You can imagine my pleasant surprise when I was invited back this year as a panelist. I’ll be participating in a panel debating flexible vs. fixed silicon solutions in the network (agenda here).

Presenting companies that fall under the Nyquist market model: LNOP, ANAD, QCOM, Luxtera, INFN, EQNX, and several more TBD. More information including registration can be found here.

The conference is being held in upstate New York this fall and always draws a very interesting speaker list – last year Steve Forbes, Michael Milken, Carver Mead and John Rutledge all delivered stellar presentations. I went in last year with a slightly negative bias of Gilder and the ‘Gilder Priesthood’ and came away with a positive impression. You might change your mind too. Hope to see you there.

Great controversy surrounds the company and investors wonder whether the valuation attached to company is justifiable. The valuation is indeed debatable and is predicated on their success penetrating the PTTs & Bellcos.

However, I feel this is a real company with a significant competitive advantage derived from their early attempt to use integrated photonics as a disruptive technology. Their key differentiation was dropping the cost of optical to electrical conversion (see “Death of the All Optical Network“) and building a system that leveraged this strength.

What is most interesting about Infinera is that they are, at their core, an optical component company. The company was founded on the concept of making a highly integrated optical transponder and then building an entire hardware/software system and carrier marketing and sales team around it.

As an optical component company, they would be throwing their lot in with JDSU (JDSU), Bookham (BKHM), Avanex (AVNX) and other companies inhabiting the arid plain of unprofitability that is the optical component business (see “Bookham, China, and the Optical Component Market“). Instead, Infinera made the component the competitive advantage, and successfully raised the $315M in venture capital needed to build an optical transport equipment company. Amazingly, this was done in a period of time when the optical sector was kryptonite to most investors.

As a result, Infinera is currently in a better position than any other transport company to deliver low-cost, high-density optical transport. Their Photonic Integrated Circuit (PIC) allows them to offer ridiculously high port densities of 100G per slot.

The 100G number is important as it gives them architectural flexibility to meet both the emerging 100G Ethernet standard or the 40G SONET/SDH standard (or a derated 100G Ethernet specification). It also lets them deliver 10G in great quantities and theoretically, low cost. These low-cost 10G links can be bundled together using inverse multiplexing to form 40G and 100G links that don’t require new bleeding edge optical components.

Conversely, most of the incumbent optical transport companies like Alcatel/Lucent (ALU), Nortel (NT), and Siemens have focused on more incremental improvements during the post-bubble years.  To make matters worse, these companies, with the exception of Huawei and ZTE, have outsourced their optical component R&D to their suppliers, leaving them with no way to differentiate themselves at the optical layer (see “Dr. Strangelove, Or: How I Learned to Stop Worrying and Love Huawei“. This has left the former titans of optical transport defenseless.

I remember a meeting with Infinera (Actually called Zepton before a savvy marketing guy changed the name of the company) way, way back where Drew Perkins laid out an array of optical modules on the conference room table. He pointed at each one, and said that they were way too big with some very colorful language. He indicated that they wanted to build a 100G linecard- this was at a time when 10G was considered state of the art. These modules wouldn’t get the job done so they were looking at ‘something else’. Chip companies asked to build 100G electronic components to feed the ‘something else’ snickered at the lunacy of it all.

If Infinera is successful, it will be a classic example of the Innovators Dilemma.

I predict this IPO will drive interest in PICs and several component vendors will attempt to piggyback on the market success of Infinera. But I expect that this new technology will not change their fortunes.

Why? Infinera built a defensible business and extracted the value from an innovative optical component. These standalone component companies will continue to slug it out in an oversupplied commodity market. Once PLCs are commoditized, Infinera will still have a differentiated system that leveraged the density strengths of the PLC. And other equipment vendors will be playing catch up.

Congratulations Infinera.

Author owns no position in the companies mentioned

The Wall St. Journal editorial yesterday captures the true issue at hand, and why the FCC is now deadlocked on approving the merger.

Meanwhile, Congressmen Ed Markey and John Dingell deserve their place in this story, having bared their teeth at Mr. McDowell in a threatening letter last week that questioned Mr. McDowell’s suitability to vet the merger. Their true interest was in strengthening the hand of Democratic Commissioners Copps and Adelstein, who seem bent on using this merger review to advance “Net neutrality” mandates that neither Congress nor the FCC have seen fit to impose through normal channels.

We’d like to see these two Congressmen demonstrate the same concern for public integrity by calling on the Commission to quickly approve a merger that everyone from shareholders to the relevant unions supports. But that probably makes us “uncurable optimists” too. The duo are eager to do favors for the campaign check-writers at Google and MoveOn.org, which want Net neutrality rules so AT&T and other telecom companies can’t charge market prices for use of their broadband pipes.

Check out the latest video from the “Save the Internet” coalition.

The only thing worse than having Washington involved in free markets is when they are trying to solve a problem that DOES NOT EXIST. My extended opinions can be found here.

I’m paraphrasing the article – read it in full here. It’s more serious, I’m just grabbing the funny bits.

How you know the bubble is back:

• Telecosm 2: Return of the Light is published
• Dave Huber starts another optical company
• You are at a cocktail party and two bankers start a debate between OOO and OEO (my personal favorite)
• CEOs/CFO’s are once again seen on CNBC talking physics and playing with toy mirrors
• Chambers Announces Cisco never left the optical business
• Average people want to hear about crystal growth yields
• Telecom people travel to Tulsa again, still no one knows why
• A California startup announces an ASIC that integrates the equivalent of 1000 Cerent 15454s on a single chip. This starts an “454 integration war” between several semi companies
• SuperComm Returns to New Orleans and calls itself SuperComm and acts like the last ten years are a forgotten memory

I know two American engineers who have relocated to China to lead optical module design teams at Chinese equipment companies. They live and work in China for Chinese companies, using their skills to build custom modules – skills no longer in demand from their American Tier-1 telecom equipment employers.

Huawei and ZTE are more than happy to make use of their knowledge. Unlike most big Optical Equipment suppliers, these Chinese companies have gone counter to the trend of buying off the shelf optical modules. Instead, they continue to homebrew their own optics for the most demanding applications.

Huawei and ZTE understand how to gain a competitive advantage in high end Optical Transport; you cannot outsource a core competency to Bookham (BKHM), Avanex (AVNX), or Opnext (OPXT).

One of the more interesting speakers at the Gilder Conference was John Rutledge – I blogged his talk here. He wrote a short summary of his impressions of China gleaned from his latest visit. He summarizes the threat China presents to the USA.

The big words in China are entrepreneurship and innovation. …The Chinese government knows they will not be able to continue growth through manufacturing. There is not enough oil, gas, and coal and the air and water quality is terrible. They have decided to grow by investing in IT (fiber optic communications networks) and human capital (education) to increase productivity. Fighting over textiles is yesterday’s war. We should be thinking about where Cisco puts their next R&D facility. China is currently bidding very hard for such operations to relocate to their country.

They may not need to bid, as many readers already know. Huawei has vaulted to #2 supplier of optical networking hardware in the world with a 9.5% share without even selling large quantities to tier-1 carriers.

I’ve been negative on Ciena (CIEN) for a long time because I believe their metro transport products are the most vulnerable to overseas commoditization. I believe that the only optical transport equipment companies that will profit and thrive in the long term are those that attempt to develop an internal component edge through R&D, like Infinera, or are big enough to sell an entire platform of products and perform the system integration – like Alcatel (ALA). Huawei is aiming to do both.

Huawei has already taken a dominant position in DSLAMs (#2) and Optical Transport (#2). So, what’s next?

I think the answer is easy - Layer 2/3 switching for the Enterprise. Cisco (CSCO) has 70% global market share in this area and derives at least 50% of their operating income (my calculations, includes optical module resale) from this business. Competition has effectively thrown in the towel with the exception of (surprise!) the Huawei / 3Com (COMS) joint venture. Huawei/3Com has secured 35% market share in China and has grown revenue 70% annually since 2004.

Important Footnote: I have expressed reservations about the accuracy of Huawei’s numbers (see Huawei 2005 Revenue).

Full Disclosure: I am short Cisco as a hedge against other positions. Long COMS

Unlike standard Wall St. conferences where portfolio companies are trotted in to give their 30 min pitch and allowed to retreat for limited questioning, companies that present at Gilder are subjected to the full force of criticism of the moderator, their peers, and as a last resort, smartasses in the audience like myself. The folks asking the questions are not analysts, but industry insiders best positioned to ask pointed questions. This creates a far better environment for evaluating companies and forming/altering your opinions on industries.

This process is by far the biggest competitive advantage the conference has. It may be too effective, and actually keep some companies away who prefer a more controlled environment.

I was not expecting such technical depth and breadth among some attendees. It was refreshing to go to a conference where everyone is not a tech drone. It’s rare you can go from talking about the requirements for VLAN tag stacking to discussing the need for consumer medical choice at a conference.

Finally, I must admit I personally enjoyed the political theme that runs under and through the conference. The message resonated with me and made it a personally enjoyable experience, though your mileage may vary. Where else can you see Michael Milken, Carver Mead, Steve Forbes, John Rutledge combined with senior management from leading optical component companies?

I’ve resolved to keep an open mind henceforth. I’ve also signed up for the newsletter. We’ll see how it goes.

During their session at Telecosm 2006, EZ-Chip (EZCH) and Raza Microelectronics read from the same script NPU companies have been using since 1998.

1. The network is filled with changing requirements, and therefore you need a flexible hardware solution.
2. There is this new market coming that will be the turning point (this changes about every 1-2 years, current flavor of the month is Metro Ethernet, last year it was access networks)
3. NPUs put companies on the same footing as a Cisco, allowing networking hardware to be commoditized. When this takes place the NPU will be to networking what the CPU is to computing.

Reasons 1 and 2 are linked. Whenever there is a new market (and quantities are low) there is a high degree of uncertainty in what features are required. If you speak with carriers today about Metro Ethernet requirements it is clear they have no idea what functions are needed. Ask them if they prefer A or B and they will answer “yes”. Ergo, an NPU is a great fit because it provides flexibility.

The problem arises once the market takes shape. The equipment provider now understands what features are needed, which are not, and which are niche. Once these requirements are well understood they can and will be codified in silicon.

When VLANs were first implemented by Xylan in the 90′s it was done in custom ASICs. Eventually, those features were cooked right into Broadcom (BRCM), Marvel (MRVL), Vitesse (VTSS.PK), and Agere (AGR) L2 switches that sell for $1 a Gigabit port today.

If something is required to be flexible and heterogeneous it is hard to see how it is adopted as a widely accepted networking standard. Unlike computing, networking hardware converges to a common set of features due to strict interoperability requirements. In Enterprise Networking, the Venn diagram of flexibility and high volume is a null set.

I see no reason why Metro Ethernet is any different than Enterprise Ethernet. Companies like Transwitch (TXCC), Zarlink (ZL) and Cortina Systems are already building dedicated silicon for Metro Ethernet.

Another barrier to the widespread adoption of NPU’s is that no common instruction set exists among all vendors. Given the frequency that NPU companies have failed there is great hesitation on the part of equipment vendors to put an single-sourced NPU in the critical path of product success. I’ve discussed this effect in depth. (see Wintegra – Following Up on a Comment)

NPU’s play a role, but only in situations where the flexibility they bring to the table matches the cost and risk premium . If dedicated hardware can process 90% of the traffic transiting a switch or router, then the remaining 10% is shunted to the side for processing by the NPU. This radically alters the market size calculations for NPU’s because instead of having one or two per linecard you have two or four per rack. Design win numbers offer no details on the magnitude of future business as a result.

Ultimately, I agree 100% with Gilder’s concept that the router will be hollowed out. This process is already underway with companies such as Vyatta providing a glimpse of the future. I believe Cisco’s (CSCO) only sustainable competitive edge is software and system integration, a la IBM.

But the hollowing out won’t be done with NPU’s. I expect that Broadcom and other ASSP companies will deliver hard silicon that is good enough for the requirements of the majority of applications, and ultimately NPUs will handle exception traffic only.

EZ-Chip is highly respected company and unlike most, has managed to survive, which says a great deal about the folks running the company. The niche market for NPU’s will sustain a small set of players, and perhaps that is Ez-Chip. When it comes to Ez-Chip serving a billion dollar market, historical precedent and market dynamics are on not on their side. I hope for the sake of employees and investors of EZ-Chip that this time is different.

Raza, which appears to me as an opaque promotion company, provided little quantitative information to make a judgement. Right now it looks like all sizzle, no steak.

Full Disclosure: I am long Zarlink and Vitesse, and am short Cisco as part of a hedge. I hold no positions in the other companies mentioned, including Ez-Chip/Lanoptics.

Full EZ-Chip presentation can be found here.

—————————–

Notes Below:

The Future of Networking Silicon: Programmable vs. Hardwire

Panelists:
Eli Fructer, President and Chief Executive Officer, EZchip
Atiq Raza, Chairman and Chief Executive Officer, Raza Microelectronics
Moderator:
George Gilder, Editor in Chief, Gilder Technology Report

Amir Eyal – EZ-Chip VP Business Development

Metro Ethernet is the target market for NPUs and is the make or break market. Discussed various market research projections about how big this market will be. Provided background on EZ-Chip products and competitors.

Atiq Raza – Raza Microelectronics (RMI)

Number and intensity of connections among devices is growing. They have a family of devices that can address these markets. RMI revenue growth is only limited by operational and manufacturing constraints. They refuse to do business with startups. My bullsh*t detector just lit up.

Panel Discussion

Bob Cousins is the moderator. Starts out with the standard NPU love-fest discussion about how everything is moved into software.

My Question: “I’m going to try and create a little controversy here. The Telecosm is littered with the bodies of NPU companies who claimed they would win because they have a technical advantage. It seems to me that the guys who are really benefiting from Metro Ethernet are the Broadcoms and Marvells, as off-the-shelf Ethernet switches are bootstrapped into new applications. How exactly do you fight this trend?”

Atiq Raza – We don’t see Broadcom in our applications. BS Detector buzzing…… The real measure of success is how many chips you ship (I interrupt) ”The real measure of success is whether you make money.” Raza: We do make money, and we will be profitable within 3 quarters. (If this was true I-bankers would be chasing him like a swarm of flies)

Amir Eyal – Requirements are continually changing, soft programmable solutions are required and hard coded Ethernet switches cannot provide the needed features. Same story…. I just don’t believe it. It’s possible I am wrong but I’ll take the risk and say that this time just isn’t different.

——————-

The day began with Steve Forbes and a lightweight ‘Don’t Worry, Be Happy’ presentation. If you are familiar with his platform, you have an idea of what he talked about and I will not go into detail. It was fantastic to see him speak in person. It’s clear why his Presidential bid failed – the guy is just too honest about what he believes in. Either you agree with his ideas or you don’t, and for the most part I do. Where do I disagree?

• I don’t buy into the concept that Gold is the uber-currency because it is an empty store of value. You buy gold, it just sits there. It’s a non-productive asset. The concept of Gold as a currency is just as creaky as the Bretton Woods hotel it was originally proposed at.
• He also waved off the risk of trade imbalances. While I am more terrified of the idea of the government ‘doing something’ I am not comfortable that the export of our wealth to China in exchange for cheap plastic furniture and other useless crap at Wall-Mart is not a problem.
• Forbes also told us to prepare for conflict with Iran. Sorry, but the stick doesn’t seem to be working to well on the Islamists. Let’s figure out how to make them rich- more Milken and less Forbes.

John Rutledge followed up. I was not familiar with him until today, and his ideas resonated with me in a particularly strong way.  He keeps a blog here.

One of his ideas that resonated with me was the concept of redistributing Iraq oil wealth to the citizens of Iraq – a la the state of Alaska. This would have the positive effect of incentivizing citizens to crush the insurgency and give people a stake in the countries future. It’s classic Reagan thinking, and I like it.

I am willing to bet it would also fracture Iraq from the OPEC alliance, as the citizens would be jumping up and down to pump oil. I wonder if this is an intended outcome…

He also talked about the risks involved with globalization and that in the aggregate it is fantastic, but there would be local discontinuities that suffered - like the guy turning a wrench in Kentucky. My opinion: you cannot have Democracy without a middle class, and it is this wealth gap that Schumpeter warned us about we he envisioned an eventual fall of entrepreneurship to socialism.

I sat with Rutledge at his table at lunch and was lucky to participate in a wide ranging discussion. I proposed that he familiarize himself with the ideas of GaveKal, and read their book The Brave New World. It’s a fantastic take on how globalization financially impacts investors, governments and corporations. It is a must read. Several people at the conference were familiar with them, and I noticed two copies of the book. Gilder should get GaveKal to come next year as their message would resonate well with the audience.

The conference wrapped with Carver Mead. He is a living tech legend. He also wrote my college textbook on VLSI (Still on my shelf, should have brought it for an autograph). He talked about our incredible progress in computing and how it has been totally meaningless in the context of artificial intelligence.

• Ben Franklin was his childhood hero
• He still has copies of Gordon Moore’s sketches that lead to Moore’s law
• No one initially believed that shrinking transistors by 1/2 would double the speed and cut power usage by a factor of 8. It was “A flagrant violation of Murphy’s law”
• Speed and density went up, but supply voltages never dropped at the same rate (I think this is because of noise immunity) so power consumption skyrocketed. This wasn’t a big problem until 10 years ago.
• Impinj (Chris Segura, are you out there?) builds a .5cm^2 RFID chip with full digital radio, Flash memory and CPU that runs on 5 microwatts. This is 1/10000 the power density of an Intel CPU.
• Artificial Intelligence is not a computing power problem, it is a problem of too many inputs. It is a classic pattern recognition problem. Our best CPU’s cannot compete with the goo in the brain of a housefly dissipating a few mW.
• Described the brain and synapses as a very big analog mux tree, with many inputs and a few outputs. These mux trees are all cross wired. Most Input pulse sequences are squelched by the tree, some are amplified.
• This computing concept is exponentially more powerful than the one we use today. There is no reason why this could not be done in silicon.

There were several company presentations during the day in addition to the above presenters. When I get the time, I’ll post my thoughts and notes on the Ikanos (IKAN), EZ-Chip (LNOP). Raza Microelectronics, and Sigma Designs (SIGM) as well as a final wrap up.

———————————————

Rutledge Notes

• It’s now fashionable to call everyone a Fascist. Gilder is a “Tech Fascist” :)
• Not fair to judge people you don’t know and blow up people you haven’t met.
• Same danger applies to American perceptions of China. They are hungry and Americans don’t realize it. He feels it is necessary to educate the American people to the entrepreneurial spirit of the Chinese people.
• Saper Vedere – Leonardo di Vinci – To know how to see. It is more important to envision the final sculpture than it is to wield the chisel. This is also a good model for investing.
• You need idiots in the marketplace to misprice assets. Otherwise it is impossible to make money. Followers of Macroeconomics provide a ready supply of idiots.
• Recessions happen when small businesses and companies can’t access capital at market rates and instead must pay higher ‘shadow pricing’ rates.
• Discussion of Ricardian growth. Ricardo assumed that physical and human capital couldn’t flow. 200 years ago he was right. Modern reality is that physical capital flows from the developed world to the emerging markets, and human capital flows from emerging markets to the developed world. Telecommunications has connected the service sectors and created very large bidirectional capital flows that can change very quickly. This creates discontinuities and investment opportunities.
• These large flows are turbulent, and while global GDP will improve, and USA GDP may improve, but many people in the US will lose out and this is a problem. Unskilled workers are in big trouble.
• Inflation is such a destructive force in China, and caused so many revolutions, that they have pegged their currency to the US. This effectively outsources inflation control, and they believe the US will do a better job then their own government.
• Financial market prices are wrong. Biggest area of discontinuity is telecom. More specifics here would have been better given the audience.

This post is one of a series as I blog the Gilder Telecosm 2006 conference. All posts can be found by searching for ‘Telecosm 2006′.

Infinera took the bold and stunning risk of angering the Gilder priesthood by illustrating that an all optical network was not the future and not the best solution.

As a comm semi guy who will always have a soft spot in his heart for TDM circuit switches and electrical crossconnects, it was a message that resonated with me. But to the assembled congregation, still clutching to a vision of an all optical future, it was analogous to Galileo arguing that the sun did not revolve around the earth.

JDSU (JDSU) has made a concerted move away from the optical module business and towards subsystems like ROADMs and EDFAs. It is even rumored that JDSU has lost future 10GE module contracts at Cisco (CSCO) (this is single-sourced, any confirmation welcome) so they may have passed the point of no return. JDSU pushed hard during the panel that the future was indeed all optical, and that they were going to provide the building blocks to do it.

Intel (INTC), as I suspected, may be exiting the commodity module business, buy they have tee-ed up a big effort to do pure optical research. The UCSB/Intel announcement was just a small step, even though it received a disproportionate amount of media attention. Kevin Kahn opened my eyes to what Intel is doing outside of the commodity optical products that I used to solely associate them with. This is worth following.

Infinera laid out why O-E-O makes sense and that they have the most efficient solution as a result of their ‘optical chip’.

I remember a meeting with Infinera (Actually called Zepton before a savvy marketing guy changed the name of the company) way back where Drew Perkins laid out an array of optical modules on the conference room table. He pointed at each one,  and said that they were way too big with some very colorful language.

The vision at Infinera appears to be the same as it was then. High integration creates high reliability, high port density, and low per port cost.

I think people were mortified to find out that Infinera, the poster boy of Telecosm 2.0, is not an all-optical player This illustrates the depth to which people have gone to really understand what this company is doing, and I realized I didn’t know what sort of electrical switching function resides inside their system. I had fallen under the spell…

The best part of the discussion was the limited debate between the merits of the ROADM and O-E-O. This was a good panel, with excellent participants and should have been allotted more time simply because much time was needed to explain the differences between what Infinera does with optical-electrical-optical (O-E-O) conversion and how ROADMs work. Some carrier input, particularly from a non-Bellco perspective, would help.

This topic will be just as relevant next year and should be revisited.

——————————————————

(Italics are my comments. Plain Text is my transcription) If folks have changes/corrections by all means add them into the comments- I am not a reporter by trade.

Speakers:

• Kevin Kahn, Senior Fellow; Director, Communications Laboratory, Intel
• Michael Ricci, Senior Vice President, Optical Communications Group, JDSU
• David Welch, Chief Strategy Officer, Founder, Infinera

Panelists:

• Frank Galuppo, Chief Executive Officer, Amedia Networks
• Jerry Rawls Chairman of the Board, President and Chief Executive Officer, Finisar

Moderator:

• Charlie Burger, Senior Technology Analyst, Gilder Technology Report

Mike Ricci – JDSU

Things are different now (Uh oh, hide your wallet when as an investor you hear anything like this). We need an optical network now, we didn’t need one 6 years ago.

ROADMs, Tunable Optics, and EDFAs are the three critical components that provide everything you need to do the all optical network. (What about a large sale optical switch? You can’t build them out of ROADMs very well) 9000 ROADMs deployed so far.

Kevin Kahn – Intel Photonics (Sitting next to me, very interesting guy)

Runs Intel Enterprise Interconnect Research. He was the guy who oversaw the work surrounding the recent Intel announcement. I applaud his organization for garnering so much public attention for what was really a heterogeneous material bonding innovation. It was a good marketing job.

JDSU is focusing on what building blocks are needed to do silicon photonics. Five key areas of focus - Modulator, Laser, Photodetector, Light Guides, Low Cost Assembly (I.E. no active alignment, more like wire bonding from the chip packaging business). Silicon is infrared transparent, so the material could be used. But it can’t do 850nm… No real progress until in silicon photonics until 2002. Kevin then walked through a neat timeline of silicon innovation milestones. Good stuff.

Dave Welch – Infinera – Chief Strategy Officer

Infinera believes that optical integration drives out cost, primarily by integrating all of the O-E-O functions into a small module.

Endgame is eventually to integrate L2/L3 capabilities right on to the transponder. Some confusion among the audience, including the Intel photonics guy in particular (who is siting next to me). Do they actually do L2 monitoring and aggregation among lambdas? I don’t think so. Another sharp guy at my table says it is a big STS-1 crossconnect. I’m starting to realize that I don’t really know what Infinera does, and may be under a marketing spell.

The Achilles heel of the ROADM is the inability to re-color wavelengths – ROADMs only switch the current wavelength. O-E-O still needed in order to change wavelength colors and perform basic traffic switching. If you don’t do this you will strand bandwidth.  This has a rather interesting reaction from the Gilder priesthood, where wasting bandwidth is the mantra. This should be interesting to watch.

Q & A

Some confusion – Charlie the Moderator thinks WDM is dead. Infinera goes through the process of illustrating how ROADMs result in stranded wavelengths. O-E-O is still valuable. He handles the explanation well. This guy is an exceptional presenter to the laymen.

Infinera’s equipment is being used throughout Infinera’s network. Gilder appears to be surprised that carriers are still deploying OEO conversion… seems like there is a bias against O-E-O conversion among the moderator/Gilder. It is clear that the concept of an all optical network is such a central Gilder theme that they are having a hard time letting go. Ricci from JDSU still pushing that all optical is the future. Perhaps Infinera has now angered the Gilder gods….

Gilder asks if optics were cheap enough, would things be all optical? I think he is trying to see if the state-of-the-art is an intermediate step to optical Valhalla. Infinera guy makes point – “Future is packet based network, so I must look at the bits.” There will be O-E-O.

Gilder trying to incite a little rivalry between Infinera and JDSU O-E-O and ROADM. Who is cheaper. Who will ship more. Who can displace the other. Panelists aren’t biting. Its clear that both have a role, but I applaud the effort to keep the debate rolling!

Question: 40G vs 4 x 10G – which one wins? Infinera: 4 x
10G. 100G will be parallel first. JDSU: Increasing pressure to supply 40G (Yes of course there is pressure but what is the volume)

Question: What about the cable companies? What are they doing? (Good question, people are sooooo Bellco focused) “Yes they are.” But no information on how they are different, if at all. More representation from cable guys is needed next year – I would bet they would side more with the JDSU vision.

Full Disclosure: I hold in Intel and have a short position (hedging) in Cisco

This post is one of a series as I blog the Gilder Telecosm 2006 conference. All posts can be found by searching for ‘Telecosm 2006′.

Storewidth is analogous to Bandwidth. Storage capacity is not the problem, it is the bandwidth to the storage, between the storage, and onto the storage that is the problem.

I was pleased and surprised by the depth of the technical debate.

It is clear that opportunities exist for companies to optimize hardware components for new distributed computing architectures. My prediction: Google (GOOG) will fund startup Memory and Disk companies to supply what they need.

(Italics are my comments. Plain Text is my transcription) If folks have changes/corrections by all means add them into the comments.

Panelists:

• Luiz Barroso, Distinguished Engineer, Google 
• Josh Coates, President, Founder, Mozy.com, Berkeley Data Systems, Inc.
• Apostolos Gerasoulis, Co-Inventor of Ask Search Technology, Ask.com
• Lane Patterson, Chief Technologist, Equinix (EQIX)
• Dayne Sampson, Vice President, Information Technology, Ask.com
• Mark Stahlman, (Financial guy involved with this area, I missed his employer/title)

Stahlman: Google is a digital services company. Put $2000 price target on Google earlier this year. Feels that centralizing computing on the network is the future.

Luiz Barroso: Responds to question if his company is worth $2000 a share. “I am very passionate and knowledgeable about some things but the stock market is not one of them.” Good for him… not an appropriate question.

Patterson: Things started to shift to this massively distributed model around 2001 when Google illustrated what could be done with racks and rack of cheap PCs. Google builds their own PCs from scratch and optimizes the hardware design for reliability cost and low power.

Gilder: This approach has worked for search applications, will it work going forward? Good Question… how do you know new apps won’t break this model?

…

Gerasoulis: Power is now the barrier cost to data centers, not computing bandwidth.

Patterson: Future is 380V DC power straight into the PC. Get rid of AC power requirements.

Stahlman: Google will eventually open up their platform to other businesses. Salesforce.com allows third parties to write plugins to their application. Google will do the same thing for consumer applications. (Isn’t Yahoo already doing this? I think this is already happening? What would you call a Google Maps mashup?)

Innovation in drives is not in the right direction for use in data centers. Density keeps going up but speed does not, so in high speed applications, full utilization cannot be achieved. Seek times are a big problem, so big datacenters focus by distributing utilization. Failure rates and mechanisms are not well understood, and it is possible we could be using less backups (Google backs up in triplicate in some situations) if the mechanism of failure is better understood. NAND flash could start bridging the gap. (Intel announced some new Flash technology that would hit this market head on here.)

Audience Question: RAM Disks based on flash would only accelerate by a factor of 2 because the OS is so poorly optimized for accessing fast storage. This problem needs to be solved.

Coates: Problem exists in server interconnect. Attempts have been made (Infiniband) but the issue is always that not enough volume exists to drive the cost down low enough to be better than a big distributed architecture using off the shelf components. There just are not that many customers for this type of equipment.

Gilder: Is it possible that these massive data centers being built next to hydroelectric dams are going to be obsolete in the near future? Good Question…. will this capex yield out.

Unknown Panelist: One solution that would be disruptive is to put portion of Google search index on every device and distribute 80% of the requests. Send out updates via FedEx (Or peer to peer internetworking)

Question for Stahlman: Is Sun (SUNW) going to be acquired by Google? Answer: Anything is possible. Dell (DELL) can no longer continue as the company they are. Without R&D they are hopeless. They will merge with an EMC (EMC) or Sun. (I totally disagree with this. Dell is the ultimate platform company.)

This post is one of a series as I blog the Gilder Telecosm 2006 conference. All posts can be found by searching for ‘Telecosm 2006′.