Fathers and Sons

Rhines, Madhavan,
Turgenev & Hemingway ...

by Peggy Aycinena

Editor's Note: An edited version of this article first appeared on-line in EDA Nation in August 2004.


Chapter 1 – Evolution and Revolution
Chapter 2 – Wally Rhines
Chapter 3 – Rajeev Madhavan
Chapter 4 – The Epilogue


Chapter 1 – Evolution and Revolution

Everything and nothing changes from generation to generation. We don’t need great Russian novelists to tell us that. Fathers more often than not represent the status quo, whereas children, or their friends, more often than not represent change. What is it about the newer generation that it’s always so anxious to pursue change? Is it change merely for change’s sake, or is it because the ways of the fathers have become tired and staid, the world has moved on and the old ways no longer apply?

For some, these are the questions that haunt the EDA industry today. And the questions are more than just academic. A lot of money is riding on the answers. EDA investors, EDA shareholders, EDA executives, EDA junior executives, the rest of the EDA staff, EDA analysts, journalists, and – last but not least – users of EDA tools, all await resolution: Is it change for the sake of change or are the ways of the fathers too tired and staid to apply to the brave new world we live in?

I think these were the issues I was trying to unearth in several conversations on Monday, July 26th. First I had a marvelous chat with the ever-engaging Wally Rhines, CEO at Mentor Graphics Corp. Following that, I had an equally stimulating conversation with the ever-energetic Rajeev Madhavan, President and CEO at Magma Design Automation.

Mentor Graphics was founded in 1981. Today, Mentor has 3700 employees and most recently reported $670 million in annual revenues. Magma was founded in 1997. Today, Magma has 530 employees and is reporting something in the order of $150 million in annual revenues. Both Magma and Mentor are publicly traded.


Chapter 2 – Wally Rhines

Q – New technologies are emerging – FPGAs, structured ASICs, ESL – how does Mentor position itself in new markets while still protecting established markets?

Wally – "Of necessity, Mentor does actually develop products internally for new markets. [Because] we are less well established than our leading competitors, we need to pursue new opportunities and take advantage of new markets. The only way an EDA company can grow is with new applications, new methodologies, and new customer bases – or totally new markets. And, we need to grow in order to generate satisfactory profitability for our shareholders and to provide benefits to new users."

"In FPGAs, we’re the only major EDA company who has made a substantial investment in this area. Marketing and supporting FPGA products requires a different kind of selling channel and marketing approach. You need to be able to make money on the sub-$10,000 sales, [and we can do that because] Mentor has two distinct sales channels. One is a direct channel where Mentor employees make the sales call, and one is the indirect channel where specialists and resellers of Mentor tools make the sales call – third parties who are willing to call on the small companies."

"When we acquired Model Technology [in 1994], the economics of their technology allowed us to be able to approach the smaller markets. ModelSim built up a worldwide channel of sales reps for us – highly skilled distributors – and we’ve built upon that. So now we have about a quarter of our revenue going through that sales channel and we’ve augmented it [to approach] the FPGA market."

"In the structured ASICs market, we support those customers with platform-based design tools. There’s a class of ASICs that’s basically a fixed-base set of functions, where you customize three-to-eight layers of metal. That device can also be referred to as a platform ASIC. We haven’t done what Synplicity has done [to approach the structured ASIC market], but we have done tools for platform-based design."

"Now, it’s true you can distinguish between structured ASICs and platform ASICs – but at the same time, there are probably 10 other categories as well within this technology. Some approaches have FPGAs in standard products; others have standard products in FPGAs. Or you have people generating new architectures for programmability, or reducing programmability in existing architectures. Whenever you see these kinds of things happening, it means the providers and the customers are searching."

"That’s when EDA companies are faced with two approaches. The first is to be engaged in the market, to have strong positions in FPGAs, for instance, or in synthesis, or simulation, or analysis. To see what’s working and to introduce new products to support it. That’s the way products are best defined – you have something to offer and you interact with customers to have them help you figure out how to make it better."

"The other approach is to wait and see how things sort themselves out in the marketplace, and then go out and acquire the winner. This second approach hasn’t been a good one for Mentor. Am I bragging or complaining? A little of both. If I had a 10x market cap, I might rely more heavily on acquisitions, but I don’t. Our discipline at Mentor is geared more towards product development."

"As far as ESL is concerned, you should check with Dataquest. I believe their numbers indicate that we’re pretty well established as number one in that space. Our major competitors have made a lot of moves in that area, but ESL requires you to move above the HDL basis of design. We’re definitely doing that. We’re providing verification with software and hardware – we’ve been quite successful in that market for 6 or 7 years – and we’re also providing products for C-based design. ESL is definitely [a reality at Mentor]."

Q – How do you sell the tools? To management or to engineering?

Wally – "We have to do both, although whether you’re selling bottom up or top down depends on the different cultures within the customers and the different economic cycles, which have a strong influence [on the process]. Traditionally, Asian and European companies have exerted more high-level management decisions with regards to purchasing EDA tools. The U.S. has always tended to be a more bottom-up culture, except with the caveat of where interoperability has forced higher levels of decision making."

"With the downturn in 2002, some of the larger companies became sufficiently troubled that financial controls started to override engineering judgement. Contracts were not renewed or were re-negotiated without paying attention to the engineering viability of the products. It was clearly a penny wise and pound-foolish mentality. As customers become less price sensitive, however, they can again make decisions based on the technical capability and advantages of the tools they’re buying."

Q – Power and signal integrity are the agonies that accompany the ecstasies of 65 nanometers. Why don't we just stop at 90 nanometers, resolve the outstanding problems there, and leave well enough alone? Does it make good business sense to be pushing the envelope?

Wally – "Let me give the Edmund Hillary answer … because it’s there. Leading-edge technology tends to pay the bills in the EDA industry. If you look at who buys the most expensive EDA tools, who adopts the new methodologies, it has historically been those who have pushed the frontiers of integrated circuits. They may be a minority, but they include companies like Intel and TI. These days, those types of companies don’t believe they have the choice to make their own tools. There’s such an economy of scale in solving [CAD tool] problems for next generation technology, that companies like Intel and TI can’t afford the development costs. As an EDA company, Mentor is solving problems where we’re listening to the inputs of many diverse users of our technology. We solve problems that the single user hasn’t anticipated as yet."

"This is a phenomenon that you’ll see in more industries than just EDA. Cirrus Logic, for instance, used to give this presentation describing how – for every generation of disk controller – Seagate Technology would develop their own disk controller, but would also engage Cirrus to develop [a competing design]. For seven generations in a row, Seagate chose the external solution. That’s because internal design groups develop solutions for the next generation, but don’t anticipate a generalized solution."

"It’s been the same for the evolution of the EDA and the semiconductor industries. Having you own team develop tools in-house is very expensive. EDA companies are able to solve custom design problems with generic design solutions. This is a law of economics that’s unlikely to change."

Q – Wally, would you rather be working at a start-up where there’s free soda and popcorn?

Wally – "Actually I went straight from my Ph.D. at Stanford to 21 years at Texas Instruments, so I’ve never actually been in a start-up. Although, Mentor does offer free coffee and free popcorn. We have always tried to preserve some of the things that are associated with start-ups. We have a putting green for employees, an exercise facility, and on certain days you can get a hair cut or a massage at the office. We also provide a very fine child development center."

"But none of this is why people work at Mentor. People are primarily looking for the opportunity in their work to do things that are unique and challenging. The software industry has always allowed people to make a difference in technology. That’s why people work at Mentor."

"Why don’t I go to a start-up? Well, by definition I would rather be working where I am. I have had plenty of opportunities to do other things – lots of opportunities, for instance, to go back to managing semiconductor companies. But, I don’t feel a strong temptation to do that. The kind of thing I do at Mentor is my creative outlet, because we do act like a start-up."

"I am aware in all of this that once you become a large company, it takes an enormous amount of effort just to stay even. But we’re still a young company and we’re continuing to grow in several markets, mostly by taking market share from our competitors. When a company has good cash flow and good earnings, it should let its investors put their money into growth opportunities [available to the company]. Of course, when opportunities are not plentiful, just as Microsoft is currently doing, that cash should be returned to the shareholders."

Q – The Big Guys in EDA are accused of a) inhibiting innovation to protect legacy markets, b) inhibiting innovation by buying small companies and letting those technologies wither, and c) inhibiting innovation by only partnering with big customers. Do the accusations become tiresome and annoying for Mentor Graphics? Is it the tolerable price of success to always be the target of broad-brush criticism?

Wally – "I think it is the price of success that there will always be critics. [That’s exacerbated by the fact that] people tend to broad brush groups into the same buckets, companies are placed into generic categories. There is validity in the criticism that large companies are guilty of not developing technology internally, and only grow through acquisitions. But how many de facto industry standards are developed by start-ups? Usually, they’re developed after the company has reached [a certain size]."

"Calibre from Mentor Graphics is certainly the de facto industry standard for physical verification – it’s one of a half dozen products that are now industry standards that were developed after Mentor became a large company. And certainly, the same thing is true of PrimeTime at Synopsys as far as being a de facto standard that emerged from a large company."

"[In the end], it’s a question of what process a company finds to be the most efficient for developing technology. If waiting, watching, and acquiring technology is a [way to establish] leadership, companies should do so and accept the criticism."

"For Mentor, however, we don’t have that luxury. It’s true that in the early 90’s, we had fallen way behind and had to jump start our internal development by doing quite a few acquisitions. But historically, Mentor has always been a product development company, and I made the decision when I came to Mentor from Texas Instruments in the early 90’s that we would be much less open to acquisitions."

"I believed and still do, that once you have the culture [for internal development] that Mentor always had, the company has to take advantage of the development skills it has in-house. Today, Mentor is 5-to-10 times the size of our next largest competitor, but even as a big company, we’ve been able to preserve our ability to innovate. I know we would be able to preserve that ability even if we were to grow to twice or three times our present size. Why? Because you always dance with the girl you brought to the party!"


Chapter 3 – Rajeev Madhavan

Q – Rajeev, I'm often hearing Magma referred to these days as one of the "Big Guys" in EDA. Is that affecting your business or technology initiatives? Was it more fun when Magma wasn't so much on the radar?

Rajeev – "Actually, from the very first product launch, we’ve been on the radar. That hasn’t changed in any way, shape, or form. There are two big EDA companies who like to see us make mistakes – we’re up against two Goliaths – and a third who sits on the fence and congratulates the other two. Since the early days, it has always been about bashing Magma."

"But also from the very beginning – if John Cooley or anybody else attacked us on things that were wrong with our products, we’ve always gone after fixing those issues and correcting them. That’s just one of the things that we’ve needed to do to succeed. We’re [almost a] 600-person company with very small carpets. It’s not easy to sweep things under the rug here. As long as we feel the pressure, it’s extremely good for Magma and we’ll continue to develop really excellent technology. From a business point of view, that’s actually our only option."

"And of course, we need to continue to win benchmarks. We need more people to do that. We have over 90 openings right now – we’re continuously trying to bring in good people, but the fact is that it’s difficult. We’re looking for people with specific experience in various areas of work, particularly in research, and often with Ph.D’s. People who have lots of publications are the ones getting noticed and often getting hired."

"But in any case, I’m actually happy that the other big three vendors continue to describe us as a start-up. Our technology is simply better. Perhaps that’s what makes us a start-up. We’re a start-up with over 500 people and run rates of $150 million this year! Do we still give out free soda like a start-up? Yes, in that way it’s still very much a start-up mentality here!"

"It was more fun before in one area, however, I will say. Today suddenly, we have issues of globalization and [responsibilities] as a growing financial organization because of things like Sarbannes-Oxley. You get the idea that corporations now are needing to spend money on all of this because of the failure of Enron and WorldCom – and as a company we’re having to do even more. We’ve always done what we had to do to provide complete reporting, even before [all of these regulations]. Nonetheless today the reporting requirements are [very costly to implement]."

Q – Is it still easy to communicate among people and groups at Magma, like it’s always rumored to be in a start-up?

Rajeev – "We want to retain open communication in the company as much as possible. We enforce this notion quite a bit at meetings and so forth. However, it’s always a problem to keep people candid as a company grows. So on a quarterly basis, we call everyone together and we beat each other up. In fact, we did a restructuring about a quarter ago. We divided the company into four groups, so we could continue to be brutally honest with each other."

"The four new groups include physical verification, design for manufacturing, front-end design, and IC implementation – with R&D shared across the four groups. Now the onus is on the head of each of those groups to make sure we understand the customer requirements in these areas and [to be honest and to speak openly with the senior management at Magma]. All of this is important for scaling the company, both on the technology side and the business side. We haven’t reached the point [where people are afraid to share the truth with the management]. In fact, if we ever reach that point – well, we just won’t!"

"I’m not saying that every one fits into the culture here at Magma, but as long as I’m here, I won’t have it any other way. People must be free to be honest. Greg Walker, Roy Jewell, Saeid Ghafouri, Venktesh Shukla, and I – everyone except for Hamid Savoj – have been CEOs in the past. We all understand the need for open discussion here. When you have four or five people in an organization that have the proven skills to be at the helm at any given point in time, that’s a very powerful position to be in."

Q – Quite candidly, and considering the current global economic picture, would Magma ever consider re-locating corporate headquarters to India?

Rajeev – "That would be impossible. The application knowledge is here. No doubt, there will be a growing number of customers in India, but it will be a long time before India or any other place in the world can reach what Silicon Valley has to offer. I owe a lot to Silicon Valley and this is where Magma belongs."

"There are entrepreneurs who come to me with that kind of model – starting an EDA company in India. But the skillsets [available there] are nowhere near what we need, even in generic software, let alone EDA. Culturally as well, there’s no expertise in India for building start-ups. Maybe there will be 10 to 15 years from now. Right now here in North America, we’re so inwardly bound and worried about losing jobs in the short term – we’re making this a bigger issue than it needs to be. Instead, we should be trying to build businesses. We should be creating business leaders, not fighting a global economy."

Q – How does Magma position itself in new markets while still protecting established markets?

Rajeev "At this stage in Magma, I spend quite a bit of my personal time on these questions. As a company, we’re driven by two things. First of all, we have an extremely good relationship with our customers. A few of them are always doing cutting-edge designs and they give us ideas of where the industry is going. In the old days, [innovation in EDA] was driven by entrepreneurial hunches. That was the driving force. Today, it’s what the customers have to say that drives us."

"The second thing is that we’re pretty much an entrepreneurial friendly company. If you came to us and told us that you had a technology that was 10x faster, for instance, than the existing simulation technology – we would be interested. We’re game to attack each and every opportunity in the various markets."

"With respect to the specific markets you mentioned, FGPAs and structured ASICs – we started a dialog 14 or 15 months ago and then bought Aplus Design Technologies. We made [this move into] the structured ASIC market even before it was being used as a word. People were telling us that SoC design was going to be platform-based and would include standard cells, FPGAs, etc. We saw that the fabric was being put into place [for that style of design]. So we made the strategic investment in Aplus, and looking at the last two quarters, we believe we’re now number one in structured ASICs."

"Now, it’s not our place to tell our customers, ‘Thou shalt use structured ASICs, or standard cells, or FPGAs.’ It’s up to our customers to determine for themselves how to mix and match these technologies. But we have a very strong technical team and, unlike everybody else who say they have a structured ASIC flow, we actually have one."

"So, we have three types of acquisitions. There’s the acquisition aimed at basic raw technology like the Aplus deal. In that instance, we’re not buying the company for the source code, but for its experience. We then implement their talent into our flow development team."

"The second type of acquisition is in the category of infrastructure – acquiring libraries, for instance, to make sure that data is available so that tools from Magma, Cadence, and Synopsys can be used very quickly."

"The third bracket of acquisition is related to what I call revolutionary or world-shattering technology. For instance, Mojave came to us with very impressive technology and an excellent team. It was clear that what they were putting together was really revolutionary. That has been the first and only acquisition we’ve done in this category, but it shows that Magma is willing to play in each of these areas. This third type of acquisition is the most forward-looking. It is not so much driven by what the customers demand as by an incredible technology."

"Cadence likes to buy technology that’s already up and running. We prefer to do it in the early stages of a company. For us, it’s the whole package. Meanwhile, of course, we still think we have a culture at Magma where we can get new tools out."

"Most EDA companies have had one success and then go and buy everything else. We’re certainly going to do what’s necessary [in the area of acquisitions], but we’re also doing a lot of new things. A software company can’t grow purely through mergers and acquisitions. If you think it’s difficult [to integrate designs] in hardware, it’s even more difficult in software. And EDA software is much more complicated than most software. And from the cultural side of an acquisition, it’s even worse [trying to merge companies]. How do we in EDA think we can build a cohesive culture within a software company by just slapping things and organizations together?"

Q – Power and signal integrity are the agonies that accompany the ecstasies of 65 nanometers. Why don't we just stop at 90 nanometers resolve the outstanding problems there, and leave well enough alone? Does it make good business sense to be pushing the envelope?

Rajeev – "We have done lots and lots of designs at 90 nanometers, and 65 nanometers is already here. In our view, by the time we announce our October/November releases of our tools, we will be completely done with 90 nanometers. That doesn’t mean there won’t be more work to do, but fundamentally we’ve done lots of chips at 90 nanometers. We know about the noise, the routing – the entire process is pretty much done. Power was a great learning experience for Magma, and our solutions were pretty much customer driven."

"Sometime between now and February, we’re going to put a public face on a number of different things. We’ll be getting a series of launches out the door, and will have more announcement and releases between now and January/February 2004 than we had in the entire history of the company. Just [having the bandwidth] to be able to communicate all of those developments to Wall Street and to EETimes is quite a challenge of its own."

"In all of this, I’m not saying that at 65 nanometers there won’t be any issues. Clearly at 65 nanometers, yield is going to be an issue – we’ve already announced partnership with PDF Solutions to address these things. But as a start-up that’s going up against two Goliaths – simply put, we’re at a confidence level such that in our guts we believe we’ll meet the expectations placed on us to deliver [solutions at] 65 nanometers. Although we’re not going to talk about it until it’s here, there’s no use shying away from it. We have the confidence to deliver!"


The Epilogue –
with apologies to Turgenev and Hemingway

The four of them stepped down off the stage where they had been sharing the limelight. The Sage One turned quietly to the Upstart. "You don't know the leaders in this industry well enough," he murmured. "You find fault with their point of view, but what makes you think it’s not a product of the very technology that you’re championing?"

"The leaders are good fellows," said the Upstart, "but their day is over. Their song has been sung to extinction. We have to move on. I can see we have to move on."

The men, whose day was over, stood nearby and listened for a minute or two, then quietly returned to their offices. One rang the other on a private line. "So it seems that you and I are behind the times. Our day is over," one said to the other.

The other responded impatiently, "In what way is he so different from us? And how on earth has he gone forward? It's those venture capitalists and analysts who have knocked such ideas into his head. I'm sure that in spite of all his staff and investments, he knows precious little about this industry."

The first one said, "No, brother, you mustn't say that. He’s clever and he knows his subject."

"And so disagreeably conceited," the second one continued.

"Yes," observed the first. "He is conceited. Evidently one can't manage without it, and that's what we’ve failed to take into account. We thought we were doing everything to keep up with the times. We studied the designs. We tried to run model companies. We ourselves have even been described as rebels at times. We’ve read. We’ve listened. We’ve tried in every way to keep abreast of the demands of the day. Yet some say our day is over. I, however, continue to think it’s barely begun."

"Exactly!" the second one replied. "Our day has barely begun!"

The conversation ended and the first one leaned back in his chair. He looked out his window and saw the highway that rose and fell past the last house in the distance. The banks of red dirt on either side of the road were sliced cleanly away and the second-growth timber stood on both sides. Their branches moved in the breeze. A storm was brewing.


July 14, 2005

Peggy Aycinena owns and operates EDA Confidential. She can be reached at peggy@aycinena.com

Copyright (c) 2005, Peggy Aycinena. All rights reserved.