GUC may not be very well-known in the U.S., but President Jim Lai says that the company's 2010 revenue of $327M placed them 14th on the Gartner Group ranking of application-specific integrated circuit (ASIC) companies. In July 2011, GUC announced they had added to their licenses for the ARM Cortex-A9 MPCore processor from ARM Holdings, with a new agreement that includes the ARM Cortex-A5 MPCore processor and ARM Mali-300 GPU. Earlier this month, Taiwanese technology news site DIGITIMES reported that TSMC and GUC had sent a team of 60 personnel to visit Apple for discussions of work on an upcoming A-series processor. GUC has approximately 500 employees worldwide.
Lai says that with more IDMs going "fablite", the 3rd party infrastructure has matured to the point that development of semiconductor intellectual property (IP), chip implementation, and manufacturing can all be outsourced. He finds that GUC's system customers increasingly require custom ASICs in order to create differentiation, and the virtual integrated device manufacturer (IDM) model based on GUC's close relationship with TSMC allows the company to act as a neutral 3rd party, since they produce no products of their own. GUC's "Flexible ASIC Model" offers fabless companies, system houses, and even IDMs the ability to outsource management of relationships with IP providers, EDA vendors, the foundry and package/test companies. GUC has a library of mixed-signal and interface IP that they have developed and can customize for particular applications. Besides ARM, GUC also works with Synopsys and other IP vendors to complete their library portfolio.
The majority of GUC's designs are currently being manufactured in TSMC's 65nm process node, with 40nm ramping up and 28nm test chips also in fabrication. GUC bases its ASIC projects on a non-recurring expense (NRE) charge without charging IP royalties directly, choosing to factor that expense into the margin that they build into the pricing of finished products. Lai tells the EE Daily News that their largest projects have approached 50 to 100M gates, in applications such as network processors. He foresees their first 20nm designs to be approximately two years away from production.
For the first-half of 2011, 44% of GUC's revenue came from U.S. customers, with 15% in Korea, 11% each in China and Taiwan, 10% in Japan and 9% in Europe. Wired and wireless communications made up 66% of GUC revenues in the first half of this year, with 21% in consumer products and 10% in computers - which Lai says is mostly attributed to Tablet PCs.
To complete the virtual IDM model, GUC provides services for engineering of packages - including thermal and electrical analysis. The company averages approximately 60 package designs per year, and ships more than 30M finished devices annually. GUC also provides test engineering, product engineering, quality and reliability, and overall supply chain planning and management. Lai also says that GUC has cumulatively delivered more than 16M system-in-package (SiP) devices.
Research in Motion (RIM) first announced development of a tablet device, the PlayBook, at their 2010 Developer's Conference (DevCon), in September last year. At the time, RIM positioned the device as "Professional Grade", based on the QNX operating system from the company which they acquired earlier that year.
RIM was then slow to get the tablet in the hands of application developers, though they held a "Meet the PlayBook" event specifically targeting that community in December 2010, where attendees weren't allowed to touch the device and only controlled demos were shown on-stage.
At the 2011 Consumer Electronics Show (CES) in January, RIM still kept the PlayBook under glass at a pre-show press event, before finally unveiling it with much fanfare on the show floor. RIM also then announced that the PlayBook would be available not only in a WiFi version, but also in a WiMAX version for the Sprint 4G network. A month later, at the Mobile World Congress, RIM announced LTE and HSPA+ versions, devices which were planned for availability in the second half of 2011. None of those devices have reached the market to date, and in August Sprint dropped plans to carry a 4G version on their network. The $500 16GB 7" device is still listed as available at RadioShack, but is poorly positioned to compete alongside other 7" tablets from companies such as Samsung and HTC, which offer more features (and apps) at a lower price, let alone the market-leading iPad-2 at the same price.
Nevertheless, at the just-completed 2011 Blackberry DevCon, RIM prominently featured the PlayBook as a centerpiece of their strategy during the opening keynote presentations. Each attendee was given their choice of a free Playbook, either the production version or an identical device pre-loaded with the 2.0 beta version of the PlayBook OS. The most significant feature of the next version of the PlayBook OS is that it will incorporate a runtime player for Android applications. During a press conference review of PlayBook Android strategy, company representatives emphasized how easy it is to port Android apps, and that they hoped once developers saw this they would be enticed to develop native versions for the PlayBook.
How the Runtime Android Player Works
 |
| Version 2.0 of the RIM PlayBook OS will include a runtime player for Android applications. |
A comparison of the PlayBook software stack (above) looks almost identical to the Google software stack (below). Since Android is available through an open-source distribution, what RIM has been able to do is to basically repackage Android on top of the QNX kernel, rather than open source Linux, keeping most of the library components and replacing (or augmenting) those that might be incompatible. With the commonality of a Linux foundation, it is - at least on the surface - pretty straightforward to see how Android apps can quickly be ported over. In the PlayBook OS, Android apps will be able to run as a secure BlackBerry PlayBook app.
 | |||
| The Android open source software stack is built on a Linux kernel. |
There are, of course, limitations. Perhaps most significantly, this approach precludes the use of Honeycomb tablet apps, because Google has not completely open-sourced that version of the Android OS. The latest version that RIM can use is version 2.3 series Gingerbread. So, while a growing list of companies have launched Honeycomb tablets, RIM is starting out at least one step behind. Now, with the introduction of Android version 4.0 - Ice Cream Sandwich - the same week, RIM is likely to fall even further behind.
Technical Limitations
During the 2011 DevCon developer sessions on the Android runtime engine, presenters described the key unsupported items:
- Native code.
- Add-on Google libraries e.g. Google Maps, Cloud to Device Messaging (C2DM), etc.
- You can invoke the native Camera app to capture a picture for inclusion in your app via android.provider.Mediastore.ACTION_IMAGE_CAPTURE.
- However, android.hardware.Camera APIs are not supported.
- No Session Initiation Protocol (SIP), Voice over IP (VoIP) or Bluetooth.
- No home screen = no widgets.
- Apps must have no launch-able activities.
- Two APKs can’t run in the same process.
- No Real Time Streaming Protocol (RTSP).
- Application icon must be re-sized to 86x86.
RIM says that a majority, approximately 65%, of Gingerbread Android apps have successfully ported without requiring a change in source code. The company is hoping that their more successful payment model will be enough to entice developers to go through the effort.
The much greater problem is drawing consumers to a device that, one year after introduction, has many more formidable competitors. Android tablet manufacturers are taking full advantage of later versions of the operating system, and they offer more advanced features at much lower prices. RIM also lacks a 3G or 4G version, and has no wireless carrier support. It is inevitable that, much like HP with the Palm OS, RIM will eventually come to realize an incomplete ecosystem is an insurmountable problem. Until and unless they address those issues, the PlayBook will just be another orphan device, Android or no Android support.
Related story
Altera has announced development of a high-definition video content analytics solution, incorporating intellectual property from Eutecus' Multi-Core Video Analytics Engine (MVE™), on a Cyclone® IV FPGA. The solution enables real-time video processing at 1080p, 30 frames/sec, according to Arun Iyengar, VP Military, Industrial, and Computer Division at Altera. The parallel processing in the Cyclone FPGA solution is capable of 60-Mpixel per second throughput, with pixel resolution that exceeds the capability of standard DSP-based approaches, he says. Eutecus estimates that to perform similar analytics, such as reading a license plate at 1024 X 1024 pixel resolution, would require a standard DSP chip to operate at a 7.3GHz clock rate.
Video content analytics differs from simple motion detection, says Ivengar, by executing complex algorithms that can check multiple user-defined rules in parallel. Since digital HD cameras produce much more data than the analog versions they are replacing, computation of the analytics in the camera, rather than in the cloud, conserves bandwidth and reduces latency over their internet protocol (IP) connections to host computers. Ivengar is seeing increasing demand for HD IP-connected surveillance cameras in Homeland Security type applications, and to provide more sophisticated traffic monitoring to detect accidents, stopped traffic and red-light violations, or identify vehicles moving in the wrong direction.
 |
| The Altera video content analytics solution embeds Eutecus Multi-Core Video Analytics Engine in a Cyclone® IV FPGA for in-camera HD video processing |
Users can update the rules and configure surveillance alerts in the analytics engine remotely, over the ethernet connection to the camera. The in-camera processing can be used to monitor video automatically, eliminating the need for manual scanning of a recording or constant monitoring of a video display, says Ivengar. The Eutecus' MVE embedded in the FPGA executes the massively parallel algorithms by combining specialized coprocessors with multiple 32b RISC Nios II cores in the Cyclone IV. In a typical camera integration, the raw pixel processing from the image sensor is first performed by a separate FPGA or DSP, The Cyclone IV then performs the analytics, and the output can be encoded by a 2nd DSP or application specific standard product (ASSP) video codec for streaming over the ethernet connection.
Eutecus provides a software GUI for the MVE that enables designers to customize the rules and parameters for event-detection to a specific application. The Eutecus API can be used by customers to interface surveillance cameras to their own video management systems, or to develop their own custom user interface.
Purchasing
The Cyclone-IV video content analytics solution is being offered on a royalty-based model, with the intellectual property sold directly by Altera, eliminating up-front non-recurring expense (NRE) charges. Altera will ship security devices along with the intellectual property, and the device must remain attached in order for the FPGA to continue functioning.
Purchasing
The Cyclone-IV video content analytics solution is being offered on a royalty-based model, with the intellectual property sold directly by Altera, eliminating up-front non-recurring expense (NRE) charges. Altera will ship security devices along with the intellectual property, and the device must remain attached in order for the FPGA to continue functioning.
 |
| Huawei's' 7" tablet for the T-Mobile network, runs the Android Honeycomb operating system. |
The most anticipated announcement (at a very lackluster CTIAena) this year was to be a Samsung "Unpacked" event with Google, where the companies were expected to unveil the first devices running the converged smartphone/tablet Android Ice Cream Sandwich operating system. However, the companies postponed that event just one week before CTIAena, they said out of respect for Steve Jobs' death, and it is now scheduled for Hong Kong(?) on October 18th.
The Samsung-Google postponement opened the door for Huawei, a $15B Chinese company that is better known for their network infrastructure equipment business. The company had also invited the press and analysts to their event at CTIAena, while acknowledging scheduling at exactly the same time as the Samsung event. While Huawei has been introducing Android smartphones in foreign markets for several months, they had no presence on a Tier-1 U.S. carrier. Until now. With many members of the press/analyst corps in attendance, Huawei announced their 1st first smartphone for the AT&T network (the Impulse 4G), and a 7" tablet running the Android Honeycomb operating system for T-Mobile's network, the Springboard.
In making the announcements, Executive VP of Product & Marketing in Huawei's Device Division - James Jiang, said that the company's strategy for the U.S. was to introduce a series of affordable devices based on Android. However, the Springboard - the U.S. incarnation of the 7" MediaPad which Huawei announced in June, actually exceeds the specifications of more well-known competitor's tablets, such as the Samsung Galaxy Tab.
The Huawei Springboard runs the latest version of Honeycomb (to the Galaxy Tab's Froyo), on a dual-core 1.2GHz Snapdragon processor (to the Tab's 1Ghz single core Hummingbird). The Springboard surpasses the Galaxy Tab in several other respects as well, with a higher resolution camera, T-Mobile "4G" connectivity, a higher resolution display, and HD video capture and output. Jiang told the EE Daily News that Huawei has been working closely with Google to develop their products with the latest versions of Android.
On September 28th, Amazon made a big splash with the announcement of a $199 color Kindle Fire, also a 7" tablet. The Kindle Fire is WiFi only, and is intended as a consumption device for content purchased from Amazon, hence a lack of any cameras. The Kindle Fire is more than one ounce heavier than the Springboard, with lower screen resolution that matches the Galaxy Tab, and only 8GB of storage.
Media and analysts at CTIAena pressed both Huawei and T-Mobile for pricing and release information, but the company would only say "Available in time for the holidays, no pricing details released yet". T-Mobile had been offering the 7" Galaxy Tab, now out of stock, for $199 with a $100 rebate. Sources have told the EE Daily News that they expect T-Mobile to offer the Springboard for "less than $200". While the downside of that pricing may be the requirement of a 2-year contract, the Huawei Springboard could definitely throw some water on Amazon's Fire in the coming holiday shopping season.
Related articles:
Sprint Nextel held a 4G Strategy/Network Update meeting in New York City on Friday, October 7th. As has been anticipated for several months, in the meeting Sprint executives announced their plans for a "aggressive" rollout of LTE - a departure from their current 4G WiMAX strategy. This represents a strategic change of direction for Sprint, whose 4G strategy has up until now been based on joint development of a WiMAX network with Clearwire. Sprint remains the majority shareholder of Clearwire, with a 53.9% share as of the end of 2010. However, the relationship between the two companies has become contentious, in what Clearwire refers to as "pricing disputes with Sprint relating to 4G usage". In the update meeting, Sprint attempted to distance themselves from Clearwire, evasively referring to Clearwire only as a "wholesale" supplier for WiMAX.
While Clearwire is working to convert their network for TD-LTE, Sprint has chosen to pursue a different strategy for their LTE migration. Sprint intends to re-purpose much of their existing spectrum holdings for LTE, and to collaborate with LightSquared for additional LTE capacity. In the October 7th meeting, Steve Elfman, President of Network Operations and Wholesale at Sprint, said that the company would begin adding LTE to its existing 1.9GHz CDMA network in mid-2012, with targeted completion at the end of 2013. Sprint will also look to add LTE in LightSquared's proposed 1.6GHz LTE spectrum, if that meets FCC approval, for deployment by the end of 2015. Ellfman also said that Sprint's LTE coverage will "hopefully" include the re-purposing of 800 MHz spectrum, which previously made up the Nextel iDEN push-to-talk network.
Much of Sprint's network update concerned their "Network Vision", a plan to implement multi-mode base-stations which the company announced last December. Sprint says that the multi-mode base stations will lower their costs of operating multiple networks, which currently represents a competitive disadvantage compared to the other U.S. operators. However, the company did not address what impact a multi-mode LTE strategy will have on end-user devices, which will presumably need to support as many as three different frequency bands: 1.9GHz, 1.6GHz, and 800 MHz. Sprint spoke only to the issue of multi-mode devices that will support 3G CDMA with WiMAX and/or 1.9GHz LTE. Sprint is looking to Qualcomm to supply 3G/4G multi-mode chipsets for dual-mode CDMA/LTE handsets, which the company expects to launch in mid-2012.
Sprint admitted that their 1.9GHz spectrum does not provide a sufficient footprint for the LTE coverage that they seek, while Clearwire has long claimed an advantage in their average of 150MHz spectrum holdings in the 2.5GHz band. Sprint's Bob Azzi, Senior VP of Networks who presented during the LTE update, had spoken of this advantage just a little over one year ago - during a 4G WiMAX Developer's Symposium that Sprint and Clearwire hosted at Stanford University.
Sprint could have pursued a strategy to continue leveraging the Clearwire spectrum advantage, by developing dual mode WiMAX/TD-LTE devices, as Sequans is doing with Malaysia-based WiMAX solutions provider Greenpacket. Sprint CEO Dan Hesse had previously declared 2010 as the "Year of 4G". By pursuing this new strategy, he is apparently willing to change that to 2013, totally giving up whatever first-mover advantage the company may have had before Verizon Wireless embarked on their own aggressive rollout of LTE.
Related articles:
- Dual-mode devices will ease the transition from WiMAX to LTE
- notes from the 4G WiMAX Developer's Symposium at Stanford (part 1)
- Clearwire's Silicon Valley WiMAX Innovation Network
The just-concluded TV Next conference in San Jose, on October 4-5, provided an opportunity to hear from a number of companies that span the broadcasting and internet ecosystems, on their visions of "Tomorrow's TV and Multi-Platform Experiences". Just over one year ago, Google announced that they, (along with Intel, Sony, Logitech, Best Buy, DISH Network and Adobe), would attempt to merge those two worlds of broadcast and internet-based content by developing a "Open Platform to Bring Web to TV". So far, the company's attempt to leverage the Android operating system in the creation of Google TV has been less than a rousing success. In fact, the effort to bring the web to TV was blocked, when the major broadcast networks responded by denying access to their websites on the Google TV hardware.
At TV Next, Shanna Prevé, Business Development Lead for Media Content Partnerships in Google Video Services and Google TV, shared five lessons that she says Google has learned from the "1.0" release, in her presentation: "Do We Finally Have Convergence in the Living Room?" Preempting questions about last year's launch announcement, she began her presentation by stating the fact that Google TV is NOT an open operating system. Only Google's chosen partners can develop for Google TV.
Beyond that, the first lesson she said, is that Google was reminded that their forte is search. The company has found that searching for content is the most commonly used function by early adopters of Google TV. This would seem to be a foregone conclusion, since a specialized content search was one of the top features that Google promised to deliver in their TV system.
But, it turned out, Google's design of a modified search engine for video was not what users wanted. Prevé said that Google's second lesson-learned has been that they can't predetermine where users will want to go on the web. Google's original assumption was that users would only want to search for video on their TVs, so they attempted to prioritize such content in their results. The discovery that people wanted to go to the same websites as on their PCs "came as a surprise", she said. The lesson here is that if you give a user a display with internet connectivity, they expect to be able to access the entire web.
Lesson #3? Though Google was excited by all the clever features in their TV 1.0, users were confused by what Prevé called the "new paradigm". As a result, Google plans to remove the list of nested menus in the 1.0 UI design, and in the forthcoming 2.0 version a "home" button will always take users back to watching TV, where a small translucent overlay of apps will appear on a menu bar.
The fourth lesson that Google learned, according to Prevé, is "when you know what you are looking for, search is great, and if you know specifically what you are looking for it is even better". According to Google's studies of usage, viewers only know what they are looking for half the time. The Google TV "What's On?" feature got much more usage than the company expected. In other words, viewers tendencies to "channel surf" did not go away when they were given other functions for finding content. It sounds like Google didn't do enough research ahead of time on how most people watch TV. Had they never heard the expression "500 channels and there's nothing on"? Typical pay-TV subscribers would greatly benefit from a better discovery process, to shorten the amount of time wasted flipping past all the infomercial channels.
Finally, the biggest lesson of them all (which Google should have known from their Android Open Handset Alliance experience), is that it takes an ecosystem to build a product that delivers value to consumers. The last painful lesson that Google learned is that "Content is King". They should have learned this lesson from the Google Books debacle. Some content owners just do not want their property organized by Google.
In conclusion, Prevé said that Google now sees their TV effort as being very very early, and that after four years of development they finally have a product that they (now) wish they could have released last year. There are many big issues still unaddressed however. Though the Google TV SDK and emulator have finally been released to developers, the Android App Store is not yet open for TV. It will be "soon" said Prevé. In regards to dealing with content owners and distributors, she would only say that Google is making a better effort to make them aware of the next TV product and its features. As expected, nothing could be said about incorporating Google TV into Motorola Mobility's set top boxes at this point either.
So, as they say in the TV world, stay tuned!
As a complement to the September 22nd EE Daily News article on companies that have filed patent lawsuits against Apple, here is a roster of Apple lawsuit targets (* indicates company is also counter-suing Apple):
- Amazon (denied in U.S.)
- HTC*
- Motorola* (multiple, for XOOM and multi-touch)
- Nokia*
- Samsung* (numerous, in many countries)
- JAY-tech (German)
- A host of accessory manufacturers, including:
- Eforcity Corporation
- Accstation Inc.
- Itrimming Inc.
- Everydaysource Inc.
- United Integral Inc.
- Crazyondigital, Inc.
- Boxwave Corporation
For both lists, if we have missed anyone (or if you are being sued by Apple), drop us an email and let us know.
Conexant has announced the launch of a new audio playback product line, starting with the addition of the KX1400, which the company has acquired from Keterex. Keterex designed the KX1400 to play 8KHz audio data directly to an external speaker via an on-chip digital audio processor and class-D amplifier that is capable of driving an external 8Ω speaker.The on-chip digital audio processor in the KX1400 supports audio data sampled at 8 kHz and encoded in the International Multimedia Association (IMA) adaptive differential pulse-code modulation (ADPCM) or 12-bit pulse-code modulated (PCM) uncompressed formats. With a Nyquist limit of approximately 4KHz and 12-bit sampling, Conexant is positioning the device for applications that do not require high-resolution audio, such as vending machines, toys, pedestrian crossing signage, appliances and interactive kiosks.
You can operate the KX1400 in a standalone mode to stream audio data from an external serial memory, or under control of an external microcontroller (MCU). In stand-alone mode the KX1400 plays a single phrase from external memory upon reset, and then returns to a stand-by power-saving state. According to the Keterex data sheet, in host-controlled mode, you connect a MCU via a 4 or 5-wire host interface. You can then send audio data directly to the KX1400 audio processor, select one of 4096 different phrases to play from an external memory, or play tones using an on-chip tone generator.
