Thursday, December 28, 2006

Let's celebrate some innovators


Since I started this blog in late August I have written about many conflicts and failures. As the year 2006 winds down to a close, I'd like to do something positive by celebrating a handful of engineers and technology managers who contributed innovative ideas in interconnect technology this year.

Please post a nomination of anyone you would like along with a short rational, or drop me a line at rbmerrit@cmp.com. I'll cull through your suggestions and write about the best of them, holding the innovators up to inspire us to great things in 2007.

Wednesday, December 27, 2006

Are you plugged in to Cable Card?


If so, I want to hear from you, pronto by a posting here or email to rbmerrit@cmp.com Word is Cable Card could see a renaissance at the Consumer Electronics Show where I will be with waaaaaaay to many other people in Vegas next month.

The FCC has mandated all new set-top boxes shipping after July 1, 2007 must have Cable Cards slots. And I have reason to believe Windows Media Center PCs are eyeing Cable Card as a way to finally plug in to digital cable and maybe satellite services, now that Vista brings enough security to assure service providers to let their digital broadcast content flow.

If you are up to speed on this, I'd like to hear about the implementation and market issues for the one-way Cable Card 1.0 vs. the two-way version 2.0 and the software-based Open Cable Applications Platform. What are people choosing and why? Thanks!

Friday, December 22, 2006

No relief in sight for digital display battle


As the year comes to a close I recall an Intel exec telling me he would push for a resolution of the battle between DisplayPort and the Unified Display Interface. Well, I guess that will have to go on the wish list for next year.

All sides appear to still be moving forward in this skirmish that pits the likes of Dell and Hewlett-Packard on the DisplayPort side with their biggest vendor, Intel, a proponent of UDI.

"All of us would appreciate it if someone would sort that out," said Bill Bucklen, advanced TV director for Analog Devices Inc. in an interview today. The company has no plans for UDI or DisplayPort chips, he added.

ADI rolled out its first HDMI version 1.3 products this week. It will demo a wireless media adapter running HDMI over the ultrawideband chips from Tzero Technologies at CES next month.

"We expect to see one or two TV makers bring this sort of adapter box to market for Christmas 2007," Bucklen said.

Here's hoping this holiday season brings you all broadband connections with those you love. --rbm

Thursday, December 21, 2006

UWB makes a mesh

Kiyon, Inc. has its mesh routing software up and running on the ultrawideband chips of startup Tzero Technologies. Siemens announced today it will ship home networking systems early in 2007 using the silicon and software.

This is the first use of mesh networking over UWB I have heard of. Even the more mature Wi-Fi folk are still hammering out their mesh standard.

Mesh networks could be a great way for UWB to expand its reach beyond the living room to become a contender for a whole home net. But the proof is in the pudding. Even Kiyon's marketing people are not ready to put hard numbers on the number of hops or latency their solution will support on the Tzero silicon.

What's more an exec with a major set-top box maker told me today the company will not integrate any wireless networking into its products in 2007 and sees "little interest" for UWB among its customers. That's one pretty big link refusing to be part of the mesh--rbm

Monday, December 18, 2006

SPI 4.2 update hits 6+Gbits

A new high-speed serial interface hits the market today with the release from the Optical Internetworking Forum of its Scalable System Packet Interface protocol today, a follow on to SPI 4.2 popular in comms chips. I wrote a round-up story on SPI-S and other major 5Gbit+ chip-to-chip links.


SPI-S and the Interlaken protocol from Cisco and Cortina Systems are the two head-to-head competitors in the group, profiled in a story earlier this year.

The big picture: at 6Gbits and beyond scopes and probes are less useful and new testing techniques are needed. That's keeping consultants like Howard Johnson and Eric Bogatin busy. It's also opening up new opportunities for companies like SiSoft that will introduce a product in 2007 to deal with the issues. Watch this space! --rbm

Friday, December 15, 2006

Staccato beats on the UWB pan

My colleague Loring Wirbel recently posted an excellent story about Staccato Communications Inc. and Korean carrier SK Telecom striking a partnership to push new uses for ultrawideband in handsets. The upshot is the duo hope to ride UWB for use in apps such as point-of-presence retailing and social networking using both Bluetooth and WiMedia protocols.

Loring has an excellent grasp of the current and historic technical, market and political issues surrounding the emerging PAN area so I highly recommend his article. But in the end, only time will tell if Staccato's efforts to market its UWB chips create some new and significant music or just add to the noise. –rbm

Thursday, December 14, 2006

Third contender for wireless HDMI

Among the snowstorm of emails about CES hitting my in-box was one about Amimon. I was not aware this San Jose startup was designing a third contender for wireless HDMI, the uncompressed HDTV link.

I have written about startup SiBeam and the group it has formed to promote 60 GHz radios to deliver 5 Gbits/s for wireless HDMI. My colleague Loring Wirbel wrote about TZero and ADI's partnership to deliver wireless HDMI using ultrawideband.

Now enter Amimon which is taking a slight but significant twist on 802.11n to deliver 1.5 Gbits/s over a 20 MHz channel for wireless HDMI. True, it's their second time trying to make a big splash at CES with an FPGA prototype. But chips that initially could cost less than $50 and dissipate less than 5W per system are expected to be ready, perhaps in time for Xmas 2007 TVs.

There are some interesting stories to be told about what will and won't work for this hotly pursued application. Stay tuned. --rbm

Wednesday, December 13, 2006

Trio of new 6Gbits/s+ links coming

I wrote on December 4 about how this month the PCI SIG may complete work on the PCI Express 2.0 spec for 5GHz chip-to-chip links. Tom Cox, executive director of the RapidIO Trade Association, was quick to point out his group is also working on a 2.0 spec defining 5 and 6GHz links.

The RIO 2.0 spec still has to pass a final ballot, but Cox and crew have been presenting the details to engineers on a road show that has traveled across the US, Japan, China and India. "I must say that 5 and 6.25 Gbps are not the urgent need of the industry today, but the technology is available and the definition is important by the standards body before a de-facto choice just happens," Cox said.

There's even more to come. Watch this space Monday for news about another 6Gbit/s+ interconnect about to hit the streets. This one relates to the Cisco/Cortina Interlaken work I blogged about on November 2.

Meanwhile, I'd love to hear what you think—oh, board designer in the field—about the requirements and challenges of pushing chip-to-chip links to 6GHz and beyond. Drop a post or an email to rbmerrit@cmp.com

Tuesday, December 12, 2006

ASI goes quietly away


I reported back in August that the Advanced Switching Interconnect, a spinoff of PCI Express technology for comms systems, was effectively dead. No one from the ASI SIG or its primary backer, Intel Corp., ever fessed up to it and the ASI SIG Web page stayed live—until today.

It's officially a dead Web site now, though there is still material about ASI to be had on Intel's Web site—for awhile anyway. Will the technology be reformulated by Intel or anyone else, perhaps under a new name such as PCIe Embedded? Could be more shoes to fall.

If anyone wants to come clean about what's happening, I am all ears. --rbm

Monday, December 11, 2006

The attack of the killer Ns


A host of 802.11n draft products are rolling out this holiday season, including a consumer-focused chip set from Metalink announced today. The company has tapped the latest process technologies to shrink its three chip set down to two that consumer 2W maximum, cost less than $20 in high volumes and offer as much as 210Mbits/second of data throughput.

As these chips begin appearing in PCs, media adapters and a retail flood of routers and gateways, it could establish a beachhead of good-enough technology in the home. I think coax, powerline and HPNA will find a big threat with .11n this year. What's your view? --rbm

Friday, December 08, 2006

Can Apple fix Moto's Rokr?


OK, this ain't exactly interconnect stuff, but I am hungry for feedback on this one. Everywhere I go people say the music phone is one of the next big things in consumer electronics, but nobody has got it right yet. Not even giant Nokia whose music phone scares me it's so ugly.

Well, Apple set the tone in MP3 players with the iPod and in portable media players with the video iPod. OK, I give Sony plenty of credit for a tie here with the pricey PSP.

So is Apple up for a hat trick with the much rumored iPhone? I have more questions than answers.

What have music phones done wrong so far? What can Apple (or Sony if it could get its head out of its lithium ion) do right? What is the right consumer experience and what does that translate to in terms of silicon and software? And how would Apple avoid getting its design and revenue model ensnared in the tentacles of a Cingular and Verizon?

I'd love to hear what you think here or at rbmerrit@cmp.com.

Bluetooth smacks down UWB in 2007

Gartner Dataquest is quite bullish on Bluetooth as the rising tide in consumer personal area networks for 2007, much less so for ultrawideband. That's an interesting reality check given UWB clearly whomps BT's backside with waaaay more bandwidth.


Bluetooth is in about a third of the 800 million cellphones this year, rising to some 72 percent by 2010, according to a briefing by Dataquest analysts in San Jose today. BT also has a play as a headset attachment for portable media players like the iPod. UWB is in approximately zero phones and MP3 players now and still under the market watcher's radar for the next couple years.

So where's UWB? "It's initially a [USB] cable replacement for PCs, but it will take time to move into consumer electronics and particularly portable devices, because UWB really taxes a battery when you are in streaming mode," said Jon Erensen, a consumer analyst for Gartner Dataquest.

Perhaps like Bluetooth before it, UWB will have to fall from its hype phase into a phase when we declare "UWB is dead" before it comes roaring back in a few years.--rbm

Thursday, December 07, 2006

Getting to know mobile WiMax


Thanks to plenty of push from intel, and now Sprint/Nextel, mobile WiMax is giving the old fixed broadband wireless sector a much higher profile. Because this is one of the technologies you probably need to better understand, my colleague Jack Shandle wrote an in-depth tutorial on it. It became one of the top eight most read stories on his Wireless DesignLine site in 2006. --rbm

Monday, December 04, 2006

100-to-5: Ethernet trumps Express


No sooner had I finished posting my note about the searingly fast 5GHz PCI Express on the horizon (see below) than in my email box comes the news that the IEEE 802.3 Higher Speed Study Group has officially voted to support 100 Gbit/second as the next speed for Ethernet.

I know from earlier talks with the group's chair, John D'Ambrosia (see my Sept. 29 posting), the group could take many paths to that data rate including multiple parallel lines. Still, whew! Things are getting hot in interconnect valley! --rbm

Are you ready for 5GHz?


The PCI Express 2.0 spec could be officially complete by the end of the year, laying down a path for 5GHz chip-to-chip signaling in a broad range of computer, communication and consumer gear.

I have two questions for you about that: Is there demand for this speed? If so, where is it? And what are the signal integrity issues computer, comms and consumer engineers face at this data rate?

Please c'mon back with some lively and substantive postings. Or if you are more inclined email me at rbmerrit@cmp.com --rbm

Friday, December 01, 2006

Wireless USB goes upband

It almost missed the fact that the USB Implementers Forum announced this week plans to support ultrawideband at frequencies above 6GHz. This apparently is aimed at opening the door to using the technology on cellphones and consumer devices without fear of interfering with cellphone and other traffic in the current wireless USB space below 5GHz.

I know from reporting back at WinHEC this spring that cellphone makers in the Bluetooth Special Interest Group were saying they didn’t want to see Bluetooth running in the 3.1-4.8GHz UWB bands that may be used for 4G cellular. Folks at Cambridge Silicon Research were suggesting they had a way around the interference--perhaps using cognitive radio techniques--but I have heard no follow up on that.

In any case the wireless USB folks aim to support 6GHz+ starting in mid-2007 and expect products out in thos frequencies in early 2008. The first wireless USB products in the existing sub-5GHz range should be out shortly and be one of the hot items of CES in January. --rbm

Thursday, November 30, 2006

Super interconnects, Part 2

Although it lost its bid to build a new supercomputer under the DARPA HPCS program, Sun Microsystems will continue developing the novel Proximity interconnect that was one part of its proposal. Jim Mitchell, who oversees the Sun project, said systems using Proximity may appear before competitors get their HPCS systems out in 2010.

Proxmity uses capacitive coupling to link chips arranged in a checkerboard-like module at speeds of as little as 2 nanosceonds. The big hurdle right now is finding an accurate yet low cost way to align the capacitive pads on the chips so the interconnect can hit its promised speed, Mitchell said.

Tuesday, November 28, 2006

Super interconnects

IBM told me yesterday its Power7 processor to be used in a 2010 supercomputer commissioned for the DARPA HPCS program will integrate interconnect on the CPU—-but they won't say which interconnect, yet.

Separately, Jack Dongarra from the University of Tennessee tells me the new benchmarks for supercomputer productivity can be used to isolate and measure the impact of good interconnects. He looked at big systems that used the same x86 CPUs but different interconnects and found Cray's SeaStar provided tons of system bandwidth not available on systems using Gbit Ethernet as an interconnect.

No one has measured the relative pluses and minuses of Infiniband, Myrinet and Quadrics yet. But there are benchmarks for some 137 systems online now, so the analysis is waiting to be done. Please do it, and come back to me here with a posting about the results! –rbm

Monday, November 27, 2006

Parallel I/O powers hard drive concept

In a sort of back-to-the-future proposal, engineers at ECC Technologies are shopping a new concept for hard disk drives based on parallel I/O that could be a substitute for RAID arrays.

So-called parallel-transfer Hard Disk Drives (pHDDs) would contain multiple, single platter head disk assemblies with one or two read-write heads per platter. They would operate in parallel, speeding up I/O and eliminating the need for separate RAID arrays. The Web site TechWorld has posted a summary of a white paper on the concept. The full paper is available to those who go through a free registration process.

ECC Technologies'agenda here is that it owns US Patent Number 5,754,563 that it claims is key to making such drives. It even makes an open solicitation to drive giant Seagate to buy the company to get a lock on the concept. I'd love to hear any comment from HDD engineers on this product idea, so please post away. –rbm

Wednesday, November 22, 2006

A peek into peta projects


Cray Inc. and IBM Corp. will split nearly half a billion dollars as part of a Defense Advanced Research Projects Agency contract announced Tuesday afternoon (Nov. 21) to fund development of petaflop-class supercomputers before the end of 2010. Sun Microsystems was dropped from the High Productivity Computing Systems program (HPCS) that aims to foster work on computers that are more powerful and easier to program than any in current operation.
It's unclear what impact the loss may have for Sun which has struggled since 2000 to be profitable and competed aggressively for the contract. Specifically, I wonder what will happen to the novel capacitive coupling chip-to-chip interconnect called Proximity as well as a high-end parallel programming language called Fortress that Sun proposed to HPCS.

I am also hoping to get fresh details about IBM's proposal which it has kept hush-hush to date. A DARPA release just said it is based on a Power7 microprocessor, IBM's AIX operating system and General Parallel File System.

Cray has been the most candid of the three about the Cascade system it proposed to DARPA planners. Cascade is essentially a cluster-in-a-box that will deliver a mix of scalar, FPGA and hybrid vector/massively multi-threaded processor boards based on future versions of Cray's current separate product lines that will be integrated into in a single hybrid system. --rbm

Thursday, November 16, 2006

White paper: HT beats PCIe in latency

The HyperTransport Consortium recently posted a white paper documenting the latency advantage of HT over PCI Express for communications systems. The paper "presents a couple of usage scenarios involving read accesses and derives the latency performance of each," according to author Brian Holden who chairs the technical working group of the HT Consortium.


The paper measures latency for short packet reads on HT at 147-168 nanoseconds compared to 240-273 for PCIe. For long packet reads HT weighs in at 576-630ns compared to 885-1,008ns for PCIe.

The paper may be timed to exploit the weakness of PCIe in comms now that the related Advanced Switching Interconnect has gone belly up (see my posting of Aug. 23-24). ASI was supposed to bring comms-friendly features to PCIe, such as support for multiple host CPUs.

Despite HT's technical edge in latency, my sense is big comms OEMs like Cisco are moving away from the parallel HT and toward PCIe in their designs because it is more broadly supported. (Intel's PC volumes far outstrip those of HT backer AMD). And PCIe performance may not be the very best, but it is good enough.

I'd love to hear any other opinions from the comms world, so post away! –rbm

Wednesday, November 15, 2006

Do-it-yourself home automation


Toss another protocol stack on the smolder pyre of home networks. Startup Threshold Corp. (Petaluma, Calif.) announced this week it is making its homegrown One-Net wireless home automation software available for free licensing.

The small company decided it needed more range, battery life and security than Insteon, Zigbee and others offered, and didn't want to pay the few thousand dollars some of those technologies require for a developers kit or alliance membership. So it did its own thing which can ride on top of any of seven different 868 or 915 MHz band transceivers from Analog Devices, Texas Instruments, Semtech, RF Monolithics, Micrel and Integration Associates.

The resulting control net can deliver 38.4 to 230.4 Kbits/second over 100 meters indoors for as little as $2-3 a node, says Threshold CEO James Martin. The startup will ship a Wi-Fi access point and about half a dozen control peripherals for One-Net before June.

My guess is there will be very few people who choose to ride this option given its low profile and small, one-company backing…but you never know. A lot of people love to roll their own! –rbm

Tuesday, November 14, 2006

Infiniband flying high in supers…

The latest ranking of the Top 500 supercomputers out yesterday shows Infiniband is gaining momentum on Gigabit Ethernet as a clustering interconnect in the world's most powerful systems.

Gbit Ethernet is still the most widely used clustering interconnect appearing in 211 systems on the list, but that's down from its use in 256 supercomputers just six month ago. Meanwhile InfiniBand was used in 78 systems more than double the 36 systems that used the link in the list six months ago. The proprietary Myrinet interconnect from Myricom came in a narrow second with 79 systems, but that was down from use in 87 systems six months ago.

Erich Strohmaier, a researcher and one of the authors of the list, said he expects IB to continue to make gains. Gbit Ethernet cannot keep up with the performance needs of the latest systems and processors he said. Click here for more details.

…but just another bird in the flock

The outlook for Infiniband is more mixed beyond the rarified world of supercomputers that demand top performance. On Monday, Mellanox Technologies, the last remaining IB merchant chip supplier, rolled out a hybrid architecture supporting IB and Ethernet. ConnectX even has limited support in software for Fibre Channel. For the full story click here.

The message is that the mainstream data center will use multiple networks. Convergence will come in silicon embracing those nets.

Myricom has already started embracing Ethernet. Following a similar trend, Fibre Channel leader QLogic acquired Infiniband switch maker SilverStorm Technologies in October. It had acquired earlier this year Infiniband card maker PathScale.

In this mixed world, Infiniband has a future as the highest performance interconnect at the high end and one of the many nets with a role to play in the mainstream.

A couple interesting data points:

--The Infiniband chip business became profitable for Mellanox last year for the first time since the company was founded in 1999 with profits of $3 million on sales of $42 million in 2005, according to S1 documents filed with the SEC in September.

--To keep its edge, Mellanox will demonstrate 40 Gbit/s Infiniband links with one microsecond latency over the ConnectX chips early next year. It is developing 10Gb/s serdes blocks designed to enable up Infiniband links at up to 120Gb/s in the more distant future. --rbm

Sunday, November 12, 2006

Startup demos 10Gbit/s chip-to-chip innovation


Startup Banpil Photonics, Inc. said today it has demonstrated a way to send 10Gbits/second signals 1.5 meters across a standard pc-board without using power-hogging pre-emphasis or equalization to amplify the signal. Its secret sauce appears to be a technique to dig a passage through the board and send the signals as radio waves.

A February patent application by Banpil chief executive Achyut Dutta describes a method for creating an open trench or slot as part of a single- or multi-layer dielectric system to reduce microwave signal loss.

A "large portion of the signal (electromagnetic wave) is allowed to pass through the air or dielectric material having the dielectric loss less than the base dielectric material itself," the application states. "The trench or back-slot of the dielectric system can be filled with air or kept in vacuum. Alternatively the trench…can be filled with the liquid crystal material, which can tune the dielectric constant and loss," said the patent application.

Banpil has also cobbled together some proprietary tools. Banpil has created a unique RF simulation, layout and modeling tool chain based on two tools from unnamed EDA vendors and a tool designed in house.

"To let others do these designs, we will need to get a Cadence or Mentor to make these tools," said Dutta, who has started talks with two unnamed companies about creating such tools. --rbm

Thursday, November 09, 2006

Voices from the data center, Part 1

Thanks to Mike Krause, an I/O specialist in Hewlett-Packard's x86 server group, for his thoughts about data center convergence. Mike was one of the early proponents of iWarp, a version of Ethernet accelerated with Infiniband-like constructs such as remote direct memory access. Here are some highlights of what he shared with me this week:

On being agnostic: "Convergence is the buzz word for many people, but buzz can take years to translate into real production environments, and customers are a skittish crowd in the enterprise. Even Cisco has taken on the message of being interconnect independent, a message I've pushed for HP for many years now--build what the customer wants and needs rather than what the vendor wants it to be."

Infiniband as the converged fabric: "The IB community [is] in a bit of flux without a credible major company betting that IB is the converged fabric of choice for the enterprise. Certainly HP and Sun have positioned IB as primarily a High Performance Computing technology for the past few years. Will IBM enter in earnest the IB fray? With companies such as Sun and Intel being anti protocol off-load and long-term integrating Ethernet into processors, the IB vendors have to be worried that lock out can occur at any time."

Infiniband in storage: "There are no announced IB native storage [products] for companies such as EMC, IBM, HP, etc. This represents a significant hole in the [IB] convergence argument."

IP over Infiniband: "Micro-benchmarks yield only 2-3 Gbits/s out of the less than 8Gbits/s max possible. Skip past the marketing hype that IB is a 10 Gbits/s link since that is not the equivalent of a 10 GbE where one refers to raw signaling and application payload."

Fibre Channel "is in a bit of a long-term vision quagmire… it is not clear that the industry will invest to go beyond 8 Gbits/s to the 16 Gbits/s that is on many road maps. The optics vendors may balk at a straight 16 Gbits/s solution. Instead, they may want to de-rate a 10 Gbit/s WDM solution to 8 Gbits/s and then just scale up in multiples of 8."

iSCSI: "…not everything in iSCSI has been fully implemented."

Note: I'd love to hear other leading voices from the OEM community speak out. You can contact me at rbmerrit@cmp.com

Tuesday, November 07, 2006

Home nets play the numbers game


What this industry needs is more unification of home networks and their quality of service techniques. What we are getting is more and more of the numbers game as each camp tries to out-position the others.

Today the phoneline camp announced they have leapfrogged coax and powerline competitors. The Home Phoneline Networking Alliance upgraded its spec to support data rates up to a combined 320 Mbits/second over two simultaneous channels. That tops coax at about 135 Mbits/s and HomePlugAV at 180 Mbits/s. CopperGate is sampling chips based on the HomePNA version 3.1.

All interesting numbers, but the number this sector really needs to see is one—one way to let traffic flow across various home nets. We will have to watch plenty of market brawls before we get to that point. –rbm

Monday, November 06, 2006

Data centers converge in silicon

The real convergence in data center networking will not be on Infiniband or even Ethernet. It will be on silicon.

Thanks to Moore's Law the many different interconnects that have been vying to be the sole conduit of the data center will all live together someday in hybrid silicon. Voltaire shows us the way today with its announcement that its Infiniband switches will now support 10 Gbit/second Ethernet too, thanks to a new hybrid Infiniband/10GE ASIC. http://www.voltaire.com/

Not long ago, Myrinet made a similar move embracing Ethernet on its proprietary clustering products. And, though it's not at the silicon level, Qlogic has been reaching beyond its mainstay Fibre Channel business to embrace Infiniband products with its acquisition of switch maker SilverStorm Technologies in October and card maker PathScale in February.

Remember when Cisco's Andiamo unit came out with its storage switches? Turned out they were not so much based around Ethernet as Fibre Channel, the rooster of the SAN roost. And guess who is the largest customer for the Infiniband chips from Mellanox Technologies? Cisco, thanks to its acquisition of Topspin in 2005.

Seems like the big players in the data center are realizing you have to have it all—Fibre Channel, Infiniband and Ethernet. And ideally, you want to get it all in a chip. –rbm

Thursday, November 02, 2006

Cisco flows along with Interlaken


Not only does Cisco Systems continue to be one of the few big OEMs still developing a lot of ASICs, it continues to aggresively push use of its own chip-to-chip interconnect on most of them.

Known internally as "Spaui" because it is a mix of SPI-4 and Xaui, the interface will appear on at least 12 of the about 15 ASICs in the works at Cisco's storage networking group. Cisco co-developed the interface with Cortina Systems and announced it under the name Interlaken in April 2006. Named after the Swiss land between two lakes (pictured), the interface defines links running up to 20 Gbits/second.

The "Spaui" link is serving Cisco well, according to Tom Edsall, a senior vice president of Cisco's Datacenter Business Unit and chief architect of Cisco's MDS 9000 storage switch.

When last contacted, the Network Processing Forum (now part of the Optical Internetworking Forum) was working on its own SPI upgrade called the Scalable-SPI spec which would be an alternative to Interlaken. "I would have to see a significant advantage to make a change," said Edsall.

When you are cranking out 15 ASICs at a time for your own proprietary systems, you can afford to go your own way. –rbm

Tuesday, October 31, 2006

Inside the first 60 GHz CMOS radios


Here's a sneak peak under the hood of what the first commercial 60 GHz radios aimed for consumer apps might look like, based on a conversation yesterday with Bob Brodersen, a pioneer in CMOS radio and chairman of startup SiBeam, Inc. :

Sampling probably in late 2007 will be a device capable of sending uncompressed high def video between a set-top box and a TV at data rates up to 5 Gbits/second. It will be packaged in a roughly one-inch square module designed to accommodate up to 30 adaptive directional antennas. It could draw up to 5W and cost a slight premium over components for a wired HDMI link. Some of the early research chips from UC Berkeley (pictured) used waveguides on silicon to keep the millimeter signals on track.

The value proposition: It could provide lower cost and higher quality than using ultrawideband. That’s because a 200-400 Mbit/s UWB link would require an extra compression step meaning more codecs, more DRAM, more data loss and latency.

Detractors say it will take two years before there are multiple sources of 60 GHz radios with demonstrated interoperability. Backers say they are designing the most integrated and lowest cost phased array devices anyone has ever attempted in CMOS.

Will this be the multimedia home networking nirvana that supplants 1394, ultrawideband and 802.11n? Stay tuned. --rbm

60 GHz crowd pushes into the living room

Move over ultrawideband, six top consumer electronics companies are defining a 60GHz radio technology to bring high definition media to the home network.

LG, Matsushita, NEC, Samsung, Sony and Toshiba announced today they have formed WirelessHD. The group will deliver early next year a specification for multi-Gbit/second home networking at a range up to 10 meters and with latencies of 5-15ms. WirelessHD.

Does this mean the PC-centric wireless USB will get squeezed between the mobile Bluetooth juggernaut (see below) and this high-end consumer play? I aim to speak with some key principals and competitors later today about this and other questions. Stay tuned.-–rbm.

Monday, October 30, 2006

We will all look like this


That’s my conclusion from the aggressive prediction posted by IMS Research today that Bluetooth could ship in as many as 500 million systems this year and as many as 1.5 billion by 2010. The vast majority of those links will be used to connect cellphones with wireless headsets, IMS said, giving us that clipped ear look made famous in Star Trek.

The big question is whether this juggernaut will make Bluetooth the PAN of choice in a highly fragmented field of short-range interconnects I have discussed below. Alternatively, PCs, printers and digital cameras may define the PAN with the much faster wireless USB chips making their way into the market this year.

Check out the full story on eet.com and let me know what you think, and what you are designing in. --rbm

Friday, October 27, 2006

Fibre Channel slow amid 4Gbit ramp

In the end, all the talk about data center convergence really amounts to a divergence and fragmentation of networks and interconnects.

For examples of this fact, look at yesterday's lackluster quarterly results from Fibre Channel leader Emulex. The company announced two new divisions to diversify into SATA, SAS, iSCSI and Ethernet because--even amid a fast transition to 4Gbit FC--the FC market is fairly stagnant and plenty of other technologies are coming on strong. Re-positioning itself for the more diverse world, Emulex is sprucing up its logo as well as its product portfolio and R&D.

Look for a full story shortly at eet.com today. --rbm

Wednesday, October 25, 2006

AMD fuses graphics and x86


Upon formally completing its merger with ATI this morning, AMD announced its Fusion program, a plan to deliver starting in late 2008 a range of combined CPU and graphics processors for everything from consumer systems for emerging markets to desktops, laptops and servers. An AMD release suggested the company will tap both its proprietary coherent HyperTransport and PCI Express to link the chips to co-processors and other devices.

Intel tried and failed to merge the x86 and graphics for low cost consumer PCs with its Timna years ago. Via had minimal success with its integrated x86 CPUs.

AMD appears to have broader market targets with Fusion than either rival. But all those targets--high end gaming, low end consumer and technical computing-- are pretty much niche markets in computing.

So my question, dear blog commentors, is can AMD make money on welding together graphics and x86 multi-core processors? I'd love to hear your opinions backed up with lively detail! --rbm

Tuesday, October 24, 2006

HT confab needs all agendas

The annual HyperTransport Technology Developers Conference kicks off Thursday, October 26, 2006, at the Santa Clara Marriott Hotel in Santa Clara, Calif. My question: can this event embrace both the association's agenda and the HT agenda of Advanced Micro Devices for OEMs and developers who want to know about both.


The association said the event will include presentations by member companies focused on a variety of HyperTransport 3.0 and HTX. No word on AMD's separate Torrenza program to license coherent HT for linking to co-processors and more. Hmmmmm?

What do you want to hear about at the HT confab? Where does HT need to go in your view, dear posters?

I am away in New York this week, so I would love comments from those of you who attend about what you liked, what you learned and what you felt was missing.--rbm

Sunday, October 22, 2006

Startup keeps IB inside the box


Startup Panta Systems officially debuts its high performance server today, flexing a muscular I/O system with many as six 10Gbit/second Infiniband ports on every four-socket Opteron card. (Go to eet.com and search on “Panta” for the story.) Panta is a big believer in IB, even extending it down to its modular in-chassis storage array at 800Mbits/s.

But note that this startup did not want to take the risk that users are ready to run IB around their data centers. The Infiniband goodies stay inside the box and Ethernet is the external conduit. Interesting reality check on so-called data center convergence. --rbm

Friday, October 20, 2006

A burning question for the data center


OK, we have all heard that the data center is going to converge on accelerated Ethernet with Remote Direct Memory Access or on Infiniband that already has RDMA. Clearly, this has not happened and still seems a far away dream given the relatively slow uptake and expense of both technologies at 10Gbits and up.

The real convergence we have seen so far is in a single software stack for both networks under the Open Fabrics Alliance. So--and this is your weekend homework, dear interconnect mavens--what do you suppose the next practical step in this data center convergence will be?

Think on it and keep checking back here for the answer coming in less than a month! --rbm

Thursday, October 19, 2006

Innovate? Great, tell me how

Does your company have a defined process for fostering innovation? If so, let me know about it by posting a comment here or emailing me at rbmerrit@cmp.com.


The fact is, most electronics companies pay lip service to innovation, but few have any defined process for making sure it happens inside their organizations. That’s the view of Curtis Carlson (pictured), president and chief executive of SRI International who wrote a book on the topic and spoke about it at a San Jose conference this week.

Carlson makes a simple, powerful point that left my jaw hanging. What could be more important to a technology company—especially one here in the seat of Silicon Valley, than having a process designed to foster and capture innovation?

OK, let’s prove Carlson wrong. Tell me about the process at your shop. –rbm

Tuesday, October 17, 2006

Joy takes a jab at computer industry


“We have a strange situation where we are delivering a new kind of computing. It’s based on $50 million worth of systems that fill whole rooms, but there are no vendors that offer these systems. Companies like Google have to build it by themselves.”

That was one of the observations of Bill Joy, former chief scientist of Sun Microsystems, speaking at the Emerging Ventures conference in San Jose Tuesday.

Joy now focuses on semiconductors and green tech as a venture capitalist at Kleiner Perkins Caufield and Byers. At the conference, he was asked to share his reflections on computing today and took issue with the high-performance computing sector.

“We are scotch taping together suboptimal machines, and its not acceptable,” Joy said. “Unix is not designed to be an OS when there are 100,000 copies. CPUs and north bridge chips are not designed for that either.

“It suggests the industry needs a fix, but I don’t’ know what it is. There is an opportunity for some great company there,” Joy said.

The fact that a handful of companies such as Google have built their own large parallel systems is a huge advantage, but “it’s something most people can’t do,” he added.


The cerebral Joy probably wasn't aware his former employer announced Project Blackbox the day before. The data-center-in-a-box effort (pictured) is aimed to help hyper-growth Web companies gear up fast. Such drastic steps appear more aimed to garner press than truly solve the problem Joy highlighted.

Separately, I bumped into Sun architect Marc Tremblay during a morning break. He said Sun’s Rock processor is on target to tapeout by the end of the year. Sun has some new multi-core tools it will roll out in tandem with Rock, he added. Stay tuned! –rbm

"EtherRAM" speeds data center I/O

Startup Gear6 emerges from stealth mode today [Oct 16] with a novel concept for boosting performance in the data center with rackable RAM cache systems that sit on an Ethernet network. The CacheFx appliance creates a scalable pool of shared semiconductor memory that fills a gap between an individual server’s main memory and the data center’s storage-area network. Future products will sit on Fibre Channel and iSCSI networks.

Even over an Ethernet link, access time for the cache appliance is 10 to 50 times faster than getting data from a storage network, the company claims.

A principal engineer from Google shared the stage at the Intel Developer Forum in September with Intel Corp. CTO Justin Rattner, noting the need for something to help fill that performance gap.

The startup’s secret sauce is in its software that makes the systems easy to use. Paving a road to what could be an eventual acquisition, Gear6 has already forged partnerships with Sun Microsystems and Network Appliance to ensure their systems interoperate with the CacheFx boxes. --rbm

Monday, October 16, 2006

The next hard(ware) step for DRM


Today’s digital rights management systems are still fairly crude, according to Brad Hunt, chief technology officer of the Motion Picture Association of America, speaking at last week’s Digital Home Developers Conference in San Jose. That’s because they protect premium content by managing whether or not you can make a copy of it. Tomorrow’s DRMs will be smarter, letting users register the set of devices they own that live on their home network so that paid-for content can be shared among them.

Hunt and others rallied around so-called link protection technologies such as DTCP-IP and Windows Media DRM for Network Devices that protect and manage copies today via software. Upcoming DRM interfaces like the Secure Video Processor work could open up the door to sharing content on across a user’s devices thanks to hardware-backed techniques.

“We are just starting to see chip sets for this emerge,” said Hunt. “We want to move from copy management to license management. We think that’s the future of content distribution,” he added. --rbm

Thursday, October 12, 2006

SI evangelist on a mission

Not sure about using VHDL-AMS? Bill Hargin, a hard charging Mentor Graphics signal integrity marketing manager, may have you in his sights.

I had a latte and chat with Bill (pictured) on his latest pass through San Jose. He has been selling big chip makers such as Intel and Texas Instruments as well as small design and training companies about the advantages of VHDL-AMS which is at the heart of Mentor’s SI tools. If he is successful in his effort to build an eco-system around the technology, it could help set it as a standard for the advanced IBIS modeling specification in the works.

“We just need to get more models out there,” Bill said.

The former mechanical engineer has a solid appreciation for today’s SI issues as well as a bent for metaphors. Here are two quick examples:

The problems in high-speed SI today are like those of a house with old plumbing: Turn a faucet off quickly and a “water hammer” effect rattles the pipes, just like reflections in a GHz transmission line, Bill says. I get the picture.

Sometimes the problems lead not just to metaphors but new products. Today’s serdes are much more complex than yesterday’s parallel interconnects because they have so many knobs and settings for effects such as pre-emphasis and equalization, Bill notes. It’s like the cockpit of a 747 compared to the driver’s seat of a Ford pickup.

So Mentor is developing a user interface to make it easy to fly those serdes when they are embedded in an Altera Stratix FPGA. The tool is about 80 percent ready for prime time and could be out in the next few months, Bill says. --rbm

Wednesday, October 11, 2006

OEMs craft their own storage CPUs


Brocade and Cisco have developed their own storage processors to handle virtualization on their Fibre Channel switches. At least that’s what I heard today from Mukund Chavan, a chief hardware architect at Emulex Corp., who considers the OEM in-house ASICs to be among his chief competitors. It’s not surprising, because years ago Cisco helped pioneer the field of network processors with some of its own in-house designs even though a small hoard of net CPU startups were begging for Cisco sockets.

Despite the OEM ASICs, Chavan believes there will still be a healthy market for merchant storage processors among the top tier storage array and switch vendors such as EMC, IBM and others. He detailed at the Fall Processor Forum today the architecture of his AV150 which began as a design from startup Aarohi Communications, acquired by Emulex in April.

If you know anything about these Brocade and Cisco chips please post a comment or drop me a line at rbmerrit@cmp.com. I may want to profile the chips and the design teams behind them on EE Times.

Wanted: platform power standards


I think the computer industry needs a more comprehensive set of power efficiency standards.

At the Fall Processor Forum yesterday both Intel and IBM talked about their recent adoption of microcontrollers to monitor and control the power, heat, frequency and voltage of a CPU. Intel uses the Platform Environment Control Interface to link to chips that handle some of those functions.

IBM sketched out the Power6, its first CPU use of an external management controller and called for a discussion about how to create an approach that would let users set power efficiency policies based on their workloads.

My impression is both companies are going down the right road in parallel. I think at some point they will realize, along with other competitors, there could be some valuable cooperation in setting a baseline of standards for this new power efficient era. --rbm

Monday, October 09, 2006

More PANdemonium


Lots of news bits about short-range wireless technologies flying around this morning.

The bad news is Aura Communications apparently closed its doors quietly back in May, although one source says its magnetic induction technology suitable for music streaming was acquired by someone. Dear blog readers, if you know more about this please tack up a comment or email me at rbmerrit@cmp.com

On the good news front, the IEEE announced today a draft standard for a Chirp Spread Spectrum (CSS) physical layer dubbed 802.15.4a. “The new standard targets applications such as Real Time Location Systems, industrial control, sensor networking and medical devices,” according to a release from Nanotron Technologies GmbH (Berlin) which backed the effort.

Nanotron makes pre-standard chips today that offer data rates of up to 2 Mbps at 10mW over 900 meters (outdoors) and 60 meters (indoors) without an external power amplifier. The final standard is expected to be published in March 2007, about the time compliant products hit.

It will be interesting to see how CSS competes with existing efforts like Bluetooth and Zigbee as well as last week’s new low-power PAN entrant from Nokia, Wibree (see posts below). In his Nearpoints blog this week, consultant Craig Mathias will say both Bluetooth and Wibree are geared “for infrequent, low-duty-cycle transfers, so power consumption alone doesn’t really make me want to run out and buy WiBree.”

Bluetooth backers will design dual-mode BT/Wibree chips as well as standalone Wibree chips for low cost devices like watches and toys, so some of the needed consolidation of PANs appears built into the concept. Thank God! –rbm

Friday, October 06, 2006

One more voice on Wibree

Analyst Craig Mathias of Farpoint Group pinged me from his travels in Europe to say the Wibreee low-power wireless technology announced by Nokia on Monday to say “I don’t think Wibree will have much impact in the general market.”

He says it is not directly competitive with Bluetooth which is mainly focused on headsets or Zigbee, aimed at industrial nets. While Wibree does sport a Mbit/second data rates at power levels supposedly a fraction of those of Bluetooth, Mathias suspects the Bluetooth community will respond with its own low-power options.

Maybe, or maybe Wibree backers such as Broadcom and CSR will just do as Nokia suggests by making Wibree a low-power model on their cellphone Bluetooth chips and making simple Wibree-only chips to attack a new opportunity in watches, toys and other cost-constrained gadgets Nokia hopes to link to with this new wireless technology. --rbm

Thursday, October 05, 2006

ARM wrestles for cellular security role


ARM wants to open up the applications programming interface for its TrustZone extensions to its ARM core in an effort to set a standard in cellphone handset security. The effort came to light yesterday at a panel discussion I chaired at the ARM Developers Conference in Santa Clara, Calif.

At a cocktail reception afterwards, ARM chief executive Warren East told me the company’s cores are in 90 percent of all cellphones and 70 percent of digital cameras. That makes the company with only a few hundred million in annual revenues a major force in mobile.

But it’s far from the only voice. The Trusted Computing Group recently finished a mobile security spec chaired by Nokia. It uses a version of the group’s Trusted Platform Module. Other groups addressing some aspects of mobile security include the Open Mobile Alliance, the Java Community Process, the Open Mobile Terminal Platform and STIP.

Reps from Broadcom and Motorola who attended a private ARM meeting on the API effort expressed hopes for one overarching mobile security standard. But a Freescale security architect took a wait-and-see attitude, as do I. –rbm

Wednesday, October 04, 2006

Out of the PAN and into the…?


Thanks and a hat tip to John Barr, a senior technologist tracking standards at Motorola, for chiming in about Nokia’s new Wibree initiative. It’s not really a personal area network, but more of a point-to-point link like wireless USB, and it lacks the rich profiles Bluetooth has worked for years to develop, notes Barr who is a strong advocate for Bluetooth.

Will Wibree make it into the PC? Have to wait and see. So far Intel declines to comment about the new technology—not a good sign for Nokia. –rbm

FPGAs make a bigger play in the x86

Interesting to watch how FPGA makers such as Altera and Xilinx are getting more cozy with the x86 computing world these days.

I had a chance to chat (remotely) with Xilinx CTO Ivo Bolsens who made a rare appearance at the Intel Developer Forum last week. Among others news bits, both Altera and Xilinx announced they had licensed Intel’s front-side bus (FSB) so they can build products linking to it.

Once upon a time, Intel kept an iron grip on its CPU bus technology, licensing only a very few chip set makers to keep the PC ecosystem thriving. But it’s a different day with Advanced Micro Devices putting on a full court press to get chipmakers hopping on its coherent HyperTransport (cHT) bus, although it sounds to me like clear terms for a cHT license are still being hammered out.

In the short term, expect to see more FPGAs in high performance computers for specialty applications. Cray has already given the green light to this idea for systems to ship in 2007, using FPGAs from DRC Computer Corp.

As the era of multi-core computing matures FPGAs and a host of other co-processors will become ever more strategic. Watch this space. --rbm

Tuesday, October 03, 2006

Another flash in the PAN


Nokia announced today a new short-range wireless technology called Wibree that is similar to Bluetooth but uses a fraction of the power. Nokia says it is aimed at connecting PCs and phones to watches, toys and other low-cost gadgets that run on button cells.

It’s getting crowded in the personal area network with Bluetooth, various flavors of ultrawideband including wireless USB pushed hard by Intel, near-field communications pushed hard by Philips, peer-to-peer and mesh variations of Wi-Fi and combinations like Bluetooth over UWB. Dell has been calling for a shakeout in the PAN for sometime now, calling the sector "a mess."

I have been saying since late 2004 there are too many radios trying to crowd into the phone from Bluetooth to RFID and Zigbee. It's ironic then that handset giant Nokia launches one more. The fact that Intel and Microsoft are not part of Wibree indicates it may be hard getting support for it built into the PC.

No doubt wireless is the long-term future of communications. And clearly the short term future is all about further fragmentation and consolidation of the technologies. Let the accordion play!

PS: You can go to eet.com and search for "PAN" to find some stories with good perspective on these issues--rbm

Monday, October 02, 2006

The Dear Abby of SI


Eric Bogatin is hanging out a virtual shingle at DoctorIsIn@BeTheSignal.com. The specialist in signal integrity training is taking your questions and posting answers at the bottom of his new monthly column on the Altera Web site. Eric will also use the Q&As in future versions of his columns for Printed Circuit Design and Manufacture Magazine. Those columns are now nicely indexed on his own Web site.

There are some decent how-to tips at both locations. The new Altera column talks about using guard traces to reduce noise by a factor of five in mixed signal applications where 100dB isolation is important.

As we move deeper into the GigaHertz Era were going to need all the help we can get. So I may link to a juicy tip from time to time. --rbm

Friday, September 29, 2006

Ethernet pulls into double diamond lane


Hold on to your computer mice. A new Ethernet high-speed study group kicked off this month. No targets or goals have been set yet but the expectation this IEEE group will have to wrestle with is whether it can define a spec for 100Gigs. Yowsa!

The group kicks off the same month the IEEE finally ratified a standard for 10Gbit Ethernet over copper. Some engineers are just recovering from the Herculean efforts that required including a major shift in cabling, signaling and on and on.

Hats off to John D’Ambrosia (pictured), the energetic and omnipresent IEEE standards maven from backplane Ethernet who was elected to chair the new group. He’ll need all the energy and mellow spirit he can muster to ride this wave.

“Everything is on the table at this point, and we do not see this as the last of the high speed Ethernet study groups” D’Ambrosia told me in an interview for a story you can find Monday at eet.com by searching for 100G.

Next big meeting is November in Dallas. Stay tuned! --rbm

Juicy tidbits from the data center

Hats off to StorageMojo’s Robin Harris for a comprehensive report with two juicy bits from last week’s Data Center Ventures conference by Dow Jones. The blog put Woven Systems on my radar screen, an early stage startup aiming to create Ethernet switches that would compete with Infiniband in latency. The other bit is rumor mill fodder that David Yen may have one foot out the door at Sun. Harris speculates Yen may not be happy trying to turn around an underperforming storage division at the still-troubled computer company. I would observe, however, David has been a loyal longtime exec that has done great things with the company’s chip and systems groups to date. –rbm

Thursday, September 28, 2006

How cozy will Emulex and Intel get?


The duo shared the job of designing a family of integrated storage controllers announced Wednesday for SAS, SATA and Fibre Channel. Emulex provided the FC software and know how and Intel provided its XScale cores, SAS and SATA capability.

My question now is, could this lead to a deeper union? Before Paul Otellini goes to the mic Oct. 17 for a quarterly earnings call to give the final report on the Intel reorg might we see the Intel storage group sold to Emulex or the Fibre Channel company bought by the X86 giant? Whadaya think? --rbm

Wednesday, September 27, 2006

The Empire strikes back

Intel has fired a shot back at AMD in the battle for the hearts and minds of third-party chip makers and OEMs.

News of proposed extensions to PCI Express called Geneseo are technically quite different that what AMD is doing with coherent HyperTransport (cHT). But functionally, they are a shot to AMD's heart.

Both efforts aim to link a coherent CPU with co-processors such as future high-end graphics chips and acclerators for XML, security and networking. But Geneseo is by no means a coherent version of Express. At best it will only provide a locking scheme and coherency hints to smooth performance. It also proposes fine grained power management support.

So what's new? AMD tries to do something the right way, and Intel tries to steal the wind from their sails with "good enough" technology. --rbm

Fighting the good fight in signal integrity

Best wishes to Todd Westerhoff, a high speed design group manager at Cisco for the past six years.Todd is moving over to the vendor side where he will continue to fight today’s signal integrity battles. As a manager at SiSoft Todd will ”help drive the future direction of their Quantum-SI product line,” he said in an email.

I gave Todd a call to wish him well and dig under the surface a bit more. Todd has been a leading light in the IBIS Advanced Tech Modeling group trying to set a standard for SI modeling with chip vendors. The group has a proposal from Cadence to use some new and proprietary technology and another proposal from Mentor to use Verilog-AMS.

”We are struggling with which semiconductor vendors support which approach and right now we don’t have any consensus,” Todd said.

Todd said another big problem in SI simulation today is serial links faster than 3Gbit/s have no open eye to measure. SiSoft has a product in the works to handle that issue, but they are not ready to discuss it yet. Stay tuned!

The other big SI issue today is in parallel link design where engineers need to conduct timing, SI and power analysis in same environment to evaluate interconnect delays at a given speed and thus make choices on physical design. What’s more, those issues have to be examined holistically for the chip die, package and board. And the issues are only getting worse as lower voltage chips emerge that naturally tolerate less noise and are thus harder to regulate in terms of power.

“Part of what I want to do [at SiSoft] is help make that clear,” Todd said,

Todd did cry foul on posters to my Sept. 8 blog who said full board-level SI simulation tools are too expensive with costs ranging from $150-200,000 per seat. Only one vendor charges anything near that range, and there are other vendors who charge from $70-150k for high-speed serial tools and $20-60 for parallel link tools, so shop around, he says.

Good luck, Todd, and stay in touch!--rbm

Monday, September 25, 2006

USB gets unwired in San Fran


This week’s Intel Developer Forum will be something of a coming out party for wireless USB with lots of demos of working prototype systems and a ready-to-go testing and logo program.

I may be breaking embargo by about 20 minutes here, but today (Tuesday, Sept. 26), the USB Implementers Forum will announced the availability of the Certified Wireless USB Compliance and Certification Program with test specs and procedures available online. Initially, Intel will host testing at one of its labs. Later, the group will announce plans for compliance workshops for 2007.

NXP, formerly Philips Semi, jumped the gun a bit, announcing a “Certified Wireless USB” controller Sept. 18. It will show at IDF today a prototype MP3 player using the controller as well as “some other [interesting] UWB stuff,” according to a comment a spokesman posted to my Sept. 14 blog.

Likewise, Alereon will demo tonight a Kodak camera transferring photos to a PC its wireless USB chip. Wonder what chip the camera uses and whether it’s a real product? I remember the Kodak techies sniffing around the Bluetooth pavilion at Comdex (anybody recall Comdex?) about five years ago, complaining that a Mbit was just not enough bandwidth for them.

In addition, WiQuest Communications today announces the availability of Windows Vista drivers in addiiton to its XP drivers for its UWB chips. So software, hardware and standards tests all appear to be ready to go with each company ready with some piece of the puzzle.

With HP, Dell and others gearing up for wireless USB, Freescale will have a hard time finding sockets for its incompatible version of ultrawideband. Who will want a device that uses the latest fast wireless link, but won’t work with their computer? –rbm

Sunday, September 24, 2006

Converging on the converged network


Several bits of news coming down today (Monday, Sept. 25) all point to the ongoing effort to craft a Swiss Army Knife of data center networking. I’ll update this posting with links to full stories on each item as soon as they become available on eetimes.com.

The Storage Bridge Bay alliance has released its 1.0 spec and announced IBM and Network Appliance have signed on. The latter partially answers one blogger’s criticism back in March when SBB debuted that the effort only has partners Dell and EMC driving it.

What’s key here is the Dell/EMC duo and their new partners have essentially put out to the industry a core spec for a mainstream modular storage array that can be adapted to many uses. That will lower OEM costs and time-to-market while providing some added ease of use. Time will tell by how much that will raise the heat on the big competitors still doing their own thing—Hewlett-Packard and Hitachi Data Systems.

And there’s plenty more shoes to fall. As I said in postings earlier this month both IBM and NEC have new proprietary products for modular storage arrays in the works.

Separately the drive to merge networking, clustering and storage on to Ethernet takes three small steps forward today. HP announced it will use the 10Gbit chip from NetXen. Chelsio rolled out its third-generation architecture and NetEffect said it will show at the Intel Developer Forum this week its chip running the Open Fabric Alliance software for accelerated Ethernet.

This accelerated Ethernet movement has needed some forward progress after a few years of slogging out specs and shifting from Gbit to 10Gbit as the market target. As Dwight Barron, an HP server technologist told me, 2007 will be the year when the rubber finally meets the road on converged Ethernet.

It is not without competition. Today the Infiniband Trade Association announces it is supporting iSCSI. IB is already getting significant traction in clustering, and now the push is on in storage. Ethernet backers say their 10Gbit parts will mow over reigning Fibre Channel which is at 4Gbit, and IB says look out for its 20Gbit parts just starting to hit the market. Their Achilles’ heel: The IB software is still a long way from being as mature as that of the competing interconnects.

The lesson? Convergence in an open market is a messy free-for-all. --rbm

ASI is dead, long live Express!

That’s the message from PLX Technology which rolls out a 48-lane PCI Express switch today (Monday, Sept.25), its new top-of-the-line product aimed at high-end servers, graphics cards and embedded systems where Express is gaining traction.


In the wake of several companies pulling the plug on Advanced Switching Interconnect products and the demise of StarGen, the last ASI silicon provider, top comms and embedded OEMs are now migrating from PCI to Express, said Akber Kazmi, a senior marketing manager with PLX.

“The major comms OEMs have no choice but to move to Express now,” Kazmi said.

So when if ever will we hear an official announcement of the demise of ASI? The ASI and StarGen Web sites are still mum, and the former ASI chair is not returning my calls. They were once so eager to talk. --rbm

Thursday, September 21, 2006

AMD clarifies what it’s NOT doing

Despite published reports based on a vague press release, Advanced Micro Devices is not making its Opteron socket open source and is not changing the licensing terms of its cache coherent version of HyperTransport.

Here’s what’s really happening, according to Doug O’Flaherty who manages the Torrenza project at AMD: AMD is close to finalizing the specification for its next-generation Opteron socket, expected to be used for chips that ship in the second half of 2007. The company has been collaborating with OEMs and partners under non-disclosure agreement on the socket design more early than usual in the design cycle, AMD claims.

The socket spec will be available when it is completed, just as AMD makes its existing socket data available for motherboard and other companies who need to design products for it. However AMD’s cache coherent version of HyperTransport will still remain proprietary, available to partners via licensing and financial terms the company has not disclosed. Between five to 20 companies have licensed the protocol so far, O’Flaherty said.

At this point, AMD is not willing to disclose any more information about what’s new in its 2007 Opteron socket or who is doing what with its cache coherent protocol. As usual in the business media, AMD has tried to generate excitement without providing much detail. --rbm

AMD hints at big plans for HyperTransport

This morning’s release from Advanced Micro Devices about licensing its cache coherent version of HyperTransport—now called Torrenza --is full of wishy-washy marketing blather and raises more questions than it answers. It appears AMD will “open source” the spec for a future version of the Opteron CPU socket as one step to licensing Torrenza.

In my Sept. 9 post below I read between the lines of IBM’s Roadrunner announcement to see Big Blue must be licensing Torrenza. AMD has been working on the technical and business details of how to do that since it announced in June its intention to open up the interconnect for third-party accelerators that plug directly into its CPU.

Today’s AMD press release sheds little light on that effort. It says Cray, Fujitsu Siemens Computers, HP, IBM, Dell and Sun Microsystems “intend to design products” for, or at least “plan to evaluate” Torrenza. Comments from CTOs of those companies followed but they were all non-specific.

The release was bursting with marketing blather about the Torrenza Innovation Socket and the Open Innovation Program. The release implied OEMs will actually be able to contribute to defining a Torrenza socket, but provided no hard details on who can get what when on what terms. I’m seeking an interview to find out exactly what is happening so I can report it in specific English. Stay tuned.

Adding to the fog, a Wall Streeter’s report in my email box says “AMD's CEO Hector Ruiz [said] yesterday at a dinner in San Francisco…Apple may eventually use AMD processors in its products ‘at some point in the future’…Hector did not suggest that there were any talks currently between the companies.” Thanks for clearing that up! --rbm

 
interconnects