What Would Joe Do?
Peggy Aycinena is a freelance journalist and Editor of EDA Confidential at www.aycinena.com. She can be reached at peggy at aycinena dot com.
Costello & Carlson: Buckle your seat belt …
January 23rd, 2014 by Peggy Aycinena
Long, long ago in a galaxy far, far away the EDA Empire began and quickly coalesced into several big players and a band of plucky startups constantly attempting to compete and stay viable.
Back in that halcyon era, Rick Carlson and Dave Millman decided to get those startups to pull as one, to try to keep the industry open and progressing, to protect the EDA industry as a place where new ideas could see the light of day and offerings from small companies could compete on a level playing field against those from the big players.
To do that, Rick and Dave came up with the idea for a consortium of Independent Design Automation Companies, IDAC, and put out the word to like-minded colleagues that this new group would benefit everybody. Creating IDAC proved more difficult than they had hoped, so letting pragmatism rule the day they approached Joe Costello for help, then CEO of Cadence, even though that meant working with one of the ‘big guys’ and hence, EDAC came to fruition.
To hear the rest of the story per Rick, recounted in a phone call in December, click here.
To hear the story recounted by Joe Costello, read below. I spoke with both Joe and Rick together on a conference call in mid-January.
Revolution from within …
Joe began: “Rick told me he’s concerned that in his recent conversation with you about the history of EDAC, he may have sounded too harsh. I said that’s not possible, because the truth about the industry is quite harsh. Just thinking about it makes my blood boil.
“Part of my perspective comes from being out of the industry for a while, but even from here I can see that things are fundamentally broken in EDA. The engine of innovation is clearly broken and everybody in the industry should be very worried.
“Also, EDA is part of the larger semiconductor industry and certainly the semi companies are starting to worry. Unfortunately, humans usually let a situation go way past any easy repair, and that has certainly happened in EDA, so now actions to fix things are going to come out of necessity, not out of opportunity.
“If you look at where EDA was when EDAC was started, that was a great time in the industry. Things were on the rise, flourishing in fact, due to a combination of things. Number one was the move to the workstation. New hardware became personal, with high-end compute power and graphics providing new opportunities for design.
“At the same time, we were going through an inflection – design could no longer be done by hand, at least not all of it. At that time, EDA tools were primarily provided by internal CAD teams, but researchers and students were becoming totally excited about opportunities for innovation, and investments in universities promoted robust motivation for new, revolutionary ideas.
“People were beginning to piece things together – ideas from academia and industry CAD teams – combine them with workstation hardware, and boom. You saw the great flourishing of the EDA industry, starting at the end of the 1970s and on into the 80s and 90s. As a result, semiconductors were on the rise, people were pulling electronics off boards and onto chips, and for over 20 years EDA was the vibrant industry pushing it all.”
Joe’s tone changed: “Today, unfortunately, it’s a very different world. EDA has settled into an ossified state, stodgy and not very fluid. Instead of having internal flows being put together by the design team – pulling tools and ideas from a variety of sources, everything from academia to startups to internal CAD teams – now it’s just a flow from Synopsys or a flow from Cadence. That’s pretty much it. Period.
“The current all-you-can-eat deals from one company or the other means design teams have to make do with these solutions, even if the designers have to wrap a lot of shit around the tools to make them work. Today, internal teams are not so much writing their own tools, or choosing best-in-class tools from small providers, instead they’re writing complex wrappers to make things work from one of the two big suppliers.
“EDA has devolved into 2 flows – at best 2.2 flows, if you spice things up with an occasional point tool from Mentor – gigantic, relatively fragile flows that are not at all receptive to modification or change. And as far as the sales process is concerned, essentially the one company just uses the other for the sake of negotiations.
“As a result, the industry ecosystem has become really barren, venture investment has completely dried up, and there is zero motivation for youth to enter EDA because if they want a job in the industry, they can only pick from one of three companies – and only two if they want to work on the flows. Essentially, the tap has been shut and nobody young is coming into EDA.”
Joe became even more grim: “And by the way, all of this is happening against a backdrop where things are only getting harder. There’s way more physics involved as we work to get down to 14 nanometers, there’s more software interaction affecting hardware development, and yet we don’t have any venture investment creating more companies – companies led by brilliant young professors and students to deal with the challenges.
“The long-term ramifications of all of this means nothing will get better. In fact, all signs point at things only getting worse. Perhaps not over the next year or two, but certainly over the next 3-to-7 years.”
“And by the way,” Joe added, “the entire industry is complicit in creating this disaster, customers included, and they will have to be part of the solution of turning this thing around. They have to want to overcome the formidable barriers thrown up by the two existing solution providers – technical barriers and business barriers – to force an opening up, to allow new ideas and new companies to be created.
“Those business barriers, those all-you-can-eat deals I mentioned earlier, they’re completely ridiculous. The two big providers are offering 99-percent discounts on the accessory tools shipped with their [principle product offering], or charging as little as five thousand bucks for the tools. No startup can afford to sell their products at that cost, so they can’t sell their tools and they can’t form new companies.
“The big companies have thrown up technical barriers and pricing barriers making it impossible for innovation to happen. This is, again, where the customers are complicit. They’re not doing anything to force the big guys to open up new technology.
“Instead, the customers send in hard-ass buying guys, their nastiest buying agents, who jam as hard as they can to get the prices as low as possible using the threat of going to the other guy to force their trading partner to push the prices to ridiculous lows.”
Joe took in a frustrated breath, and I said, “You’re telling me this is how the industry looks, and it looks scary, so what’s the answer? You’ve got a big problem, so what’s the solution?”
Joe responded, “There may be a better world opening things up a little bit, because I don’t think this can be solved just by EDA companies. EDA doesn’t exist in a vacuum – the customers are part of the ecosystem and they have to lay down the law.
“They’ve been in this commodity price-crunching mode for so long – squeezing people to get the best deal, penny wise and pound foolish – that they’ve developed an addiction to low-priced tools, and a co-dependent relationship has developed between the customers and the big tool suppliers which has helped create this whole scenario.”
“So you’re advocating for a complete change of course for the industry,” I asked, “the kind of radical change that usually involves violence and the shedding of blood?”
Joe said, “Yes, people with great power hang on for dear life and, yeah, there can be violence in removing them. But there is another human analogy that applies in this case. Does the EDA industry want to use up all of their critical resources for short-terms gains, use up everything [including the brain trust and the good will of investors] on the short term?
“They can do that up until another crisis causes the kind of violent ripping you’re talking about. Yes, sometimes when civilizations hit the wall they shatter exiting structures, but there’s another kind of ripping that tears things asunder and does not come through internal machinations. That’s what we have here.
“In the case of evolving EDA back to a viable industry, we need to look at where all of the chips are going to come from over the next 10 years. If you look at that, you’ll see what cool things on the horizon are going to fuel the market – the rise of mobile and portable devices. Certainly we already know there’s been a sea change at companies like Intel, Qualcomm, Broadcom, and Samsung, and it’s those guys who will dominate the market.”
“You’re suggesting,” I said, “that this new brand of customer will somehow change things?”
Joe said, “Yes, the big EDA customers today need to take stock and think about change. I’ve seen the beginning of that kind of change at Intel. They’re starting to get scared and are realizing the way they’ve played the game for the last decade is completely deforesting things [in EDA innovation], depleting the good minerals in the farmland, polluting the water and the air. So much so, that now it’s more than a quality of life issue in EDA, it’s a matter of life and death.”
“If the new brand of customers are unhappy with the EDA establishment,” I asked, “won’t they return to internal tool development to meet their needs?”
Joe said, “Yes, that could be part of the thing that creates the solution – if they say they can’t wait any more for the third-party providers. It’s true there are smart guys in the customer companies, but I still think it’s outside people, dedicated and focused, who will solve the tools problems better in the end.
“I still believe in an independent EDA industry for sure, but the customers must require that the flows be open instantly. When they do, an important door will be opened that will allow small, innovative providers to once again easily add their tools to the flow.
“And by the way, that’s important from multiple points of view. First, it forces openness. And second, EDA is a budget-line item today. Instead of the conversation at the customer being about building their own tools, which will cause management to ask why it’s being done that way, the conversation will be with the providers.
“When the customers say – Holy Shit! There’s a design problem here that can’t be ignored and you’ve got to work with everybody to provide the solution! – all parts of the EDA ecosystem will be able to respond.”
“When you say open the flows,” I asked, “are you advocating for open source EDA?”
Joe said, “Well, that is possible. But when things have been really successful in open source, they’ve been inspired by a genius who has absolute passion for a technology. Certainly that could happen, but there’s not too much motivation for a genius to move into this world.”
“You mean like Linux,” I asked, “and its move to widespread adoption, despite protests from the IT guys in the organization who didn’t want engineers bringing their Linux OS in from home?”
Joe said, “Linux is a perfect example of the power of the customer to change an industry. In EDA, the customers can demand changes like that. In fact, they’re more than half responsible for the state of the EDA industry – they just don’t realize the power they have to force changes, the kind of changes demonstrated by the Linux example. However, again, I don’t see one guy causing that level of change in EDA.
“There are so many complex technologies involved in this space, it’s more likely that the big customer companies – the Intels, Qualcomms, Broadcoms, etc. – are going to push for change in EDA. The customers are going to claw away at the current leaders in EDA, push and prod and break the glass to see change happen. They will push until the old system is ripped down.”
Rick Carlson added, “We started out this conversation about the history of EDAC. Dave Millman and I were there at the beginning and realized then that a small group of independent companies could possibly change the world. It took us a few hours over several sessions to figure out how to make it work, and then we knew the big companies really needed to be in the conversation and Joe joined us.
“But from the beginning, there was been a lack of openness about innovation and interoperability among the other players in EDA. In fact, Alain Hanover in 1996 shut down the EDAC interoperability lab because he said the EDA companies could handle that internally. Look how that’s worked out!
“Increasingly over the years, there has been less and less open dialog within the EDA industry. Instead it’s always a small select group who have decided everything.”
Joe added, adamantly, “You need leadership to change things. Somebody has to stand up and take the lead!
“Yeah, revolutions are about breaking glass and shedding blood, but they are also about someone articulating a viewpoint, standing up for it and giving people an idea to rally around. It happens in political revolutions – in any crisis – a leader emerges to take control, so that’s certainly something to watch for. You are going to see a different kind of leadership emerge on the customers’ side.
“It’s like the thing they ask you every election cycle: Do you feel better than you did 4 years ago? Similarly, people are starting to ask the EDA industry, and the customers of the EDA industry: Do you feel better off than you did 4 years ago? Is your anxiety about design and the solutions being offered less or more?”
Joe took a deep breath and concluded: “We’ve had a very wide ranging and important conversation here, so let me summarize a few of the key take-aways.
“First, some of the critical foundation elements of a previously rich EDA ecosystem are dying or dead, and the number one red flag here is that there is no outside investment going into EDA anymore. When all outside investment stops, you know you have a big problem!
“Currently in EDA, there are no IPOs, no investment banker interest, no VC investment at all, angels investment is almost gone, and even the customers aren’t investing – other than paying the all-you-can-eat annual fees.
“When there’s no outside investment in an industry, the other parts wither and die. Young, hot graduates won’t be coming to EDA, the talent pipeline starts drying up, there’s less academic research in the area, and the overall idea pipeline simply dries up.
“Again, let’s ask why professional investors won’t put money in. It’s because it’s too hard to start a successful company in this atmosphere.
“It has always required a reasonable amount of money to get an EDA company started, but even after you’ve done that and created a successful product, it is extremely difficult to sell at the beginning. And when the EDA flows are dominated by two companies and are closed, it’s even harder to get your new tool working with them. It’s a huge effort and very few customers are willing to do it.
“However, even if you can get your tool integrated, there are still two pricing barriers. One, the all-you-can-eat programs offer greater than 99-percent discounts to the customers on the smaller tools in their flows. If your startup tool lists at $100,000 but you want to compete, you would have to sell your new product for only $1000. That, coupled with the fact that all customer buying goes through hard-nosed purchasing agents who drive those discounts, means you aren’t going to see the kind of money you need to support your startup.
“And, even if do you manage to build your company, the revenue levels and growth again are like to be limited from above, you can’t go public, and you’ll have to be acquired by one of the 2.5 companies that have been busy trying to kill you.”
“Okay,” I said. “Again, what is the solution?”
Joe responded, “How do you change all of this? In the end, the customers are the ones who will suffer, because they won’t get the right level of brain power needed to solve their problems. They can try do to it themselves, as in the past, but that doesn’t make a lot of sense.
“The truth is, the customers are going to have to force changes in their suppliers to open up this world again to innovation. The customers need to force the flows to be open, so new companies can again play, and they need to pay real prices, instead of all-you-can-eat, for tool innovation to encourage new technologies.
“They need to invest in EDA innovation, instead of just sending in purchasing people hell-bent on lowering the cost. This is what will be required to get the industry healthy enough to get outside investment rolling in again.
“Here’s the message to the customers: If you don’t force these changes, you will pay the price in a different way over the next few years. There will be inadequate technologies to handle your designs, and there will be no quick solutions available.
“It takes a long time to fix this kind of ecosystem and the problems that plague it today. There’s isn’t a lot of time to make these changes. The customers have to insist they happen now!”
It was obvious from our several lengthy phone calls that Joe Costello and Rick Carlson, and other restless seekers, are working to change the moribund state of affairs in the EDA industry. A revolution is clearly coming, and these guys are hinting broadly that they plan to drive it. They plan to be part of the Brave New World that will emerge from it.
Buckle your seat belt, it’s going to be a bumpy ride.
2 Responses to “Costello & Carlson: Buckle your seat belt …”