Independent Analysis of the Networking-Silicon Industry
Editor: Linley Gwennap
Contributors: Tom Halfhill, Joseph Byrne
In This Issue:
-Market Share Report Now Available
-Exar Jumps Ahead with ODU0/ODUflex Muxponder
-Vitesse Tops 14Gbps
-Freescale Amplifies QorIQ Family
-Embedded Processor Market Up 31%
-Processor Conference on October 5-6
To read these articles on-line, click here
Market Share Report Now Available
The new Linley Group report, Communications Semiconductor Market Share 2010, is now available for immediate shipment. This report provides all-new market share data for more than 20 categories of communications semiconductors, including Ethernet products, broadband interface chips, embedded processors, FPGAs, interconnect chips, and home-networking technologies, such as powerline communications (PLC) and Moca. New in this edition is market data for 10 Gigabit Ethernet switch ASSPs.
For more information, visit our web site.
Exar Jumps Ahead with ODU0/ODUflex Muxponder
By Jag Bolaria
This week, Exar announced availability of the MXP2 muxponder chip. It multiplexes multiple client ports to 2x10Gbps OTN ports. The client ports can be Ethernet, Fibre Channel, SONET/SDH or HD video, which are mapped to OTN ports. The OTN containers can be ODU0 for Gigabit Ethernet to ODU2 for 10Gbps Ethernet. It also supports ODUflex allowing the OTN container to better match client-bandwidth requirements. In addition to ODUflex, the MXP2 supports GMP mapping and network-delay measurements. The latter is useful for latency-sensitive applications such as Fibre Channel.
Target applications for the MXP2 include ROADMs, Packet-Optical Transport Systems (P-OTS) and MSPPs. The device supports OTN ring functionality by switching ODU0, ODU1, ODU2, and ODUflex containers. The dual 10Gbps OTN ports support protected line cards in OTN ring networks. The device has a low power consumption of 8W (typ).
The MXP2 is the industry’s first ODU0/ODUflex muxponder and it supports the broadest range of client ports. Relative to competing ASSPs, such as PMC-Sierra’s HyPHY, the MXP2 offers ODUflex, delay measurement, and adds muxponder functions in addition to the transponder functionality of the HyPHY. Today, most muxponders are built using FPGA devices. The MXP2 is a single chip in comparison to the 2-3 FPGAs required to do a similar function. With a unique set of features, leading integration, and the lowest power consumption, the MXP2 is primed for success once Exar qualifies it for production.
Vitesse Tops 14Gbps
By Jag Bolaria
This week, Vitesse announced availability of the VSC7224 multiprotocol quad-channel extender. The chip can be used to extend signal reach on a board or over a cable for different networking protocols. It supports four channels, which can be Ethernet, Fibre Channel (FC), or InfiniBand. Each channel has a frequency range of 155Mbps to 14.1Gbps—enabling the device to support services such as 4 x 10GBase-KR, 4 x 16G/8G FC, and 4 x QDR/FDR InfiniBand.
The VSC7224 includes a four-tap decision feedback equalizer to enable a reach of 65 inches on FR4-material printed circuit boards, or up to 20 meters over 24-gauge twin-ax cable. This level of signal conditioning may allow customers to upgrade legacy chassis to faster data rates. The chip can be used as a redriver or a retimer with its integrated clock/data recovery circuits. The typical power dissipation is 100mW/channel in the redriver mode and 180mW/channel in the retimer mode.
Currently sampling, the VSC7224 offers leading throughput with competitive power consumption. The flexibility of using it as a redriver enables designers to further reduce power consumption. Support for all the leading client interfaces allows designers to use this device in different systems and in multiservice environments. With the VSC7224, Vitesse further strengthens its position as a leading supplier of signal-integrity components.
Freescale Amplifies QorIQ Family
By Tom R. Halfhill
Freescale has announced a whole new series of QorIQ-family processors that will deliver about four times the performance and twice the power efficiency of today’s best P-series chips. Scheduled to begin sampling early next year, the AMP (Advanced Multiprocessing) series will debut with Freescale’s first multithreaded CPU core, up to a dozen CPUs per chip, higher clock frequencies, faster offload engines, resurrected AltiVec extensions, and other goodies.
The first T-series AMP processor will be the T4240, the flagship of the fleet. It will have 12 dual-threaded CPUs (or 24 “virtual cores,” as Freescale describes them). The new CPU is the Power e6500, Freescale’s second-generation 64-bit core. Clock frequencies will reach 2.0GHz—enough to sustain packet forwarding at 48Gbps in data-plane applications. For control-plane processing, future T5-series chips with six CPUs will aim for clock speeds as high as 2.5GHz. The T4 and T5 series will be Freescale’s most powerful multicore designs, surpassing today’s dual-CPU P5020 and eight-CPU P4080. Future T1- and T2-series processors will be lower-priced, lower-power products that reach 1.6GHz and dovetail with the existing P1 and P2 series.
Freescale is promising an ambitious production schedule. After the T4240 begins sampling in 1Q12, the company plans to unveil a new T-series chip every quarter. Freescale usually starts volume production about a year after initial sampling. In the past two years, the company has introduced 24 P-series processors, so a similarly relentless rollout will unleash a barrage of T-series processors intended to pulverize the competition. Freescale wants to exploit its greater development resources and larger sales network to beat smaller but aggressive companies like Cavium and NetLogic. The only larger competitor is Intel, which lacks highly integrated chip designs.
Although Freescale is withholding some important technical details about AMP for now, the product roadmap looks impressive and achievable, given the company’s history. Competitors will continue to offer processors with more CPU cores—currently, up to 32 single-threaded CPUs from Cavium and 16 quad-threaded CPUs from NetLogic—but Freescale’s broader product line covers the largest swath of the market, leaving the smaller, high-end niche to other companies. Also, for the first time, Freescale and NetLogic will manufacture their leading products at the same process node (28nm). This level ground will tilt the competitive balance more heavily toward their design skills.
The complete Microprocessor Report article is available here: http://www.mpronline.com/mpr/h/article.php?id=10691
Embedded Processor Market Up 31%
By Joseph Byrne
Sales of embedded processors rebounded strongly in 2010, renewing a linear growth trend interrupted by the 2009 downturn, according to the newest market-share report by The Linley Group. Intel and Freescale claim the top rankings; no other company comes close to their size. Freescale is as big as all the smaller suppliers combined.
Despite their great size, Intel and Freescale gained share, growing 32% and 39% respectively compared with 31% for the overall market. The fastest-growing suppliers in the top 10, however, were Cavium and NetLogic, owing to the success of their multicore processors targeted at communications. Cavium’s embedded-pro¬cessor revenue increased 91%. Were the company to maintain this growth rate for three more years, it would be almost as big as Freescale was in 2010.
An upstart, however, has practically no chance to continue such a torrid growth rate. As a company gets larger, a given revenue gain translates to a smaller rate (e.g., an increase from $1 million to $2 million is 100% growth, but an increase from $100 million to $101 million is 1% growth). Freescale, moreover, has yet to ramp production of its multicore processors, which will take a bite out of Cavium’s and NetLogic’s burgeoning sales.
For now, the real king of the multicore hill is Intel. We estimate that the company commanded nearly half of the embedded-multicore market in 2010. Intel’s processors, however, tend to find different uses compared with processors from Cavium and NetLogic. Whereas the latter companies target data-plane functions in security and various other communications systems, Intel’s multicore processors are more likely to perform control, analysis, and other high-level applications.
Looking at communications specifically, the competitive landscape changes. Freescale takes the pole position, and Intel comes in second. Cavium overtook AppliedMicro in 2010, with NetLogic rounding out the top five. In this segment, Freescale did particularly well, growing 38%. Not only did the company rebound with the market, it also benefited from new designs that are ramping.
Complete market-share information appears in our new report, Communications Semiconductor Market Share 2010, available on our web site. http://www.linleygroup.com/report_detail.php?num=1
Processor Conference on October 5-6
The Linley Tech Processor Conference will return to San Jose on October 5-6, 2011. The two-day event will focus on processors and related technologies for networking and communications applications. Industry leaders will deliver in-depth information on the newest chips and technologies. Attendees will hear presentations addressing various types of processor architecture and design, semiconductor intellectual property, and related technologies used in these applications.
The 2-day, single-track conference will address the following topics:
* High-speed embedded processors
* Processors for control and data planes
* Processor IP
* Software for embedded processors
* Network and platform security
* Memory technology for processors
* Processor ecosystems
* Multicore processors
* Single/dual-core processors
* Communications processors
The conference will include lively panel discussions, opportunities to talk with the presenters, meet Linley Group analysts, Microprocessor Report editors, and an evening reception with sponsor exhibits, demos. It will feature presentations from AppliedMicro, Cavium, Freescale, NetLogic, MIPS, EZchip, Netronome, LSI, Xilinx, Continuous Computing, Authentec, Tilera, 6Wind, MoSys, ARM, Intel, and The Linley Group.
The conference will include:
> Sponsor exhibits
> Evening networking reception
> Door prizes
> Hosted speaker tables at lunch (meet the presenters)
> Validated parking
> Continental breakfasts and gourmet lunches
The conference is intended for system designers, network-equipment vendors, OEM/ODMs, carriers/service providers, press, and the financial community. Qualified attendees receive free admission. Additional details are available on our web site.
About Linley Wire
Linley Wire is a free electronic newsletter published by The Linley Group, a technology analysis and strategic consulting firm. Linley Wire provides our analysis of recent news on semiconductors for networking and communications. Articles are posted weekly to our web site and sent monthly via email. To access the web content directly, visit our web site.
Linley Wire is not affiliated with any outside vendors. We do not rent or sell our mailing list; it is used only to send you the newsletter and information on our events.
We encourage you to forward this newsletter to colleagues who may benefit from receiving this information. They may subscribe by visiting our web site.
Copyright 2011, The Linley Group