br />
a platform due to the availability of validated IP cores<br />
for the CAN-bus, serial interfaces (SPI master/UART), and<br />
PCI-Express, I2C, and GPIO. All that’s needed to enable the<br />
interfaces is the IP core and corresponding high-speed mezzanine<br />
cards (HSMCs).<br />
Figure 2: The MICROSPACE MSMST is an industrial-temperaturerange,<br />
PCIe/104 embedded single-board computer based on the<br />
<strong>Intel®</strong> Atom processor E600C series.<br />
Making this a particularly useful solution, the PC/104<br />
standards, of which PCIe/104 is one of the newest, are welladopted,<br />
highly successful, computing bases. As a result,<br />
designers have found it easy to use this familiar off-the-shelf<br />
open platform. It is an optimal platform to marry PCI-Express<br />
with a low-power computing core and an FPGA. Plus,<br />
there’s an established x86/FPGA technology ecosystem of IP<br />
cores and HSMCs. OEMs also will benefit from a reduced<br />
bill of materials (BOM), eliminating the need for additional<br />
carrier boards. Furthermore, the Kontron MICROSPACE®<br />
MSMST is by-design industrial-temperature-grade (-40° to<br />
85°C) SBC. It’s therefore a viable choice for harsh environments,<br />
such as outdoor POS/POI systems, transportation,<br />
energy, and military applications.<br />
Christine Van De Graaf is the product manager<br />
for Kontron America’s <strong>Embedded</strong> Modules Division.<br />
Van De Graaf has a decade of experience<br />
working in the embedded-computing technology<br />
industry and holds an MBA in marketing<br />
management from California State University,<br />
East Bay, Hayward, CA.
Configuration Is Key to Success<br />
with <strong>Embedded</strong> Virtualization<br />
Delivering new functionality for graphics or networking<br />
In the “old days,” configuring an embedded system was<br />
simple. A processor had a single central-processing-unit<br />
(CPU) core executing an operating system (OS). Depending<br />
on the product’s needs, a general-purpose or real-time operating<br />
system (RTOS) would be chosen. If both were needed,<br />
the design would require two processors. Yet today’s powerful<br />
single-core and multicore processors can actually be<br />
configured in many different configurations.<br />
Figure 1: Shown are several CPU-core configuration options<br />
A multicore processor can be managed by a single symmetrical-multiprocessing<br />
(SMP) operating system, which<br />
manages all of the cores (see Figure 1). Alternatively, each<br />
core can be given to a separate OS in an asymmetrical-multiprocessing<br />
(AMP) configuration. SMP and AMP both have<br />
their challenges and advantages. For example, SMP doesn’t<br />
always scale well, depending on workload. For its part, AMP<br />
can be difficult to configure with regard to which OS gets<br />
access to which device. Operating systems assume that they<br />
have full control over the hardware devices that they detect.<br />
Often, this creates conflicts in the AMP case.<br />
A technology that facilitates the configuration of these<br />
multicore processors is embedded virtualization. <strong>Embedded</strong><br />
virtualization introduces a thin, real-time virtual-machine<br />
monitor (or hypervisor) directly on top of the hardware.<br />
This hypervisor then creates virtual boards (partitions) that<br />
contain the operating systems. As a result, system designers<br />
can utilize a wide variety of configurations—including<br />
By Chris Ault, Wind River<br />
mixes of AMP, SMP, and core virtualization—to build their<br />
next-generation embedded systems. The hypervisor manages<br />
the hardware and partitions within which OSs execute.<br />
Each partition is given access to resources (processing cores,<br />
memory, and devices), can host an operating system (guest<br />
OS), and is protected from the other partitions. Simply put,<br />
an embedded hypervisor provides technology to partition or<br />
virtualize processing cores, memory, and devices between<br />
the multiple OSs that are used to build a system.<br />
In the information-technology (IT) industry, virtualization<br />
is well-known, well-understood, and well-embraced.<br />
In embedded devices, however, virtualization is nascent. It<br />
shares some of the benefits offered to the IT industry while<br />
providing other advantages that are unique to the embedded<br />
industry.<br />
For example, virtualization in the IT industry focuses on<br />
hardware abstraction to virtualize access to all devices on<br />
the host server. In doing so, it provides maximize guest-OS<br />
consolidation and homogeneous host environments. The<br />
resulting compute platforms appear identical to all guest<br />
OSs, regardless of the physical host and its hardware. Yet<br />
virtualization in the embedded industry focuses on a different<br />
set of benefits. The OSs in an embedded product need<br />
to collaborate to deliver complete device functionality. Each<br />
OS uses its own set of hardware devices, memory, and processing<br />
cores. In addition, it needs to communicate with the<br />
other OSs in the device. At the same time, the OS usually<br />
has to operate within tight memory limitations and adhere<br />
to strict timing requirements. Sometimes, it also needs to be<br />
certified to certain safety standards.<br />
<strong>Embedded</strong> virtualization can be adopted to utilize a mixture<br />
of different OSs to build an embedded device (see Figure<br />
2). For example, it may make more sense to manage and control<br />
sensors and actuators with an RTOS. Yet graphics and<br />
the networking aspects of the product could be better suited<br />
if they were supported by a general-purpose OS that offers<br />
improved graphics support and connectivity.<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 35<br />
TECHNOLOGY APPLICATION<br />
Multicore Virtualization
TECHNOLOGY APPLICATION<br />
Multicore Virtualization<br />
Figure 2: Specific operating systems are suitable for specific<br />
purposes<br />
Challenges like the following can occur when a complete<br />
solution has one subset of hardware devices controlled by<br />
one OS and another controlled by a different OS:<br />
�� ����������������������������������������������<br />
multiple OSs<br />
�� �������������������������������������������������������<br />
multiple OSs<br />
�� �����������������������������������������������<br />
moving to a virtualized environment<br />
�� ������������������������������������������������������<br />
without making widespread software changes<br />
Figure 3: Shown is an example of an embedded hypervisor<br />
configuration<br />
With an embedded hypervisor that can be configured to<br />
present system-level definitions and hardware-device mappings<br />
to the virtual boards and guest OSs, the developer<br />
gains a mechanism to describe all device configurations in<br />
one location (see Figure 3). Such device assignments include<br />
partitioning physical-memory ranges and local physicalhardware<br />
devices, assigning interrupts, and allocating CPU<br />
cores to guest OSs. Contrast this with virtualization in the<br />
36 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
IT industry, where most hardware devices are virtualized<br />
and visible to all guest OSs for maximal virtualization. <strong>Embedded</strong><br />
virtualization puts the system developer in control.<br />
Only he or she can make sure that the system partitioning<br />
is done in such a way that the final system behaves in the<br />
desired fashion.<br />
Figure 4: This hypervisor configuration shows the specific<br />
configuration of devices<br />
The system-level configuration used by an embedded hypervisor<br />
partitions the system into multiple virtual boards<br />
(see Figure 4). Each virtual board executes a real-time or<br />
general-purpose guest OS. The virtual board is managed by<br />
the hypervisor. Based on the configuration presented to it<br />
at boot time, the hypervisor controls the cores on which the<br />
virtual board executes, the memory range, and devices that<br />
can be accessed by the guest OS. The memory, PCI attributes,<br />
and interrupts can be directly mapped into individual<br />
guests. The hypervisor isn’t involved in the datapath to or<br />
from the devices. The resulting performance is equal to native<br />
non-virtualized performance.<br />
Figure 5: Achieve innovation by migrating to multicore and<br />
extending functionality<br />
With the ability to explicitly describe the devices detected<br />
by the OSs in the virtual boards, a system developer can port<br />
legacy applications—hosted by older OSs running on singlecore<br />
CPUs—onto new multicore CPUs and hardware (see<br />
Figure 5). This can be done by hosting the legacy application<br />
in a virtual board, which is presented with the same devices
(and address ranges, interrupts, etc.) as detected on the old<br />
hardware. An embedded hypervisor that supports unmodified<br />
guest OSs facilitates this task without requiring software<br />
changes to the trusted legacy application stack. Device vendors<br />
are able to port their software assets—intact—to new<br />
hardware. They can therefore<br />
leverage multicore CPUs that<br />
offer increased performance/<br />
power ratios.<br />
In this way, migration<br />
can be leveraged to add new<br />
functionality to an embedded<br />
product while isolating the<br />
existing application from the<br />
new feature extensions. Migrating<br />
a legacy application<br />
from a single-core CPU to one<br />
of the cores on a multicore<br />
CPU provides extra compute<br />
power, which can be used for<br />
this new functionality. For<br />
example, some new features<br />
that can be implemented could<br />
be those that exploit the enhanced graphics or networking<br />
capabilities of a general-purpose OS, such as Linux or Windows.<br />
In summary, multicore CPUs and embedded virtualization<br />
offer many opportunities, such as power savings and<br />
improved compute performance, in embedded products.<br />
With these opportunities come challenges in partitioning<br />
and assigning the hardware devices of the embedded board<br />
to the various CPU cores and virtualized OSs. With an embedded<br />
hypervisor, it’s possible to maintain complete control<br />
of the embedded system’s hardware devices to retain the<br />
real-time requirements in a virtualized environment. The<br />
embedded hypervisor manages and partitions hardware de-<br />
Featured Videos<br />
“<strong>Embedded</strong> virtualization introduces<br />
a thin, real-time virtual-machine<br />
monitor (or hypervisor) directly<br />
on top of the hardware.<br />
This hypervisor then creates<br />
virtual boards (partitions) that<br />
contain the operating systems.”<br />
vices based on the configuration presented to it at boot time.<br />
By partitioning the hardware devices through configuration<br />
that’s enforced by a hypervisor, the real-time requirements<br />
of the embedded system are maintained. At the same time,<br />
the benefits of embedded virtualization and multicore CPUs<br />
can be realized.<br />
With the ability to explicitly<br />
describe the devices<br />
detected by an OS within a<br />
virtual board, legacy applications<br />
can be easily migrated<br />
from old to new hardware,<br />
thereby offering increased<br />
performance/power ratios.<br />
With increased CPU performance,<br />
new features can be<br />
delivered by the development<br />
teams. With the isolation enforced<br />
by the hypervisor and<br />
virtual boards, new functionality<br />
can be delivered using<br />
general-purpose OSs with improved<br />
libraries for graphics or<br />
networking.<br />
Chris Ault is a senior product manager with<br />
Wind River focusing on virtualization solutions.<br />
Prior to joining Wind River, Ault worked in various<br />
roles ranging from software engineering to<br />
product management at Mitel, Nortel, Ciena,<br />
AppZero, and Liquid Computing. His focus was<br />
on application and server virtualization products,<br />
technologies, and sales. Ault holds electronics, computer<br />
science, and economics degrees from Carleton University and<br />
Algonquin College. He resides in Ottawa, Canada.<br />
Max Domeika talks to Low-Power Engineering about<br />
the impact of power and how that is affecting everything<br />
from embedded to multicore software<br />
http://chipdesignmag.com/lpd/blog/2009/12/10/the-view-from-intel/<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 37<br />
TECHNOLOGY APPLICATION<br />
Multicore Virtualization
17143_W_DMD11<br />
�<br />
The New Resource Exclusively for Designers<br />
and Developers of Medical Electronic Devices<br />
Co located with: Supported by:<br />
Produced and managed by: UBM Canon | ubmcanon.com<br />
CONFERENCE & EXPOSITION<br />
February 8-10, 2011<br />
Anaheim Convention Center<br />
Anaheim, CA<br />
Meet hundreds of the nation’s leading suppliers presenting<br />
the latest in emerging technologies that can be incorporated<br />
into the next generation of medical electronics, including:<br />
� Analog ICs<br />
� Communications & Networking Devices<br />
� Components<br />
� Development Tools<br />
� DSPs<br />
� Graphics/Graphics-Interface ICs<br />
� Memory & Storage<br />
� Microcontrollers & Microprocessors<br />
� Operating Systems<br />
� Power<br />
� Programmable ICs<br />
� RF<br />
� Semiconductor Processes & IP<br />
� Software<br />
� Test & Measurement<br />
� Wireless<br />
...and much more!<br />
Primary Event Sponsor<br />
<strong>Embedded</strong>/Multicore Sponsor<br />
System-on-Chip Sponsor<br />
Conference Sponsor<br />
Meet the Experts Sponsor<br />
Semiconductor Sponsor<br />
For more information on attending or exhibiting, please visit<br />
Register online today using Promo Code: AA
Product Showcase Index<br />
6WIND<br />
6WINDGate Multicore Packet Processing Software .................... 39<br />
AXIOMTEK<br />
AXIOMTEK CAPA800 Lunar Pier Refresh 3.5-inch<br />
<strong>Embedded</strong> Board .......................................................................... 40<br />
AXIOMTEK’S FIRST QueensBay Platform! PICO822 ................... 40<br />
COMMELL<br />
LV-67F---Intel ® Core i7/i5/i3 processor-based<br />
rPGA988A Mini-ITX .......................................................................41<br />
Emerson Network Power<br />
COMX-315 Compact COM Express ® Module .................................. 41<br />
CPCI7203 PICMG 2.30 3U SBC .......................................................... 42<br />
iVME7210 Dual-Core VMEbus SBC .................................................. 42<br />
NITX-300 Series Ultra Low Power Nano-ITX Motherboard ..........43<br />
ITOX<br />
ES121-LR Fanless 12VDC System Supports<br />
Triple-Independent Displays ........................................................ 43<br />
MSI Computer<br />
MSI IM-PV-C with Intel ® Atom Processors for Low Power<br />
and High Performance .................................................................. 44<br />
Pinnacle Data Systems, Inc.<br />
6WINDGate Multicore Packet Processing Software<br />
6WINDGate is the Gold Standard in packet processing<br />
software for networking equipment, wireless infrastructure,<br />
security appliances and data centers. It provides up to 10x<br />
the packet processing performance of a standard networking<br />
stack, significantly improving the price-performance and<br />
power-performance ratios of networking equipment.<br />
6WINDGate is compatible with standard Operating System<br />
APIs (e.g. Netfilter, Netlink etc). This ensures that clients<br />
can migrate either from a single-core to a multicore<br />
platform, or from one multicore platform to another, without<br />
needing to rewrite their existing software. Clients minimize<br />
the development time for their base multicore software<br />
platform, focusing on their unique product differentiation<br />
and accelerating their time-to-market.<br />
With a full set of Layer 2 through Layer 4 protocols for routing,<br />
switching, security and mobility, optimized for multicore<br />
systems, 6WINDGate is a drop-in replacement for standard<br />
networking stacks. The majority of packets are processed in a<br />
fast path environment, executing outside the operating system<br />
for optimum performance. Available protocols include:<br />
�� ��������������������������������������������������<br />
�� �������������������������������������������������������������<br />
���������������������<br />
�� ������������������������������<br />
�� �������������������������������<br />
�� ��������������������������<br />
�� ������������������������������������������<br />
ComputeNode CompactPCI Chassis Products ........................ 44<br />
CP86-N1 Intel ® Core 2 Duo Processor-Based Blade ................ 45<br />
PMC-SD18 and XMC-SD18 SATA HDD/SDD Storage Modules ... 45<br />
TS1200 Telco Server Platform .................................................... 46<br />
TS2200 Telco Server Platform ................................................... 46<br />
XMC-GBX Quad Gigabit Ethernet Adaptor ................................ 47<br />
Extreme Engineering <strong>Solutions</strong>, Inc.<br />
XPedite7470 Features Quad-Core 2nd Generation<br />
Intel Core i7 processor ............................................................... 47<br />
6WINDGate supports multicore processors from Cavium,<br />
Freescale, Intel and NetLogic.<br />
6WIND<br />
Immeuble Central Gare<br />
1 place Charles de Gaulle<br />
Montigny-le-Bretonneux, 78180<br />
France<br />
+1 (650) 968-8768 Telephone<br />
6wind-sales@6wind.com<br />
http://www.6wind.com<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 39<br />
Product Showcase
Product Showcase<br />
AXIOMTEK CAPA800 Lunar Pier Refresh 3.5-inch <strong>Embedded</strong> Board<br />
Axiomtek announces the CAPA800, a 3.5-inch embedded<br />
board either supporting a low power single-core <strong>Intel®</strong><br />
Atom processor N450 a D410 1.66GHz or a dual-core<br />
<strong>Intel®</strong> Atom processor D510 1.66GHz. The CAPA800<br />
is equipped with the <strong>Intel®</strong> I/O Controller Hub 8M to<br />
deliver enhanced system performance, I/O functions,<br />
and also supports a full DDR2 2GB memory. Along with<br />
an Integrated <strong>Intel®</strong> Graphics Media Accelerator 3150<br />
that supports advanced 3D graphics and 18-bit LVDS<br />
and VGA. The board has an expansion capability that<br />
includes a socket for a PCI Express Mini card. The system<br />
simply requires a +12V DC power input. Axiomtek’s smallfootprint<br />
CAPA800 embedded board is ideal for in-vehicle<br />
infotainment, industrial control, automation, gaming,<br />
medical devices, self-serve terminals, digital signage,<br />
and many other small and fanless devices. For additional<br />
information or a product quote contact us!<br />
Features<br />
�� ���������������������������������������������������<br />
�� ����������������������������������������<br />
�� ���������������������������������������<br />
�� ����������������<br />
�� �����������<br />
�� ���������������<br />
AXIOMTEK’S FIRST QueensBay Platform! PICO822<br />
Axiomtek announces its FIRST QueensBay Platform; the<br />
PICO822 offers excellent performance at a remarkably<br />
low power consumption rate, making it ideal for spacelimited<br />
and portable applications. The PICO822 aims at<br />
providing flexibility for expansion such as MEMS, GPS,<br />
wide-range DC-in, SDVO, PCle. We omitted soldering the<br />
I/O connectors onto the board edge so it would allow<br />
for a flexible design that will fit in with any developers’<br />
current or future projects. Our extremely small form<br />
factor embedded platform offers an excellent solution<br />
for embedded markets such as medical applications, invehicle<br />
PCs, portable POS, Car PC, entry-level gaming,<br />
and in-flight infotainment.<br />
�� ���������������������������������������������������<br />
with <strong>Intel®</strong> Platform Controller Hub EG20T<br />
�� ������������������������������������������������<br />
�� ����������������������������������������������������<br />
Gigabit Ethernet, SMbus, & 8 Channel Digital I/O<br />
�� �������������������������������������������������<br />
graphics engine<br />
�� ����������������������������������������<br />
�� �����������������������������<br />
�� ���������������������������������������������<br />
temperature support<br />
40 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
AXIOMTEK<br />
18138 Rowland St<br />
City of Industry, CA 91748<br />
USA<br />
1.800.GO.AXIOM Toll Free<br />
1.626.581.3232 Telephone<br />
sales@axiomtek.com<br />
www.axiomtek.com<br />
AXIOMTEK<br />
18138 Rowland St<br />
City of Industry, CA 91748<br />
USA<br />
1.800.GO.AXIOM Toll Free<br />
1.626.581.3232 Telephone<br />
sales@axiomtek.com<br />
www.axiomtek.com
LV-67F--- Intel ® Core i7/i5/i3 processor-based rPGA988A Mini-ITX<br />
������� �������� ��������� ��������������� ���� ����������<br />
leader of Industrial Mini-ITX mainboard, is releasing the Mini-<br />
�������������������������������������������������������� ®<br />
�����������������������������������������������������������<br />
ITX mainboard based on the Mobile Intel ® ����������������<br />
along with a compatible processor that is one of Intel’s next<br />
generation 64-bit, multi-core processors built on 32 nm Intel<br />
������������������������������������������������������������<br />
������������������������������������������������������������<br />
and graphics core on the processor die provides a two-chip<br />
system architecture. Innovative two-chip solution provides<br />
Intel ® Turbo boost technology and Intel ® � ����������������<br />
����������� ������ ���������� ������������ ��� ������ �����<br />
workload and lower power than previous three-chip platform.<br />
This platform is ideal for developing high-performance<br />
systems for industrial control and automation, retail, gaming,<br />
print imaging and digital signage.<br />
�����������������������������������������������������������<br />
���� ��� ����� ��������� ���� ����� ��������� ����� �������������<br />
��������������������������������������������������������<br />
�������������������������������������������������������������<br />
������������������������������������������� ® ���������������������<br />
����� �������� ������ ������� ������������ ����������<br />
�������� ��������� �������������� ��������� ������ ��������� ���<br />
������������������������������������������������������������������<br />
������������������������������������������������������������<br />
����������������������������������������������������������<br />
COMX-315 Compact COM Express ® Module<br />
The Emerson Network Power is a Compact COM Express ®<br />
module based on the Intel ® Atom processor E640 for<br />
deeply embedded, low power applications. At 95 mm x 95<br />
mm, compact format COM Express modules are designed to<br />
integrate all the core components of a common PC in a form<br />
factor small enough for specific embedded applications on a<br />
carrier board. COMX-315 includes graphics, sound, storage,<br />
network and multiple USB ports. The COMX-315 will be Type<br />
2 compatible to fit on existing Type 2 carriers, but will not<br />
include the legacy PCIbus and IDE parallel interfaces.<br />
Based on the Intel Atom processor E640, the COMX-315<br />
module is a highly reliable and long lifecycle product with<br />
ultra low power consumption, making it ideal for portable and<br />
hand held applications that do not require forced-air cooling.<br />
Typical applications for COMX-315 include portable test<br />
equipment, handheld devices and mobile based applications.<br />
Product Features<br />
�� �������������������������������������<br />
�� ����������������������������������<br />
�� ������������� ® Atom processor E640<br />
�� ����� ® �������������������������������<br />
�� �����������������������������<br />
�� �����������������<br />
�� ����������������������������������������������<br />
��������������������������������������������������������������<br />
Intel ® ��������������������������������������������������������� ®<br />
����������������������������������������������������������<br />
quality audio. Also support for eSATA enables the full SATA<br />
���������� ������ ��� ��� ������� �������� ���� ����� ������� ���� ���<br />
enables high reliablity for personal data, or maximum storage<br />
performance for intensive applications. Mini PCI, Mini PCIe<br />
and PCIE slots easily using any other add-on cards.<br />
����������������������������������������������������� http://<br />
www.commell.com.tw<br />
Taiwan Commate Computer Inc.<br />
886-2-26963909 Phone<br />
886-2-26963911 Fax<br />
info@commell.com.tw<br />
sales@tcommate.com.tw<br />
www.commell.com.tw<br />
Emerson Network Power<br />
2900 S. Diablo Way, Suite 190<br />
Tempe, AZ 85282<br />
USA<br />
1 800 759 1107 Toll Free<br />
1 602 438 5720 Telephone<br />
<strong>Embedded</strong>ComputingSales@Emerson.com<br />
Emerson.com/<strong>Embedded</strong>Computing<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 41<br />
Product Showcase
Product Showcase<br />
CPCI7203 PICMG 2.30 3U SBC<br />
The CPCI7203 3U SBC features the integrated dual-core Intel ®<br />
Core i7 processor for use in high performance, spaceconstrained<br />
applications. On-board memory includes up to<br />
8GB DDR3 and 256KB non-volatile F-RAM. F-RAM does not<br />
require batteries or periodic refreshes and offers many more<br />
read/ write cycles and faster performance than flash memory,<br />
which benefits critical non-volatile data storage, data logs<br />
and dynamic program updates. The Trusted Platform Module<br />
(TPM) enhances data security and encryption capabilities.<br />
The CPCI7203 is a low-power, high-performance SBC that<br />
offers full hot swap compliance per PICMG ® 2.1 and supports<br />
the PICMG 2.9 System Management specification and PICMG<br />
2.30 CompactPCI PlusIO specification.<br />
The CPCI7203 supports a range of operating system and<br />
software options. It is ideal for a wide range of industrial,<br />
medical and military/aerospace applications, such as<br />
railway control, semiconductor processing, robotics, image<br />
processing, vehicle communications and on-board flight<br />
information systems.<br />
Product Features<br />
�� �������������������������� ® Core i7 processor (up to 2 GHz)<br />
�� ������������������������������������������������<br />
�� ������������������������<br />
�� ������������ ® 5 Series chipset<br />
�� �����������������������������������������������������������<br />
�� �����������<br />
iVME7210 Dual-Core VMEbus SBC<br />
������������������������ ® Core i7 processor variants and the<br />
Mobile Intel ® �����������������������������������������������<br />
of industrial, medical and military/aerospace applications<br />
�������������������������������������������������������������<br />
signal intelligence. The dual-core processor has integrated<br />
memory and graphics controller. On-board memory includes<br />
up to 8GB DDR3 soldered memory and 256KB non-volatile<br />
Ferroelectric Random Access Memory (F-RAM). F-RAM does<br />
not require batteries or periodic refreshes and offers much<br />
greater read/write cycles and faster performance than flash.<br />
��������������������������������������������������������������<br />
������������������������������������������������������<br />
drive accessory option. Connectivity includes four Gigabit<br />
���������������������������������������������������������������<br />
�������������������������������������������������������������<br />
������������������������������������������������������������<br />
and two SATA ports. Compatible operating systems include<br />
��������������������������������������������������������������<br />
��������������������������������������������������������������<br />
alliance partners.<br />
Product Features<br />
�� ��������������� ® �����������������������������������������<br />
�������������������������������������<br />
�� ����� ® ������������������������������������������<br />
�� ���������������������������������������������<br />
�� ��������������������<br />
�� �����������������������<br />
�� ��������������������������<br />
42 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
Emerson Network Power<br />
2900 S. Diablo Way, Suite 190<br />
Tempe, AZ 85282<br />
USA<br />
1 800 759 1107 Toll Free<br />
1 602 438 5720 Telephone<br />
<strong>Embedded</strong>ComputingSales@Emerson.com<br />
Emerson.com/<strong>Embedded</strong>Computing<br />
Emerson Network Power<br />
2900 S. Diablo Way, Suite 190<br />
Tempe, AZ 85282<br />
USA<br />
1 800 759 1107 Toll Free<br />
1 602 438 5720 Telephone<br />
<strong>Embedded</strong>ComputingSales@Emerson.com<br />
Emerson.com/<strong>Embedded</strong>Computing
NITX-300 Series Ultra Low Power Nano-ITX Motherboard<br />
The NITX-300 series of Nano-ITX motherboards from<br />
Emerson Network Power feature the Intel ® Atom processor<br />
E6xx series. These ultra low power motherboard solutions<br />
offer passive cooling capability for reliable operation. They are<br />
designed for use in a variety of applications such as embedded<br />
instruments, medical carts, audio visual display systems,<br />
and other applications that require an easy-to-use Nano-ITX<br />
motherboard with support for a variety of operating systems.<br />
With a size format of 120 mm x 120 mm, Nano-ITX form factor<br />
motherboards are very suitable for low power embedded<br />
applications. The NITX-300 series has a low height profile<br />
to fit into most enclosures and has a wide range of built-in<br />
connectivity including LCD and/or CRT displays; SATA for<br />
physical or solid state disks; a PCI Express x1 expansion slot<br />
and a PCI Express Mini Card slot for Wi-Fi/WiMAX; USB and<br />
Gigabit Ethernet networks; audio; and multiple serial ports.<br />
Product Features<br />
�� ��������������������������������������������������<br />
�� ������������������������������������������<br />
�� �����������������������������������������������<br />
�� �����������������������������������������<br />
�� �������������������������������<br />
�� ����������������������������������������������<br />
�� ���������������������<br />
�� ������������������<br />
Emerson Network Power<br />
2900 S. Diablo Way, Suite 190<br />
Tempe, AZ 85282<br />
USA<br />
1 800 759 1107 Toll Free<br />
1 602 438 5720 Telephone<br />
<strong>Embedded</strong>ComputingSales@Emerson.com<br />
Emerson.com/<strong>Embedded</strong>Computing<br />
ES121-LR Fanless 12VDC System Supports Triple-Independent Displays<br />
This new embedded system offers higher-performance per watt<br />
than any previous low power platform. The ES121-LR utilizes a<br />
dual-core Intel ® Atom processor D510 and Intel ® I/O Controller<br />
���������������������������������������������������� ® 435<br />
���� ������������ ����� ����������������� ������ ������ ��������<br />
support. The low-profile chassis is constructed of heavy duty<br />
steel and an extruded aluminum top cover providing noise-free,<br />
passive system cooling.<br />
The S3 Graphics Chrome ® 435 ULP accelerator provides two<br />
DVI-D and two VGA display interfaces using DVI-I connectors,<br />
���� ���� ������� ����� �������� ����������� �������������������<br />
display imaging is accomplished using the processor’s VGA<br />
display interface and any combination of the S3 Graphics<br />
��������������������������������������<br />
Key System Features<br />
�� ��������������� ® Atom processor D510<br />
�� ���������������������������������<br />
�� ������������������������<br />
�� ��������������������������������������<br />
�� ����������������������������������<br />
�� ��������������������������������������<br />
�� ���������������������������������<br />
�� ��������������������������������������������<br />
�� ������������������������������������������<br />
�� ����������������������������������������<br />
�� �������������������������������������<br />
�� ��������������������������������<br />
This low-power 12VDC embedded system is ideal for<br />
applications requiring a stable revision-controlled platform,<br />
such as digital signage, gaming, industrial control automation,<br />
POS and Kiosk, and medical equipment applications.<br />
ITOX Applied Computing<br />
8 Elkins Road<br />
East Brunswick, NJ 08816<br />
732-390-2815 Phone<br />
888-200-4869 Toll Free<br />
732-390-2817 Fax<br />
sales@itox.com<br />
www.itox.com<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 43<br />
Product Showcase
Product Showcase<br />
MSI IM-PV-C with <strong>Intel®</strong> Atom Processors for Low Power and<br />
High Performance<br />
MSI announces the latest mini-ITX motherboard, IM-PV-C,<br />
to IPC market. IM-PV-C is based on the latest <strong>Intel®</strong> Atom<br />
processors N455/ N475/ D425/ D525 and the <strong>Intel®</strong> I/O<br />
Controller Hub 8M which is an ideal low power and high<br />
performance solution for industrial applications.<br />
MSI IM-PV-C comes to meet the requirements of a low power<br />
platform, especially in power input design. MSI IM-PV-C<br />
supports ATX and wide range DC 12V ~ 24V inputs as the<br />
different BOM option. For supporting high resolution video,<br />
MSI debuts IM-PV-C with great 3D graphics performance for<br />
a high definition up to 1080P videos.<br />
With two DDR3 800 MHz So-DIMM sockets, IM-PV-N with<br />
Intel Atom processors D425/ D525 can support up to 4GB<br />
system memory and IM-PV-N with Intel Atom processors<br />
N455 / N475 can support up to 2GB system memory. For<br />
the storage application, it supports 3 SATAII ports. Onboard<br />
I/O advantages include 5 COM ports, 8 USB 2.0 ports,<br />
and multiple video output (DVI, VGA and LVDS) for dual<br />
independent display. Networking is provided by one <strong>Intel®</strong><br />
82574L GbE LAN and one <strong>Intel®</strong> 82567V GbE LAN. Moreover,<br />
it features one compact flash slot, one PCI and one mini-PCIe<br />
slot for easy expansion. It also supports onboard watchdog<br />
timer for added system redundancy and security.<br />
ComputeNode CompactPCI Chassis Products<br />
PDSi’s ComputeNode line offers a range of NEBS Level 3-compliant<br />
CompactPCI chassis in sizes from 1U to 4U. These carrier-grade<br />
chassis include a horizontal design, superior air cooling, cPCI and<br />
cPSB (PICMG 2.16) backplanes, redundant hot-swappable fans,<br />
hot-swappable front-accessible AC or DC power supplies, and<br />
rear single or dual power feeds. All 2U and larger ComputeNode<br />
platforms include PDSi’s unique Alert!Node (or Enhanced<br />
Alert!Node) alarm card, an intelligent chassis management<br />
controller for comprehensive fan and power monitoring. The<br />
Alert!Node card does not occupy a CompactPCI slot, front or rear.<br />
OEMs and Independent Software Vendors (ISVs) can also take<br />
advantage of PDSi’s design, integration, and support services,<br />
including custom board and system design, validation<br />
and certification, production assembly and test, as well as<br />
extended service programs.<br />
Product Features<br />
�� ������������������������������������<br />
�� �������������������������������������������<br />
�� �����������������������<br />
�� ���������������������������������<br />
�� ��������������������������������<br />
�� ���������������<br />
�� ������������������������������<br />
�� ��������������������<br />
�� �������������������������������������<br />
�� ��������������������������������������<br />
�� ��������������������������������������������������<br />
extended availability assured<br />
44 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
MSI Computer<br />
901 Canada Court<br />
City of Industry, CA 91748<br />
626 913 0828 ext. 193<br />
626 913 0818 Fax<br />
JonasC@msicomputer.com<br />
www.msicomputer.com<br />
Pinnacle Data Systems, Inc.<br />
6600 Port Road<br />
Groveport, OH 43125<br />
USA<br />
1 614 748 1150 Telephone<br />
info.sales@pinnacle.com<br />
www.pinnacle.com
CP86-N1 Intel ® Core 2 Duo Processor-Based Blade<br />
PDSi’s Intel-based CompactPCI x86 Processor Blade (CP86-N1)<br />
provides a robust, high-performance general purpose compute<br />
platform for use in CompactPCI PICMG 2.16 systems. This latest<br />
addition to PDSi’s ComputeNode family of carrier-grade<br />
CompactPCI solutions is built around the 45nm technology Intel ®<br />
Core2 Duo processor and server-grade Intel ® 5100 Memory<br />
Controller Hub (MCH) chipset and I/O Controller Hub (ICH) 9R<br />
supporting ECC memory. This powerful, compact blade offers the<br />
highest performance and dependability in its class.<br />
The ComputeNode CP86-N1 blade includes a standard PMC/<br />
XMC site for I/O expansion and features an onboard SATA drive<br />
plus high resolution graphics. I/O capability covers a very broad<br />
range of interfaces that can be accessed through one of PDSi’s<br />
companion rear transition modules such as CP86-RT01. Two<br />
1000Base-T Ethernet ports provide the PICMG 2.16-compliant<br />
fabric interfaces, making the CP86-N1 fully compatible with any<br />
ComputeNode cPSB chassis.<br />
Product Features<br />
�� �������������������������������������<br />
�� �����������������������������������������������������������<br />
MCH chipset and / ICH9R<br />
�� �������������������������������������������<br />
�� ���������������������������������<br />
�� ������������������������������������<br />
�� �������������������������������<br />
�� �������������������<br />
�� ��������������������������������������������������<br />
extended availability assured.<br />
PMC-SD18 and XMC-SD18 SATA HDD/SDD Storage Modules<br />
These new SATA Storage Modules are offered in both PMC<br />
and XMC formats. Both provide high capacity SATA storage<br />
using compact 1.8 inch hard disk (HDD) or solid state drives<br />
(SSD) - up to 160GB of storage is available with either drive<br />
type. Whether configured with an economical rotating HDD or<br />
with a highly shock-resistant SSD, these low profile modules<br />
�����������������������������������������������������������<br />
AdvancedTCA®, and PCI Express processor boards without<br />
risk of mechanical interference.<br />
���� �������� ������� ����� ����������� ��������� �� �����������<br />
external drive interfaces. OS support includes Windows,<br />
Linux, Solaris x86, and Solaris SPARC. Critical military and<br />
aerospace applications will appreciate the high operating<br />
shock resistance (1000+ G) and high MTBF (over 1 million<br />
hours) when configuring these modules with the latest SSD<br />
technology.<br />
Product Features<br />
�� �����������������������������������������������<br />
�� ��������������������������������������<br />
�� ��������������������������������<br />
�� �����������������������������������������������������<br />
�� ��������������������������<br />
�� ����������������������������������<br />
�� ��������������������������������������������������<br />
extended availability assured<br />
Pinnacle Data Systems, Inc.<br />
6600 Port Road<br />
Groveport, OH 43125<br />
USA<br />
1 614 748 1150 Telephone<br />
info.sales@pinnacle.com<br />
www.pinnacle.com<br />
Pinnacle Data Systems, Inc.<br />
6600 Port Road<br />
Groveport, OH 43125<br />
USA<br />
1 614 748 1150 Telephone<br />
info.sales@pinnacle.com<br />
www.pinnacle.com<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 45<br />
Product Showcase
Product Showcase<br />
TS1200 Telco Server Platform<br />
The TS1200 is a cost-effective NEBS® Level 3 compliant rackmount<br />
or stand-alone platform. It comes with either two Four-Core<br />
or two Six-Core <strong>Intel®</strong> Xeon® processors 56XX series, up to 96<br />
GB of DDR3 memory, and features IPMI System Management.<br />
The compact TS1200 enclosure, 1U high by 21.25” Telco depth,<br />
features redundant rear DC or AC power supplies and two (2) front<br />
hot-swap 2.5” SAS, SATA or SSD disks.<br />
The TS1200 uses cost effective CPUs without active heat<br />
sinks, and provides superior carrier grade cooling and<br />
serviceability. Its modular high capacity redundant hot-swap<br />
fans assure quiet operation and can be quickly serviced. Fans<br />
are directly accessible from the front – no need to open a top<br />
cover. Designed for NEBS compliancy, PDSi platforms can be<br />
custom-configured and managed to provide long-life support.<br />
PDSi can also manage full NEBS certification with customer<br />
specific PCI cards installed.<br />
Product Features<br />
�� ����������������������������������������������������������<br />
�� �������������������������������������<br />
�� �����������������������������������������<br />
�� ���������������������������������������������������<br />
necessary<br />
�� ���������������������������������������������<br />
�� ��������������������������������������������������<br />
optional<br />
�� ����������������������������������������������������<br />
TS2200 Telco Server Platform<br />
The TS2200 is a cost-effective NEBS® Level 3 compliant<br />
rack-mount or stand-alone platform. It comes with either two<br />
Four-Core or two Six-Core <strong>Intel®</strong> Xeon® processors 56XX<br />
Series, up to 96GB of DDR3 memory, and features IPMI System<br />
Management. The compact TS2200 enclosure, 2U high by<br />
22.65” Telco depth, features redundant rear DC or AC power<br />
����������������������������������������������������������<br />
The TS2200 uses cost effective CPUs without active heat sinks,<br />
and provides superior carrier grade cooling and serviceability.<br />
Its modular high capacity redundant hot-swap fans assure<br />
quiet operation and can be quickly serviced. Fans are directly<br />
accessible from the front and sides - no need to open a top<br />
cover. Designed for NEBS compliancy, PDSi platforms can be<br />
custom-configured and managed to provide long-life support.<br />
PDSi can also manage full NEBS certification with customer<br />
specific PCI cards installed.<br />
Product Features<br />
�� ����������������������������������������������������������<br />
�� �������������������������������������<br />
�� �����������������������������������������<br />
�� ���������������������������������������������������<br />
necessary<br />
�� �����������������������������������������������<br />
�� ��������������������������������������������������<br />
optional<br />
�� ����������������������������������������������������<br />
46 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
Pinnacle Data Systems, Inc.<br />
6600 Port Road<br />
Groveport, OH 43125<br />
USA<br />
1 614 748 1150 Telephone<br />
info.sales@pinnacle.com<br />
www.pinnacle.com<br />
Pinnacle Data Systems, Inc.<br />
6600 Port Road<br />
Groveport, OH 43125<br />
USA<br />
1 614 748 1150 Telephone<br />
info.sales@pinnacle.com<br />
www.pinnacle.com
XMC-GBX Quad Gigabit Ethernet Adaptor<br />
This new quad gigabit Ethernet XMC is a high-performance,<br />
low-latency network adaptor providing four high-speed Ethernet<br />
interfaces for use with VITA 42.3-compatible VME, PCI Express,<br />
CompactPCI®, and AdvancedTCA® processor boards. It is<br />
available in three configurations offering a mix of front and rear<br />
port access.<br />
Wide internal data paths eliminate performance bottlenecks.<br />
The parallel and pipelined logic architecture is optimized<br />
for Gigabit Ethernet and efficiently handles packets with<br />
minimum latency. Using widely accepted <strong>Intel®</strong> 82571EB<br />
Ethernet controllers, this adaptor offers up to four 10BASE-<br />
T/100BASE-Tx/1000BASE-T copper ports with front-mounted<br />
RJ-45 connectors and full status indicators. Alternatively, up to<br />
four SERDES ports are accessible through the Pn4 connector<br />
for use via an appropriate copper or fiber-based rear transition<br />
module.<br />
Product Features<br />
�� ���������������������������������������������������<br />
�� ��������������������������������������������������<br />
RJ-45 front connectors with status indicators<br />
�� ��������������������������������������������<br />
�� �������������������������<br />
�� ������������������������������<br />
�� ���������������������������<br />
�� �����������������������������������������<br />
Pinnacle Data Systems, Inc.<br />
6600 Port Road<br />
Groveport, OH 43125<br />
USA<br />
1 614 748 1150 Telephone<br />
info.sales@pinnacle.com<br />
www.pinnacle.com<br />
XPedite7470 Features Quad-Core 2nd Generation Intel Core i7 processor<br />
The XPedite7470 is a high-performance, low-power 3U VPX-<br />
REDI single-board computer based on the 2nd generation<br />
������� �������� ���������� ���� ������ ����� ��������� ����� ����<br />
PCI Express x4 VPX P1 interconnects and two Gigabit Ethernet<br />
ports, the XPedite7470 is ideal for the high-bandwidth and<br />
processing intensive applications of today’s military and<br />
avionics applications.<br />
The XPedite7470 accommodates up to 8 GB of DDR3 ECC SDRAM<br />
on two channels to support memory-intensive applications.<br />
The XPedite7470 also hosts numerous I/O interfaces including<br />
Gigabit Ethernet, USB 2.0, SATA, graphics, and RS-232/RS-422/<br />
RS-485 through the backplane connectors.<br />
The XPedite7470 can be used in either the system slot or<br />
peripheral slots of a VPX backplane. Wind River VxWorks,<br />
������ ������ ����������� ���� ������ ������ �������� ���������<br />
(BSPs) are available, as well as Windows drivers.<br />
Product Features<br />
�� �����������������������������������������������������<br />
Vector Extensions (AVX) delivers highest performance for<br />
demanding DSP applications<br />
�� �������������������������������������������<br />
�� ���������������������������������������������������������<br />
bandwidth system interface<br />
�� �������������������������������������������������������<br />
of application requirements<br />
�� �����������������������������������������������������<br />
storage requirements<br />
�� �������������������������������������������������������<br />
application I/O requirements<br />
�� �������������������������������������������������������<br />
support software development<br />
Extreme Engineering <strong>Solutions</strong>, Inc.<br />
3225 Deming Way, Suite 120<br />
Middleton, WI 53562<br />
608-833-1155 Telephone<br />
608-827-6171 Fax<br />
sales@xes-inc.com<br />
www.xes-inc.com<br />
www.embeddedintel.com | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | 47<br />
Product Showcase
LAST WORD<br />
PCI Express Versus RapidIO: The<br />
Winner Is…<br />
The debate known as the “fabric wars” started around the<br />
turn of the century and has been raging ever since. Over<br />
the past decade, there have been dozens of entries vying to<br />
be the system-interconnect fabric for embedded systems.<br />
Who can forget Advanced Switching Interconnect (ASI),<br />
StarFabric, Hyper Transport, InfiniBand, and Parallel RapidIO?<br />
Now, the list of contenders is down to just three: 10<br />
Gigabit Ethernet (10 GbE), RapidIO, and PCI Express (PCIe).<br />
Ethernet isn’t going away. RMDA support is part of the<br />
10 GbE specification and cut-through switches are becoming<br />
available. (Here, the switch starts forwarding a packet before<br />
the whole packet has been received, as opposed to store-andforward<br />
switches.) Clearly, 10 GbE will serve the needs of<br />
many applications. It will take some time, however, before 10<br />
GbE is ready for the heavy lifting of the data plane for highend<br />
applications.<br />
Until then, RapidIO and PCIe are the most viable highperformance,<br />
low-latency, low-overhead, embedded-systems<br />
interconnects. RapidIO has been seen by many to be technically<br />
superior—primarily because of its peer-to-peer support.<br />
PCIe was designed as a serial replacement for the parallel<br />
PCI bus. At the software level, PCIe preserves compatibility<br />
with PCI. To use PCIe in embedded applications with multiple,<br />
independent CPU subsystems, non-transparent bridging<br />
is required (just as CompactPCI required non-transparent<br />
PCI bridging).<br />
The best technology doesn’t always win, however. One of<br />
the most famous examples of this is Betamax versus VHS.<br />
(For you younger readers, these are the video-tape formats<br />
that predate DVDs.) Videophiles recognized that the quality<br />
of Betamax video was superior. But VHS won out due to a<br />
longer record time, which the market wanted.<br />
The embedded market is leveraging the PCIe infrastructure<br />
from the PC world, which is analogous to how it<br />
leveraged the PCI bus infrastructure. All new 32- and 64-bit<br />
processors support PCIe. In addition, I/O controllers and devices<br />
that previously had PCI bus interfaces have migrated to<br />
PCIe endpoint interfaces. PCIe switch chips from a number<br />
of companies are on the market. Even field-programmable<br />
gate arrays (FPGAs) include hard-coded PCIe endpoint cores.<br />
The picture for RapidIO is much different. The majority<br />
of companies that implement RapidIO in silicon are processor<br />
or DSP companies, for which RapidIO is a very good<br />
interconnect. There are far fewer switch options available.<br />
48 | <strong>Embedded</strong> Intel ® <strong>Solutions</strong> — Winter 2011 | www.embeddedintel.com<br />
By Dave Barker, Extreme Engineering <strong>Solutions</strong> (X-ES)<br />
Overall, the RapidIO ecosystem is much smaller than the<br />
PCIe ecosystem.<br />
In this competition, success is all about market share.<br />
Because of the PC market, PCIe has market share while<br />
RapidIO does not. By itself, the embedded-computing market<br />
isn’t big enough to make a market for an interconnect<br />
technology that lives within a chip. Several years ago, this<br />
point was made clear to me at a RapidIO Trade Association<br />
meeting. We were discussing the future of RapidIO prior to<br />
the release of the RapidIO 2.0 specification. Several CPU<br />
vendors present said the only way that they could support<br />
RapidIO in their chips was if it continued to stay in lockstep<br />
with PCIe and leverage the same lower layers of the protocol<br />
(e.g., 8b/10b encoding) and SERDES. Otherwise, there wasn’t<br />
enough market demand to justify implementing RapidIO at<br />
the expense of another feature. Even though RapidIO was<br />
technically sound, I realized that it couldn’t win in the market<br />
over PCIe.<br />
Since then, a number of advancements have made PCIe<br />
very relevant for embedded applications. PCIe switches are<br />
becoming available, which allow multiple non-transparent<br />
ports to be configured. PCIe 2.0 was ratified in 2007,<br />
doubling the data rate to 5.0 Gb/s. With proper software<br />
support, processors in embedded systems can communicate<br />
over PCIe as peers in a network. In late 2010, the PCI SIG<br />
made the PCIe 3.0 specification available to members.<br />
It’s time to declare the fabric war over. And the winner is<br />
PCIe—for now. Depending on the longevity of PCIe and the<br />
viability of 10 GbE, the landscape may look different in a few<br />
years. But the next big debate may be over optical interconnects.<br />
Ready for the “optical war?”<br />
Dave Barker is the director of marketing at<br />
Extreme Engineering <strong>Solutions</strong> (X-ES). Previously,<br />
he headed marketing and business<br />
development at VMETRO and was the VME<br />
product manager at the Motorola Computer<br />
Group. Barker has a BS in computer science<br />
from the University of Pittsburgh and an MBA from the University<br />
of Phoenix.
GET THE EM BEDDED INTEL ® SOLUTION YOU NEED.<br />
GET THE PARTNER YOU W ANT.<br />
THAT’S THE<br />
CRITICAL DIFFERENCE.<br />
With the <strong>Embedded</strong> Computing business of Emerson Network Power,<br />
you get leading-edge integrated platforms, chassis, blades, modules<br />
and software building blocks. And with Emerson as your partner, you<br />
get to market first with reliable products that last. You get innovative,<br />
proven technology all backed by global service from a trusted partner.<br />
See the difference for yourself at<br />
EmersonNet w orkPow er.com/ <strong>Embedded</strong>Comput ing<br />
EmersonNet w orkPow er.com/ <strong>Embedded</strong>Comput ing<br />
1.800.759.1107 or +1.602.438.5720<br />
Em erson is a trademark of Em erson Elect ric Co. Intel is a trademark of Intel Corporation or its<br />
subsidiaries in t he US or other countries. ©2010 Emerson Electric Co. All rights reserved.