GSA Silicon Series: New Markets, New Economics, Feb 10, 2010

This very informative, analyst – only panel discussion (Feb 10th in Santa Clara, CA) assessed the outlook for new and growing markets throughout the semiconductor industry. While several markets were covered, this summary will focus exclusively on the hot communications related markets and products. During the meeting, analysts from FBR Capital Markets, iSuppli, Databeans and Gartner shared their perspectives and predictions on what's hot and why in several markets of interest to communications and network practitioners.

Craig Berger of FBR Capital Markets singled out smart phones and Set Top Boxes (STBs) as key semiconductor industry drivers. He said that China's 3G infrastructure build-outs would have ripple effects across the industry and accelerate in late 2010. China is expected to spend over $60B in 3G related telecom equipment over the next two years. That's certainly impressive! Craig opined that he expected India to "ramp up" 3G network production, but didn't say when (India's 3G licensed spectrum auctions have been postponed for well over one year now). In addition to 3G, Mr. Berger commented that wireless interfaces, such as WiFi, Bluetooth, and GPS, were propogating much more broadly into hand held devices that didn't have much or any IC content previously.

FBR sees a continued proliferation of smart phones, with +15% growth in 2009 and +25% expected in 2010. Handset chip content is actually increasing versus recent years. Other gadgets are expected to do well – Tablets, eBook readers, and MIDs. Chip makers that will benefit from this advanced wireless handheld trend are Qualcomm, Infineon, Broadcom, Marvell, Nvidia, Intel, and STMicro, according to Craig.

Distributed computing is another trend to watch says Mr. Berger. As broadband becomes faster and increasingly available, PCs do not need to have large amounts of on board storage, expensive operating systems and processors. Instead, they can be 'dumb terminals‘ that just interface to Internet or cloud computing based servers.

Success is created when products are specifically designed to address a particular market niche. For example, Bluetooth is really 20 different sub-markets, depending on application and usage models. Broadcom has done particularly well in these sub markets by offering a wide variety of components and combo chips that address the different requirements of the sub markets.

Mr. Berger noted four "Cool and Innovative Products for 2010:"

1. Intel’s Nehalem Quad-Core CPU on 32nm

2. Qualcomm’s Snapdragon on 45nm on a single chip includes:

  • 1GHz ARM CPU
  • 600MHz DSP
  • 3G baseband modem
  • GPS (7th Generation)
  •  720p Video Decode
  • 12 megapixel camera support
  • Multiple audio codecs and MediaFLO TV support

3. Broadcom’s Bluetooth/WiFi/FM (BCM4329) on 65nm includes:???? 802.11n A/B/G

  • Bluetooth 2.1 +EDR
  • FM receiver and transmitter (transceiver)
  • On chip power management unit (PMU)

4. Broadcom’s Single Chip Integrated Access Device (IAD) on 65nm combines many voice, video, data functions in 1 chip:

  • ADSL2+ DSL
  • IEEE 802.11n WiFi ( 2×2)
  • Gigabit Ethernet switch (4 PHys)
  • VoIP functionality (DECT)
  • 400MHz dual core processor

Susie Inouye of Data Beans stated that the consumer electronics industry was being fueled by new media. Hot devices include: Blackbery, iPhone, Droid smart phones; eReaders, flat screen TVs, and netbooks. Her predictions were that:

  • Pads/Tablets will be disruptive to eReader and Netbook Segments of the PC market
    • iTunes to begin selling books
    • Book Rentals/"Borrowing"
  • Lower Cost Notebooks would be negative to the Netbook Market (and what about "smartbooks"?)
  • Android OS to Gain Market Share
  • Volume Ramp for LCD/OLED Displays
  • Volume Ramp for Capacitive Touch

The most interesting presentation from a communications perspective was from Lee Ratliff, Senior Analyst for Broadband Digital Home and IPTV markets at iSuppli. His presentation was entitled, "Uniting the Digital Home" – something Cisco and Taiwanese equipment companies have been trying to do for some time. Lee started his talk with a question, "What will drive broadband data rates to the digital home?" He noted that in Japan, the average broadband residential subscriber gets 65M bits/sec – much more than the highest speeds available to U.S. home users. Lee believes the U.S. and rest of the world is headed for much higher residential broadband rates, driven by new applications and appliances that are "arriving all the time." This what will necesitate what he refered to as "High-QoS Home Networks." For the foreseeable future, Lee says those networks will be wired (they won't be wireless). Power Line, Coax (MoCA), and Phone Line all be used for media distribution to the TV, STB, DVR, PC, gadgets galore, etc. New Cat 5 (or higher grade) twisted pair has proved too expensive and difficult to install, so it won't be used.

The Market Drivers of High-QoS Home Networks include:

  • IPTV- through a managed service provider network with high QoS (not the public Internet)
  • Multi-Room DVR
  • DBS single-wire distribution
  • Consumer Electronics: Media extenders, Internet Enabled TVs (IETV), and much richer and broader VoD

The IETVs will be connected to the Internet mostly via Ethernet (we assume this to be GE rather than the 10/100 BaseT most of us use to connect our PCs to a cable or DSL modem). IEEE 802.11n may be used in some cases to connect to a home WiFi router.

Two new emerging standards of note are ITU G.hn and IEEE 1901, although Lee believes only the former has real potential to become a mass market. G.hn provides an evolution from the various wired home networking technologies like MoCA, Home PNA, and Homeplug Powerline to a new unified standard that supports cable, power line and phone line transmission.

Cable and power line are expected to be the dominant PHY layer for home networking gear till at least 2013. Coax (the MoCA standard) will dominate the North America market because of the prevalence of coax cables in homes. Power Line home transmission is expected to dominate for the rest of the world, especially Europe where it's being used now to distribute IPTV within the home. (Please contact Lee Ratliff directly – lratliff at isuppli.com – for his market projection slides).

As the home networking industry matures, the small innovative start-ups are likely to be acquired by larger companies. Such consolidation is well underway:

  • Sigma Designs / CopperGate
  • Atheros / Intellon
  • Lantiq / Aware
  • Broadcom / a power line data/video distribution company later this year?

Service Providers are the dominant channel for home networking – much greater than retail. Lee says that service providers are opting for wireline home transmission (over wireless) because it's reliable, stable, and offers high performance. He believes that WiFi will be used in rare cases. Consumers shopping in retail stores will continue buying WiFi enabled gear, because it's ubiquitous, easy, and portable. And standards confusion with G.hn and IEEE 1901 may give IEEE 802.11n "room to run" with many home CE makers using 802.11n instead of a new wireline standard. Wireline bridge products might be used, but only when WiFi fails to work throughout the digital home.

Sergis Mushell, Principal Research Analyst at Gartner Group talked about Cloud Computing's impact on the semiconductor industry. Cloud services will provide scalable, elastic IT related capabilities provided as a service to external customers using Internet technologies. But one needs to take note of the different characteristics of public and private clouds.

Intel's microprocessor architecture vs licensed ARM cores was said to be the battleground at the intersection of the mini-notebook, netbook and smart phone markets. These smaller computers would be more prevalent as cloud computing matures. Indeed, the big beneficiaries of cloud computing will be low end notebooks, smart phones and mobile network infrastructure to support all the new wireless gadgets and gizmos that will access the cloud (e.g. an eReader to download books, magazines or newspapers). Clouds will have negative impact on client storage and mainstream PCs (both desktops and laptops). Much better security will be needed across many product lines.

Data Centers represent an opportunity, because that's where the cloud servers will reside. The servers must operate at high capacity, be robust and reliable, but not have the high power consumption that results in expensive power bills and cumbersome cooling requirements. Storage security will be particularly important.

Sergis Mushell told the audience to look for opportunities in new devices that place cloud computing capabilities (via broadband wireless access) in the hands of consumers. He predicts that service providers will offer cloud services to consumers in addition to triple play or quad play services offered today.

——————————————————————————————————————————

 Please leave a comment if you have any questions or issues with this article.  The author will try to respond as time permits.

0 thoughts on “GSA Silicon Series: New Markets, New Economics, Feb 10, 2010

  1. If the Comcast commercials are to be believed, at least their footprint will soon start to see the 100 Mb/s speeds.  Maybe we aren't too far behind the 65 Mb/s cited in this article and all of the cool applications that will follow.  It is amazing how these seemingly disparate technologies meld together to create things unimaginable (at least by me) a short time ago.  

    I saw a blue-tooth enabled stethoscope the other day.  Apparently it digitizes the signal and sends it to a PC where it can be analyzed and/or sent directly to a doctor.  This could have some real cool applications for rural areas, where doctors are not readily available.  It's not too much of a stretch to envision this device being another "app" on a smart phone; all things that are possible thanks to the chip-level innovation described in this article.

  2. Ken,  Thanks for your commment.  What most people don't realize is that it's advances in silicon technology that makes all these new gadgets and gizmos possible.  As a former IC chip architect and microprocessor apps engineer, it is truly astounding to observe all the functionality in Qualcomm's Snapdragon and the two Broadcom combo chips referenced in the article.
    On the subject of cloud services to mobile devices, I believe we'll need much more bandwidth and QoS than we now have in 3G networks.  We'll also need much better security across the board.

  3. Great article!  Especially liked the section on High QoS home networking.  Wonder if we can get QoS within a home network when we can't even get it over the broadband Internet or Clearwire's version of mobile WiMAX.  What will be the QoS mechanism – will it be at the MAC or IP layer?

  4. Can we be sure that all devices on a home network understand IP v6 or Diffserv or other higher layer QoS?  I don't think so.  PHY layers are for transport w/o any QoS, so that leaves MAC layers where different QoS mechanisms are used.  If multiple QoS methods are used in the home network, there'd need to be mapping from one to the other.  So I don't see how there can be any real QoS in home networks.
    Ironically, the only network technology with very strong QoS was ATM, but it has been relegated to carrying PPP over Ethernet over AAL5 traffic on DSLs.

  5. High-QoS home networking is certainly feasible and desirable, even while we continue to face QoS challenges on the public internet. I believe that for at least the next five years high-QoS home networks will be driven by service providers seeking to enable advanced data services like IPTV or multi-room DVR. Since the networks are designed, installed, and managed by the service provider, QoS is almost entirely within their control.  Despite the fact that the home might have a <25 Mbps broadband connection, the home network needs to be capable of much higher performance – measured in 100's of Mbps today – and far better QoS.  Intra-home activities like large file transfer, multi-room DVR, video trickplay, video intercom, and video surveillance will drive bandwidth requirements.  The industry is learning to cope with poor (or non-existent) QoS on the public internet through buffering, VOD delivery (as opposed to linear streaming), and the use of private IP networks (like Uverse). However, having QoS makes things a lot easier. For instance, most multi-room DVR satellite STBs do not have hard-drives – they depend on high-QoS to avoid buffering and the expense of the HDD.
    As for the implementation of QoS in home networking, each standard implements it differently.  Here is a good whitepaper covering MoCA QoS:
    http://www.mocalliance.org/industry/white_papers/PQoS_White_Paper.pdf
    Lee

Leave a Reply

Your email address will not be published. Required fields are marked *

I accept that my given data and my IP address is sent to a server in the USA only for the purpose of spam prevention through the Akismet program.More information on Akismet and GDPR.

This site uses Akismet to reduce spam. Learn how your comment data is processed.