Mentor Graphics Acquires Embedded Alley, Unveils Embedded Open Source Strategy

Continuing the trend of consolidation started by Intel’s acquisition of Wind River, Mentor Graphics announced today that the EDA vendor was acquiring Embedded Alley, a supplier of embedded Linux and Android solutions and services.  In bringing Embedded Alley into the Mentor Graphics family, the company (finally) makes its entry into the open source embedded marketplace, complementing its deeply-embedded legacy proprietary RTOSes, Nucleus and VRTX, with the Embedded Alley Development System for Linux and for Android, along with complementary services.

The announcement comes on the heels of Embedded Alley’s very successful launch of tools and services for deploying the Google Android platform in non-mobile applications, specifically for MIPS and Power Architecture CPUs in designs ranging from portable multimedia to home entertainment to automotive to instrumentation to industrial control.  These applications align closely with Mentor Graphics embedded target markets, and Embedded Alley competence in Linux and Android clearly  motivated the acquisition.

Mentor Graphics and Embedded Software – A Brief History

Mentor Graphics is unique among EDA suppliers in making consistent, long-term investments in embedded software to complement the company’s more traditional design tools and services offerings.   With the acquisition of  Embedded Alley,  a relatively young company, but one with already strong traction in embedded open source, the mature Mentor Graphics Embedded Software Division substantiates its long-held vision for full device life-cycle coverage, from silicon design through device software deployment.

The history of that business unit reflects Mentor’s investments towards that same end:

• 1990-1994 : Mentor partnered closely with Microtec Research (my employer until 1993)
• 1994 : Microtec Research acquires Ready Systems (provider of VRTX, founded by Jim Ready of more recent MontaVista renown)
• 1995 : Mentor Graphics acquires Microtec Research and established Embedded S/W Division, with compiler (MCC), debugger (XRAY) and RTOS (VRTX) offerings
• 2002 : Mentor Graphics acquires Accelerated Technology, providers of the Nucleus RTOS and tools
• 2006 : Mentor acquired Embedded Performance Inc, provider of JTAG hardware debug tools

Key Points

The announcement carries implications for reshaping the embedded software landscape:

• The Mentor Graphics Embedded Software Division has always been a part of the a much larger software products company.  The announcement signals Mentor’s design to grow its embedded business, but as a services offering
• While Nucleus (and VRTX before it) enjoyed most of its design wins as a deeply embedded OS (e.g., in mobile, for baseband processing in over 1B handsets), the Linux and Android technology purveyed by Embedded Alley is definitively for hosting highly visible value-added applications
• Unlike the Intel acquisition of Wind River, this merger is not designed to pull silicon into new designs (e.g., Atom in 3G mobile) but rather to expand a broad base of non-mobile embedded projects using Linux and Android (although I am sure that mobile is not 100% off the table), to complement Mentor’s EDA tools and broaden the company’s ability to address OEM product life cycles
• Also unlike the Intel-Wind River deal, in which Intel’s vast open source capability and experience was joined to Wind’s lesser but still mature Linux services capability, Mentor is looking to Embedded Alley to reinvent their embedded business around Linux, Android and open source.

Wednesday’s Press Event

The acquisition was announced on Wednesday, July 29 in San Francisco at the Design Automation Conference (DAC), at an exclusive press-only event.  Presentations by Mentor Graphics and Embedded Alley executives were followed by partner talks from ARM, MIPS and those companies’ silicon licensees.  The lineup also featured a lively pitch by Jim Zemlin of the Linux Foundation.

The New Embedded Linux/Android Opportunity, Post Wind River

With Embedded Alley on board, Mentor is well positioned to exploit the highly visible gap left when Intel snapped up Wind River.  Despite Alameda’s strong messages of continuity, OS, tools and services companies have been circling around semiconductor suppliers hoping to secure relationships and design wins for ARM, MIPS and Power Architecture that Wind is unlikely to pursue as part of Intel and the Atom juggernaut.

While there is no denying that Intel’s acquisition leaves Wind River’s traditional silicon relationships open to new challengers, those relationships were never exclusive.  Mentor and the rest of industry will still have to execute on their hopes/intentions to exploit the gaps left by Wind-Intel, with competitive technologies, platforms and services, something few have been able to do to date, on both tilted and level playing fields.

Deep Background and Fun Facts

I myself have a long history with Mentor and its Embedded Software division.  I actually worked for Microtec Research from 1988 to 1993 as product manager for XRAY and for the popular 68000 tool suite.  I exited right before the merger with Ready Systems and the subsequent purchase of the company by Mentor, and moved to Brazil (1993-1996, the Lost Years).   In 2005, I actually presented on embedded Linux at the Mentor Graphics sales kickoff in San Francisco.  I was asked by then VP of Marketing Robert Day to give the team a “wake up call” and I did my best to open their eyes.   In the face of persistent denial on the part of the Nucleus OS and tools sales teams, I explained that Linux and OSS were already eating their proverbial lunch, and that OEMs deployed the open source OS not BECAUSE of its attributes but rather IN SPITE of those same attributes.

The acquisition brings a chapter of embedded software history full circle:  as cited above , Microtec acquired Ready Systems and was then bought by Mentor.   In 1999, Jim Ready exited Mentor Graphics to found MontaVista Software (I was there too!).  Almost the entirety of the Embedded Alley team, from CEO Pete Popov to CTO Dan Malek to COO Matt Locke to the firm’s two dozen other staffers worked for MontaVista before founding/joining Embedded Alley.  President Paul Staudacher worked at Mentor until he joined MontaVista in 1999.

And now, Embedded Alley is “back”, inside Mentor, where they will hopefully reinvent the company’s embedded business and probably a goodly portion of the EDA and embedded industries as well.

Filed under: Uncategorized

Google Chrome and Android – Why not AndroChrome?

Last week’s buzz over Google’s announcement of its Chrome desktop (and presumably notetop) OS led me and my readers to ask why the Silicon Valley behemoth needs two mobile OSes.  If device OEMs are working to retool, refactor  and deploy Android for netbooks, why can’t Google do the same?

Google’s rationale for launching Chrome as a separate OS from Android centers on differences in visual form factor and connectivity, as aligned with Google’s commercial goals. Google’s building and backing of Android  was motivated by the company’s desire to promulgate its search-and-advertising revenue model to ubiquitous 3G mobile hardware.

While Google could have continued with distribution of  various mobile apps (Maps, etc.) and mobile browser-based search on competing platforms (especially iPhone), the company appears to want to create a more integrated mobile experience and also be able to gather user data uniquely associated with mobile use cases.  Google also needed a mobile OS that had stand-alone, stateful capabilities, given the still-evolving level of connectivity and available bandwidth on 3G networks, especially in North America. Building Android to run Java (Dalvik is J2SE compatible) also gives OEMs deploying the platform the ability to leverage both existing mobile developer expertise (still very Java-centric) as well enterprise and desktop Java-savvy developers.

The desktop, by contrast, already has a working search-and-advertising model at larger visual form factors, but also enjoys high bandwidth connectivity.  For tethered devices with broadband, it makes more sense to build a web-based OS, in this case with the Chrome browser and a stripped down version of Linux.

Google’s choices reflect a dynamic I have documented in some of my recent blog entries, emphasizing that there is a difference in coming “down from notebooks” and “up from handsets”.  As sales of traditional (too soon to say legacy?) desktop PC and fully-loaded notebooks have flattened and perhaps begin to slide, the trend from PC manufacturers has been to cost-down those devices.  A web-based OS for PCs, notebooks and now netbooks supports that trend, requiring less memory, minimal mass storage, and less potent client CPUs (even as multi-core silicon becomes cheaply available).

On the “up from phones” curve, Android supports smarter smartphones, while still being a mobile phone OS (telephony, power management, footprint, LBS, etc.).

I myself do not see a clear point of convergence, just convergent trends.  The key for success for notebooks and netbooks with diminishing provisioning will be Google’s and others’ ability to leverage the cloud. To date, minimalist netbooks, running versions of Linux, have failed to make the jump from local to cloud computing for two reasons:  because PC OEMs and their channels still upsell on specs, leaving space for Microsoft as users opt to extend/mimic the desktop experience with Windows, Office, etc. and more strategically, because the value proposition of the cloud does not yet translate to the average end user.

With all the fuss over Chrome as a gauntlet thrown down to Microsoft, I think these and other basic questions are going unanswered.  I’d be eager to hear your views.

Filed under: Google, Linux, mobile, netbooks