Seeking Alpha
We cover over 5K calls/quarter
Profile| Send Message| ()  

Executives

Kenny Green - CCG Investor Relations

Eli Fruchter - Chief Executive Officer

Amir Eyal - VP, Business Development

Analysts

Daniel Berenbaum - MKM Partners

Joseph Wolf - Barclays Capital

Jeff Schreiner - Feltl and Co.

Jay Srivatsa - Chardan Capital Markets

Doug Rosenberg - RBC Capital Markets

Dan Harvard - Deutsche Bank

Paul McWilliams - Next Inning Technology Research

Sundeep Bajikar - Jefferies

EZchip Semiconductor Ltd. (EZCH) New Product Announcement Conference September 5, 2012 10:00 AM ET

Operator

(Operator Instructions) As a reminder, this conference is being recorded, September 5, 2012. I would now like to hand the call over to Mr. Kenny Green of CCG Investor Relations. Mr. Green, please go ahead.

Kenny Green

Thank you, Operator, and good day to everybody. I would like to welcome all of you to this conference call in which EZchip is publically reviewing the details of its new product. On behalf of all the investors, I would like thank EZchip management for hosting this call.

I’ll note that this call will -- is simultaneously webcast together with the presentation, which is accessible from the link on EZchip Investor Relations website. The presentation is also available for download from the same link.

With us on the line today are Mr. Eli Fruchter, CEO; and Mr. Amir Eyal, VP, Business Development. Eli will open the call and will walk through the presentation. Following that we will provide time for you questions. We request that in the first round, question will be kept to two questions per person, time permitting, we will poll again for questions after the first round.

In terms of Safe Harbor, the remarks of management during the presentation and the question-and-answer session may contain projections or other forward-looking statements regarding future events or the future performance of the company or the industry. Please note that the Safe Harbor statement in today’s press release also covers the content of this conference call.

And with that, I’d like to hand the call over to EZchip’s CEO, Mr. Eli Fruchter. Eli, please go ahead.

Eli Fruchter

Thank you, Kenny. Good day, everyone. And welcome to our NPS conference call. It is a big day for us as we believe we are not announcing yet another NPU generation like we have done every three years in the last 12 years. We are announcing today a new concept, a paradigm shift, the NPS, a revolutionary NPU that changes the old by which customers will select and use NPUs in the future.

Market leaders are often busy designing the next order generation and miss out on changing market trends, end up late to respond and lose their leadership. NPS is introduced by us, the leader in high-speed NPUs reading the market trends right and in time and setting above very high, yet again leaving the competition far behind.

We are announcing today the NPS, assembly of network processors for smart network. The NPS is a new breed of C-programmable NPUs for the next wave of high performance, smart carrier, data center and cloud networks.

Carrier data center and cloud networks are under constant pressure to be smarter and faster, i.e. more services that has speed. Network equipment as a result needs smarter and greater processing capability or smarter and greater NPUs.

Two types of processors are available today, NPUs and CPUs. When I’m talking about CPUs, it can relate to CPU with one core, two cores, four cores or multi-core. NPUs are optimized for fastest layer 2 and 3 processing, which use micro code and can process layer 4 to 7. CPUs and C-programmable general processor for the upper layer 4 to 7, but it reduce throughput and therefore limited in performance.

NPS breaks the barriers of traditional NPUs and CPUs by using C-programmable core to process layers 2 to 7 at extremely high-speed for a combination of unmatched performance and flexibility.

Processors in networking equipment perform two main tasks, data and control plane. Data plane processors must deal with every individual packet and maintain vast performance. Control plane processors deal with applications and management rather than packets and execute millions of lines of code. The difference in functionality imposes different architecture.

Data plane architecture require hundreds of optimized processors to deal with simultaneously with millions of packets and flows, plus Traffic Manager to deal with quality of services. Control plane processors don’t require Traffic Manager and require fewer and larger code with large memory cache to execute millions of lines of code.

Today CPUs even with improved RISC cores are targeting also the data plane, however, their architecture is control plane oriented and they provide insufficient performance for the data plane.

In low-speed, the CPU can execute control and data plane processing, while high-speed require different processor type for the control and data plane. Today high-speed designs can consist of three field of processors, with NPU at layer 2, 3, a CPU with multiple core at layer 4 to 7 and a control CPU for management. NPS eliminates one field collapsing three field design into two fields with the huge performance advantage and the significant reduction in both assumption and costs.

The ideal solution is with the NPS in the data plane focus on speed and the CPU in the control plane running millions of lines of code, the combination of the two will provide the best core performance results.

The NPS is unique in many aspects with technology being the first and foremost. In our 10-year old NP family that made EZchip the leading NPU supplier, we have developed special task optimized processor for layers 2, 3, called TOPs.

The TOPs were significantly and faster compared with off-the-shelf RISC processors used by other NPU vendors at that time. It turned out that the RISC based architecture could not scale to higher speeds and EZchip became the leader in high-speed NPUs with our TOP based architecture.

We are doing it now again, but for layer 4 to 7, the breakthrough is achieved through our CTOPS, C-programmable Task Optimized Processors that are 10-fold smaller and faster compared to the general C-programmable RISC processor, we have optimized the CTOPs for layer 2 to 7 data plane processing at extremely high-speed and made them C-programmable.

The NPS architecture is in contrast to CPUs that have order of magnitude for fuel engines integrated into chip because that are based on much larger general process call as large memory caches design to execute millions of lines of code and target both data and control processing.

The NPS sufficient design allows incorporating into the chip not only a great number of codes but also EZchip’s market proven hardware traffic manager and more. The end results is a very powerful network processor with superior power efficiency and integration.

In the NPS-400 we have integrated 256 CTOPs with 16 threads each for a total of 4K thread or 4000 virtual processors in the single chip. With additional hardware based Traffic Manager, DPI, Security, Search Engines, 800-gigabit of I/O, all in one chip, that deliver 400-gigabit performance at layers 2 to 7 using C-programming and the Linux operating system. NPS is targeting carrier, cloud and data center of every possible networking related application, including SDN, OpenFlow and network virtualization.

NPS is a revolutionary NPU because it leapfrogs traditional NPUs both merchant and in-house, and it comes with EZchip’s highly differentiated TM integrated. For carrier equipment NPS is adding the layer 4 to 7 market as a target market for us with features such as application recognition and IPSec VPNs both on smart line cards for standalone security appliances.

For data center equipment, NPS scales the performance beyond the level achievable by other processor for sever load-balancing, firewall, SDN, OpenFlow or network virtualization application.

By targeting data center equipment, NPS is doubling EZchip’s TAM, as well as SAM or Served Available Market by 2016.

The NPS is targeting a $1.1 billion TAM in 2016, that is in addition to the traditional NPU TAM and we believe will targeting the high-speed portion of that TAM which represents approximately $400 million in 2016. Those two numbers are similar to the NPU numbers in 2016 and combined they double our TAM and SAM.

NPS is part of our comprehensive roadmap and we are announcing today the first two members of the family. The NPS-400 and NPS-200 both developed in our Kiryat Gat facility.

Our plan is to sample the NPS at the end of next year, generate samples relative revenues in 2014 and serve meaningful production revenues by 2015.

NPS has very clear differentiator, first, NPS opens the layer 4 to 7 equipment market for us in carriers and data center, and doubles our TAM. We can win with NPS because we are providing a 10-fold faster solution throughout CTOPs technology with Superior integration and power consumption.

Second, NPS is the next wave of NPUs for next-generation Edge Routers. We believe the addition of layer 4 to 7 and C-programming leapfrogs the competition both merchant and in-house and will provide NPS customers with significant time to market and feature advantages.

Our customers will have the option of two winning alternatives in the coming years. One which is for basic line card without NP product and the other for advanced smart line cards with NPS.

We believe the 10 years from now or line cards will be smart and we will make every effort, that such line cards will be based on NPS. We believes that NPS, NPS is leaving other solutions, merchant and in-house far behind and that NPS will became the NPU choice for next-generation Edge Routers.

Third key differentiator is our technology, NPS is based on our strongly differentiated CTOP processing technology that was optimized for the data plane with 256 processors, with 16 thread each for the total of unmatched 400 virtual processors in the single chip. Most other solutions are in the range of less than 100 virtual processors which can explain the performance differences. And on top of that it includes the TM and other accelerators that make NPS an extremely highly integrated part.

Smart networks will require smarter routers and we believe that advance services that run on services cards today will have to move to the line card to the little higher performance otherwise many services cards will be required in every router to provide the advance services, not a practical solution.

With NPS customers can build smart line card that can run all the services extremely high-speed and eliminate the need for services cards all allow one services card with the CPU that can handle the less demanding general task.

We believe that this trend can increase our market penetration and possibly win an outer vendors that are not customers today. Customers will be able to use NPS line cards in conjunction with other line cards NP-based or in-house based for the mix basic and advanced line cards. In the long-term, we believe that all line cards will be advanced or smart to keep up with the advances in the network.

To summarize, what is NPS? It’s the NPU for next-generation smart network with strong L2-7 packet processing capabilities.

Why is it needed? In carriers network for next-generation smart L2-7 line cards and network appliances and in the data center for 10-fold performance increase in L2-7 processing.

Competition in Edge Routers leapfrogs in-house ASICs and merchant NPUs and in the data center its scales performance 10-fold beyond today’s levels.

Why NPS? It’s smart, simple, scalable, flexible, C-programmable. It can be used at layer 2 to 7, with a 10-fold performance increase, superior integration, power efficient and more. What makes NPS better? It’s optimized for the data plane with the CTOP architecture and integrated TM and more.

All carrier and data center equipment vendors, meaning that all existing customers can be customer for NPS, plus new business units within our customers that don’t use us today and obviously the new customers.

And with that, I would like to open the call for the Q&A session. Operator?

Question-and-Answer Session

Operator

Thank you. (Operator Instructions) The first question is from Daniel Berenbaum of MKM Partners. Please go ahead.

Daniel Berenbaum - MKM Partners

Hi, guys. Thanks for taking my questions. Thanks for doing this call. My first question is on customer buying. So both specifically for your product and for the idea of an integrated layer 2 to 7 processor on the data plane. Can you talk a little bit about, is this -- is there a hold for this product coming from your customers that they want to see all of these features integrated and do you already have any design wins? Have you already emulated this on an FPGA and you already have any design win that you could talk about?

Eli Fruchter

Hi, Daniel. Thank you for the question. So every product we do, we do when we get request, demands from customers and we work pretty closely with customers and get obviously feedback from them.

So, yes, there is a demand to increase performance in layer 4 to 7 in advanced services versus what can be done today and that exactly what we are doing. And obviously, on the technical side, we are simulating what we do. We verify what we do to make sure that what we say that we do is actually going to be on the chip.

Daniel Berenbaum - MKM Partners

So with those stimulations that you’ve already done, are there any design wins yet or do you have any customer commitments to design this into products?

Eli Fruchter

We don’t have customer commitments yet. We are still little bit far from silicon and it’s not typical for new product line to get real design win that early. Once customer start to design with a chip like that, they will have many engineers working on that and they will need from us more than what we deliver today. We believe that we will start to see design wins for the NPS in the first half of next year, towards the end of the first half.

Daniel Berenbaum - MKM Partners

Sort of the middle of next year is when you would expect to see the design wins, middle of 2013.

Eli Fruchter

Yeah. I would say that during the second quarter, I expect to start seeing design wins.

Daniel Berenbaum - MKM Partners

Okay. And then, so the follow-up question is, where is the competition here? I mean obviously with edge routing product, your competition has really been plus an ASIC. It seems that by moving layer 4 to 7 and we’ve heard other companies talk about trying to do a layer 2 to 7 data plane processor. Are you sort of moving into a bigger pond where you are maybe a smaller fish? Is that the right way to think about it?

Eli Fruchter

We are certainly moving into a bigger pond but I believe that we will be a big fish in that pond, because I think that the architecture that we are developing is very different than other architecture that are using.

I would say general purpose RISC processor and can put maybe 16 or 32 processors in the chip and those chips also typically do the general processing or the control plane processing, and therefore, a very limited in what they can do on the data plane.

We believe that control plane and data plane need two different solutions, two different processor and one cannot sell both when it comes to high speed. It can work at slow speed and we see slow speed designs today that will use one CPU, a single, dual or multi-core that will do everything.

In high speeds, we see more and more processors and as I described in one of the slides, we see 3 tiers with NPU doing layer 2 to 3 and multi-core layer 4 to 7, and the CPU doing the management of the control plane. So what we believe, we can achieve with our architecture is really collapsing it into only two layers or two tiers of processors.

Daniel Berenbaum - MKM Partners

But certainly there are other merchant. There is other established semiconductor companies that are trying to do the same thing. Let say, you, I mean, just to understand, I mean you would have more merchant competition in this space than you have with your current products? Is that the right way to think about it?

Eli Fruchter

Yeah. This is the right way to think about it. But as I said and I will repeat that road that we take is completely different. We don’t compete with general-purpose CPU. We compete with an NPU, which uses task optimized processors. We have, as I said 4,000 virtual processors on the chip and nothing gets close to that.

Daniel Berenbaum - MKM Partners

Okay. Great. Thanks very much.

Eli Fruchter

Thank you, Daniel.

Operator

The next question is from Joseph Wolf of Barclays Capital. Please go ahead.

Joseph Wolf - Barclays Capital

Thanks. My first question is, if you look at, I guess and maybe I’m simplifying this. Two pieces of what this brings, which I think is the C-programmability and the layers 4 to 7, to what looks to be the regular speed roadmap of where the NP, the classic NPU product was?

Could you go through some of the cost of development? By my rough calculations you spent about or have invested about $10 million on this. Was most of this a software, hardware, where was the real breakthrough, was it at the programmability, what were the customers looking for in this design?

Eli Fruchter

I think this is from an R&D and cost perspective. It’s part of what we’ve done so far. About 40% of R&D is working on that chip right now and it would be 50% of the year and going forward. So this is becoming a significant part of R&D.

And a big portion of that is the development of the CTOP. That’s really real unique in what we do. That’s the development of a tiny processor that is extremely high-speed. It is C-programmable as I said and can operate that layer 4 to 7, and that processor makes a big part of the differentiation between what we do and what others do.

And I want to repeat again, I believe that we are not only targeting layer 4 to 7 here, that’s the new market for us, that we’ve not targeted so far. But we are also providing solutions for Edge Routers our main market, which we believe is really leapfrogging our competition. And I believe that that will actually strengthen our leadership in that market, which is our main market today. So we are actually achieving two things with one part.

Joseph Wolf - Barclays Capital

Well, I guess, my second question feeds right into that. If you look at the carrier market versus data center, it would seem that NPS is the migration path for your current NPU customers and there aren’t that many customers for you to win?

And so when you look at that, or you look at the data center and you talked about a design win by 2Q ’13, what’s the real application or do we see it yet for 4 to 7? It seems very, very high-speed. It’s got a new architecture. Even more than the hardware, it feels like there has got to be a sales team or some sales spending and some marketing to get into, getting some design wins here? If we look at that 2Q ’13, is that a carrier win and when do we see a data center win?

Eli Fruchter

We believe that we will see wins in both at that timeframe and I don’t know. Amir, do you want to add something to it?

Amir Eyal

Joe, perhaps for the applications though, when we are talking about the data center applications, there are number of applications where we see good potential and those are in general, it can be described as security, load-balancing and SDN. I would say were probably the key ones for us.

Joseph Wolf - Barclays Capital

Okay. And just I guess just to round out, in terms of sales investment and marketing, does that ramp up 2013 to build awareness for the new architecture?

Eli Fruchter

So, what we said our OpEx for the next year that we spoke about includes everything. Remember that our current customers are also customers for the NPS, not necessarily the same business units but other business units within the same customers and we are well connected there, and we don’t need to make any special efforts to get in front of those additional business units. We actually already did and there is a group of other customers. And in our budget in the OpEx that we guided, it is included.

Joseph Wolf - Barclays Capital

Okay. Thank you.

Eli Fruchter

Thank you, Joseph.

Operator

The next question is from Jeff Schreiner of Feltl and Co. Please go ahead.

Jeff Schreiner - Feltl and Co.

Yeah. Good morning, gentlemen. Thank you very much for taking my questions. I guess I just would like to give a little more insight into who’s really the largest shareholder in terms of market share within layer 4 to 7 right now and doesn’t this start to bump EZchip in terms of direct competition with its current partner, Marvell?

Eli Fruchter

So, many of our existing customers are large shareholders of layer 4 to 7. Cisco is obviously one of them, almost in every segment in networking, not just ours. But it also opens new customers that are not customers today, like people that will do just load balancing or others that will do just security, firewalls and all that kind of stuff. What was the other part, Jeff?

Jeff Schreiner - Feltl and Co.

As you get into the 4 to 7 -- layer 4 to 7 market, we know about Cavium, I think Broadcom is playing there. Is it Marvell, your current partners also playing in these markets?

Eli Fruchter

And I would say to less extent, I think that both come in the Cavium to higher extent.

Jeff Schreiner - Feltl and Co.

Okay. And then my second question. Just trying to understand that we have a new kind of TAM and SAM for EZchip based on the release of the information today. We know how high the percentage of the SAM and high-speed Edge Routers you currently have. What’s the reasonable percentage that you are targeting or you thinking maybe able to achieve of the new $400 million SAM from NPS?

Eli Fruchter

Well, it’s really very early to say, Jeff. We are a year or more than a year from sampling the part. We do not have design wins yet, but obviously we believe that we can get a significant portion of that market because we think that we provide the strongly differentiated technology. But it’s too early to talk about target numbers then.

Jeff Schreiner - Feltl and Co.

All right. Thank you very much.

Eli Fruchter

Thank you, Jeff.

Operator

The next question is from Jay Srivatsa of Chardan Capital Markets. Please go ahead.

Jay Srivatsa - Chardan Capital Markets

Thanks for taking my questions. Eli, given that you’re still a year away from sampling and no design wins yet, what’s the impetus to announce the product now and are you concerned that you might be empowering some of your competitors to start looking at a similar product to what you are announcing today?

Eli Fruchter

I think that people -- since we want to expose it to more and more customers because as I said, I believe that in the second quarter of next year, we will be able to start and get design wins. I think that it’s not the secret. Once you approach many customers, the information becomes available.

And secondly, it seems that we have -- we spoke about it for a long time to investor because it obviously increased our OpEx and we needed to explain what it is. And we got lot of questions about what we do and we felt that it’s the right time. We are already working on that thoughtfully although now it’s a long work. Its -- I don’t think that -- I think that we are ahead of any potential competition in that space.

Jay Srivatsa - Chardan Capital Markets

All right. My second question was related to process geometry. Have you been able to finalized what type of the process you are going to be using and how is that -- how would that compare to some of the competitor solutions out there?

Eli Fruchter

We are going to use 28 nanometer for that part and they think that most part that are designed to the high-speed thoughtfully. Greater processing are designed using 28 nanometer.

Jay Srivatsa - Chardan Capital Markets

Thank you. Good luck.

Eli Fruchter

Thank you, Jay.

Operator

The next question is from Doug Rosenberg of RBC Capital Markets. Please go ahead.

Doug Rosenberg - RBC Capital Markets

Hi. Thank you. Congratulations, guys. It’s more of a follow-up. The first question is, I understand that it became more parallel to the NP -- the traditional NPU families, but do you -- are you afraid of some sort of cannibalism of the NPS taking some of the market away from the NPUs, NP-5, NP-6?

Eli Fruchter

Hi, Doug. Thank you for the question. So as I said, I think that we are -- with NPS we are providing customers with alternative. We provide them to NP alternative for basic line cards and we provide them their NPS alternative for smart line cards. Then if -- some how I believe that all line cards will be smart and between now and then, we provide both alternative and during that time, we will see design wins for both.

Doug Rosenberg - RBC Capital Markets

Okay. And just as trying to understand a little bit more onto the numbers, the TAM and SAM, is there anything, anymore details you can get. You gave us to how you came -- how you came the numbers or always there are going to different pricing I guess versus the -- I mean, it’s going to be a little bit more expensive I am guessing?

Eli Fruchter

Yeah. So, we get those numbers from companies that do market research and we use -- really use number from Informatics Research, from a delivery group and we had also our input with that and right now we still don’t have that pricing for the NPS, but the assumption that it will be a little bit more expensive is correct.

Doug Rosenberg - RBC Capital Markets

Okay. Thanks.

Eli Fruchter

Thank you, Doug.

Doug Rosenberg - RBC Capital Markets

A couple next time.

Operator

The next question is from Dan Harvard of Deutsche Bank. Please go ahead.

Dan Harvard - Deutsche Bank

Hi, good afternoon. And thanks for taking the question. Couple of questions. Firstly, how long would be the design cycle for new customer? And then secondly, on the manufacturing sides, do you already have potential partner in place? Thank you.

Eli Fruchter

So the design cycle, its very similar to the design cycle of the NP line. It’s a pretty long. And I would say that if we will start to see design wins in the second quarter of next year and we are talking about meaningful bucks in 2015, it means that they its going to take -- the design cycle is about 18 months, which is typical and we have seen it in the NPU as well.

Dan Harvard - Deutsche Bank

So, does this imply that the potential Q2 design wins would be with existing customers rather than new customers?

Eli Fruchter

Not necessarily. It is possible that it will be a mix.

Dan Harvard - Deutsche Bank

Okay. And then on the foundry side?

Eli Fruchter

On the foundry, we have not -- we are not giving any information on that now, we will know soon.

Dan Harvard - Deutsche Bank

Okay. And then just one more question obviously in terms of reaching out to new customers you have to have some kind of go-to-market strategy and that will be less than more divest customer based in your existing base. How are you planning to go ahead with that?

Eli Fruchter

As I said before, we are also -- we are already doing it. We are already approaching, obviously we approach all of the existing customers, decent business units within those and that’s obviously easy for us. But we also now identify the potential customers for product that are not customers today and we are approaching them.

Dan Harvard - Deutsche Bank

Okay. Thank you and good luck.

Eli Fruchter

Thank you, Dan.

Operator

The next question is from Paul McWilliams of Next Inning Technology Research. Please go ahead.

Paul McWilliams - Next Inning Technology Research

Good morning. I guess you mean Eli and Amir. Thank you for taking my call. How many NPS’ can be put on a single line card?

Eli Fruchter

So it should be very similar to -- if you are talking about NPS as the next generation NPU online. It would be very similar to the NP line. And that would be up to better line comp.

Paul McWilliams - Next Inning Technology Research

Okay. Now, that would be facilitating low as 2-7 services in this case, correct?

Eli Fruchter

Yeah.

Paul McWilliams - Next Inning Technology Research

Okay. What would be the throughput of that line card without over subscription?

Eli Fruchter

So the bulk is the form of gigabit. So it will be eight times 400 without other subscription and it could be double than that with other subscription.

Paul McWilliams - Next Inning Technology Research

Okay. So 3.2 terabits?

Eli Fruchter

Yeah.

Paul McWilliams - Next Inning Technology Research

Very impressive. I realize I’m limited to two questions so I will go back in queue but I do have more I would like to ask.

Eli Fruchter

We’ll wait hopefully until we answer all your questions.

Paul McWilliams - Next Inning Technology Research

Thank you.

Operator

The next question is from Sundeep Bajikar of Jefferies. Please go ahead.

Sundeep Bajikar - Jefferies

Hi, guys. Thanks for taking my question. Couple of follow-ups. So first is it hard to think of the incremental TAM for the NPS as being from the Security and DPI type of layer 4-7 functionality. And if so what exactly are you adding for security or DPI that’s different from competition today?

Amir Eyal

Hi, Sundeep. This is Amir. The additional term is definitely coming from applications that require DPI and Security but not only as we mentioned before. In order to take advantage of applications that need layer 4-7 or advanced applications you need layer 4-7 possessing that do not necessarily require specific hardware acceleration.

If you need to for instance to terminate session et cetera, it doesn’t necessary require special hardware but you need very elaborate processing capabilities. So that we have enabled with the new CTOPs that we have in the device.

Regarding the specific [cards] ratios that we have and so insecurity we have more or less what you might find in some CPU better targeting layer 4-7 application. So there is nothing because especially unique in that with the Gulf because DPI we believe we have achieved some breakthroughs in the hardware accelerator that we have and algorithms that we use that enabled very efficient regular expression pattern matching.

And then top of that we have significant accelerators that we haven’t really touched on yet but one of them simply certainly was mentioning is the traffic management, which is the unique technology that we have when compared to other CPU that are targeting their 4-7 applications.

Sundeep Bajikar - Jefferies

Thank you so much. That’s very helpful. Just another follow-up. If the NPS lines is the successful what impact would it have on the microcode programmable NPUs. I guess the question is are you expecting your customers to maintain two parallel lines of software development for your NPUs one for the NP and one for the NPS?

Amir Eyal

Sundeep, I think that we are giving them the alternative. So if they want to continue with NP line because of all the folks to investment that they make on that line then they can do it. They can -- they use NP-4 now, they can use the NP-5 and NP-6 afterwards and so on. And if they at some point want to get advantages of designing smarter line cards then they will move to the NPS and we will obviously offer assistance in tools to make the transition easy for them.

Sundeep Bajikar - Jefferies

Okay. Thank you so much, guys.

Amir Eyal

Thank you, Sundeep.

Operator

(Operator Instructions) The first question is a follow-up question from Paul McWilliams of Next Inning Technology Research. Please go ahead.

Eli Fruchter

Operator, we cannot hear Paul.

Operator

Paul, please go ahead.

Paul McWilliams - Next Inning Technology Research

Hello?

Eli Fruchter

Yeah. Yeah. We can. Now we can hear you.

Paul McWilliams - Next Inning Technology Research

Okay. How substantially would a customer have to change their control plane design and/or software to take advantage of the NPS. Let’s say a customer like Cisco in an ASR 9000 driver?

Amir Eyal

Hi, Paul. This is Amir. So obviously whenever you introduce and you talk with the new architecture into assistant and there need to be some adaptation done. I can tell you based on our experience working with the NP line. We have many customers that have adopted their controlled plane to look next two NPUs. This is not a very big task. It’s not something that is negligible but is not something that should be a significant factor when a customer decide to use or not to use particular solution.

Paul McWilliams - Next Inning Technology Research

So a customer like Cisco and I am just using ASR 9000 as an example could have the option available to their customers who have a NPS smart line card that they could use or NP layer 2-3 line cards that they could use in the same machine?

Amir Eyal

I don’t know fairly one specifically about one customer or one particular type of system but it certainly we see a scenario where customers will -- where our customers will offer to their customers. While we would consider simpler line cards that we will be based on the NP and higher price, higher future, the line cards based on the NPS. And as Eli pointed out earlier in the call, we expect transition period where more and more smart cards will be used instead of the basic cards.

Paul McWilliams - Next Inning Technology Research

Very good. That’s very logical. How much TCAM is in the NPS. I am driving to how much external TCAM a company might need to fully leverage the NPS?

Amir Eyal

So the TCAM is -- I don’t want to get into specific but its measured in a few megabits that integrated or embedded TCAM we have in the device. Discrete TCAM device out in the market are measure in tens of megabits. So the intent we note to provide the same size of TCAM embedded into the device however we do offer capabilities to extend the TCAM into larger space using algorithmic method. And the additional space actually resides on external DRAM to the NPS.

Eli Fruchter

Really depending on the application this might be relevant or useful for the customer and in some cases it might not. Also their applications that need a large TCAM to earn and often applications that need small TCAM. So those application, the embedded TCAM would be sufficient. So there are variety of solutions here that are available for the customers.

Paul McWilliams - Next Inning Technology Research

Very good. Could you help -- I think what would be helpful for a lot of people here including me, would be a description of the services that you do in layer 4-7. What services does that enable, not just the technical terms but the services we might recognize them?

Eli Fruchter

So switching and routing, I think that we can escape but moving further on, we’re talking about things that are security related. So any data center and any carrier network needs to number one provide security for the servicing side inside the network. On one hand and on the other hand, if we’re offering services to end customers, it needs to provide security for those services that are offered to the end customers.

That’s one type of -- actually security is a name that encompasses many types of applications. Another big type of application would be the load balancing where you really want to be able to split the load based on different criteria between service or between switches in the network.

And the smarter you do the load balancing, the better distribution or better balancing effect you have of the traffic load in the networks. That’s what we come in. The whole new requirement especially in mobile networks where people -- people I mean, wireless providers are looking to look at the traffic that goes over the network more closely. So they need capability for application recognition to be able to number one analyze the traffic that’s going on the network and make decisions on how the network should be architected for the future.

And number two, allow them to perhaps provide different billing method for peers. For application recognition is a big deal. In every network, especially we see that becoming extremely required in the further wireless infrastructure. Does that answer your question?

Paul McWilliams - Next Inning Technology Research

Yeah. Very good. I appreciate that. My last question here is on virtualization. It would seem that based on the number of threads, the 4K threads, you could handle and virtualize environment in NPS text virtualization to a considerably higher level than the current alternatives could. Is that the right way to look at it?

Eli Fruchter

That’s wrong way to look at it through the number of threads but certainly virtualization is a big deal in other ways as well. The trends now, the buzz now with software-defined networks and OpenFlow, we could be able to control the network from a central location and virtualize the different elements or notes in the network and a processor like the NPS that has a lot of processing element certainly end of process, flexible in capabilities tends to support those capabilities in various sectors whenever you can -- really as you point out because of the many processing, different process source that we have on the device, many virtual elements and virtual switches can be implemented on one or more NPS devices.

Paul McWilliams - Next Inning Technology Research

Thank you. That’s very helpful. That wraps it for me.

Eli Fruchter

Thank you, Paul.

Operator

The next question is a follow-up question from Daniel Berenbaum of MKM Partners. Please go ahead.

Daniel Berenbaum - MKM Partners

Just follow up on an earlier questions, so when I asked about ASPs, you -- if I understand that correctly, so that ASPs on the new products would be higher than your current ASPs. Did I understand that correctly?

Eli Fruchter

Yeah. That’s correct.

Daniel Berenbaum - MKM Partners

Will there -- when you talk about offer NPS 400 and 200, will there eventually be a broader range of products where you could potentially see lower ASPs in order to be able to address more of the market. Is that something you would consider and along those lines, how should I think about the impacts of gross margin from these products?

Eli Fruchter

So obviously, all those options that you mentioned are open and we can go to the upper speeds and we can go to the lower speeds and broaden the family. I believe that we are talking about the gross margin that are similar to the gross margin that we have in our current product.

So I believe that long term, we would be able to continue and see the same level of gross margins also and then NPS become significant portion of our revenues.

Daniel Berenbaum - MKM Partners

Okay. Great. Thanks very much.

Eli Fruchter

Thank you, Daniel.

Operator

The next question is a follow-up question from Jeff Schreiner of Felt and Co. Please go ahead.

Jeff Schreiner - Felt and Co.

Yeah. Sticking with that Eli, are we going to see the similar type of profile or ASP cycle that has been present within the NPU markets where as you introduce the new product, it sees a decline in the legacy products but then gives you almost a doubling of the ASP in the new offering. Would that repeat with the NPS product?

Eli Fruchter

You know, it’s too early to say Jeff. Theoretically, it can follow the same trend but that’s too early to say. We’re still -- more than two years from any significant revenue than on the NPS and even longer than that until we let that at the next part of that family. So it’s really too early to just comment.

Jeff Schreiner - Felt and Co.

And then can you talk Eli about, we hear today now. We’ve dealt with NPUs as the major product line for EZchip for many years and now you are introducing the NPS line and hopes of expanding your end-market opportunities. Is there further expansion of end market opportunities for EZchip with its NPU or NPS product lines are based on the current market you are trying to serve today?

Amir Eyal

Hi. Jeff, this is Amir. Yeah, there are always opportunities. We are always looking at where we should put our R&D resources. Right now the focus is NPS but certainly we’re seeing additional opportunities. If you, kind of, fit from a high altitude look at what we’re doing, we’re essentially always trying to integrate or add additional functionality of a line card of an appliance into our devices and provide better integration, better power and better cost savings to our customers.

So that’s one way to look at it but another way to look at it is to provide solutions that work next or enhance the chips that we do today. So we’re looking at both options always.

Eli Fruchter

And Jeff, let me ask you when you ask that did you mean beyond the additional count that we already targeting with NPS. We are doubling the time. You are talking beyond it, right.

Jeff Schreiner - Felt and Co.

Yeah. I always want so. So I was hoping that…

Eli Fruchter

That’s okay. Just wanted to make sure but one step at a time.

Jeff Schreiner - Felt and Co.

Okay. Last question for me, gentlemen, thank you for your time today. Just want to go back to the software investment and try to get a level of the investment required for NPS relative to NPU for your customers if they were to make that switch?

Eli Fruchter

You mean how much resources they would be required to allocate.

Jeff Schreiner - Felt and Co.

Yeah. Just trying to look at kind of total cost for them to make the switch in terms of what they’ll have to invest on their end to move over to NPS after the investment they made into NPU?

Eli Fruchter

I think it’s about the same. I think in many ways we’re making it easier for our customers to work with the NPS because of the C programming language that you don’t need to spend time to teach the customers because of the operating system which is available on the NPS and was not available on NP products. So there were significant enhancements or changes for the program that work with our devices. It’s a major enhancement for them.

Jeff Schreiner - Felt and Co.

Okay. Thank you very much for your time gentlemen.

Eli Fruchter

Thank you, Jeff.

Operator

The next question is a follow-up question from Joseph Wolf of Barclays Capital. Please go ahead.

Joseph Wolf - Barclays Capital

I think this was touched on a little bit with the virtualization answer but when I think about this product, is it geared towards the multiple box vendors that are now doing the different types of application, just talking about DPI and Security or is this product something that would enhance a vendor to do multiple things directly on the router?

Eli Fruchter

So you know, even today you always find that some applications are provided in those two forms factor. You find the special appliances that are designed to do security or whatever and you find the routers that provide this through an add-on as services cards. So really moving forward both approaches will continue and both are viable and we enable both of them.

Joseph Wolf - Barclays Capital

Okay. Great. Thank you.

Operator

The next question is from Jay Srivatsa of Chardan Capital Markets. Please go ahead.

Jay Srivatsa - Chardan Capital Markets

Yeah. Thanks for taking my follow-up. Eli, as you look at transitioning some of these customers who are potentially using some existing solutions out there. I guess, the question is what is the switching cost for the customer and how easy or how hard would it be for them to go from existing products to an NPS from EZchip?

Eli Fruchter

Just it will really depend on what the NPS will loss. If the NPS will loss sort of 10% improvement then they will not bother. They will not make the investment but if it will provide the 10-fold increase in performance and they will not do it, I think they will not be competitive.

So right now, I believe, the port we are actually building is something that will give a use performance advantage and that’s why it seems that they will make this hopeful investment therefore to move to the NPS.

Jay Srivatsa - Chardan Capital Markets

Thank you very much.

Eli Fruchter

Thank you, Jay.

Operator

There are no further questions at this time. Before I turn the call over to Mr. Fruchter for the concluding statement, I would like to remind the participants that the replay of this call will be available on the company website at www.ezchip.com. Mr. Fruchter, would you like to make a concluding statement.

Eli Fruchter

Yeah. I just want to thank everyone for joining us today. Again I feel that it’s a big day for us. It’s -- we are already two years in that effort and is going to take another year before we see something but we believe that we are developing product that is very, very strongly differentiated. Thank you very much for joining us today.

Operator

Thank you. This concludes the EZchip conference call. Thank you for your participation. You may go ahead and disconnect.

Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to www.SeekingAlpha.com. All other use is prohibited.

THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.

If you have any additional questions about our online transcripts, please contact us at: transcripts@seekingalpha.com. Thank you!

Source: EZchip's CEO Hosts New Product Announcement Conference (Transcript)
This Transcript
All Transcripts