Article Info

EZ-Chip and Raza Micro at Gilder Telecosm 2006

NPU companies consistently make the case the market is moving into their domain and that technology is their edge, right up to the point they go out of business.

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

Eli Fructer, President and Chief Executive Officer, EZchip
Atiq Raza, Chairman and Chief Executive Officer, Raza Microelectronics
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.



Comments are disallowed for this post.

  1. I would tend to agree with on the upside for NPUs, although, I will be the first to admit that I haven’t been following the 10G market to closely.

    I can say that that BRCM, MVRL, Agere, and to some extent Vitesse are adding processor cores (or even multiple cores) to a lot of their products, giving the flexibility to adapt to changing standards.

    With draft 802.11n BRCM and MRVL came out with chips that combined the radio to somewhat of a generic ARM/MIPs processor that allowed them to implement a soft MAC, which was important since the standard has not been ratified. Of course, their next gen products will codify the MAC in silicon, enabling lower cost solutions.

    The same thing is happening with the 10/00/1000 switches. Vitesse introduced a 24-port switch with an ARM9 earlier this year The other have done, or are doing similar things with their switch offerings.

    As far as CSCO, its software and integration are big competitive advantages that continue to help it win deals. However, I will add that while the US and Western European markets will continue to evolve and grow, a large precentage of growth in the Ethernet WAN/MAN market will take place outside of of those regions. Only, CSCO, Alcatel, Huawei, and maybe NT and Juniper have the global presence in terms of sales and support. Of those, only CSCO, and maybe Huawei have the right product mix to address those markets.

    So while you can make a nice little router with a Dell PC and Vyatta software, you still have to market, sell and support it.

    Posted by Sak | October 10, 2006, 11:09 PM
  2. I would agree, no Vyatta replacing the CRS-1 anytime soon.

    Posted by Andrew Schmitt | October 10, 2006, 11:35 PM
  3. * good work on blogging the conference!
    * NPU’s good points — it goes way back to Maker. The story has always been the same :-( If it works, it usually only works for one thing :-( ( same as ASIC only more expensive )
    * Infinera – Drew Perkins has always been in the right place at the right time – Fore, Lightera, and now Infinera. He knows ….
    * I agree with the point that — The network is built for the applications. ( ie Google, YouTube, MSN, eBay, Porn, et al ). The application layer is what is “cool” and why everything else is req’d.

    Posted by Iain | October 11, 2006, 12:52 PM
  4. If you look at who has market share in Metro Ethernet, everyone of these vendors is shipping custom-designed, programmable silicon. What kind of market share do Broadcom and Marvell have in this space?

    Posted by Schlettie | November 14, 2006, 4:44 PM
  5. Any opinions about Cavium?

    Posted by protocol_junkie | November 14, 2006, 5:13 PM
  6. Yes. Cavium (Octeon) is suited for L5-L7 processing at a few gigabits/sec, max. It is completely unsuitable (i.e., way too expensive) as a L2/L3/L4 forwarding engine.

    Posted by Schlettie | November 14, 2006, 8:37 PM