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.
IC Manage: A Roadmap out to Forever
September 7th, 2017 by Peggy Aycinena
Dean Drako founded IC Manage in 2003, a company whose products are targeted at IC designers who need help coordinating their efforts, integrating third-party IP into their design equation, and accelerating design. Interestingly, at the same time Drako was founding IC Manage, he was also founding Barracuda Networks, and ran both companies simultaneously for a number of years.
Today 14 years later, Drako still serves as President and CEO of IC Manage, but is ‘only’ on the board of Barracuda. Lest you think his plate is not full enough, however, he’s also currently President and CEO of Eagle Eye Networks.
Prior to our phone call last week, I researched Drako on Wikipedia: “Drako has written a number of articles on Open Source, Big Data, and SoC design. He is a frequently invited speaker on the topic of entrepreneurship [and] is a holder of 27 patents, including patents in network security, network protocols, digital circuits, software, biochemical processes, and sporting equipment.”
Yeah, pretty much just like talking to Elon Musk.
WWJD: Wikipedia has brought me up-to-speed on your history.
Dean Drako: [chuckling] That makes me a little nervous, because anybody can write what they want in those articles.
WWJD: I see there that you’re an advocate for open source?
Dean Drako: Yes, I’m appreciative of open source, and often thought someday when I retire I might just land out on a beach and do something like open-source coding.
WWJD: And for which sporting equipment and which biochemical process do you own patents?
Dean Drako: Wow. I worked long and hard on some snowboarding equipment. I was trying to solve some of the binding problems of how the boots connect to the board. In biochemical, I was doing micro-fluidic research.
WWJD: Wow, so why IC Manage? What was the motivations for founding that company?
Dean Drako: One of the earliest companies I created was Design Acceleration, which provided the first debugging tool for Verilog. We were the very first with an integrated debug environment for Verilog.
Eventually, after close to 10 years, the company was sold to Cadence in 1998, so I’ve always been steeped in EDA. Before that I was a chip designer at Apple. I’ve always been in hardware.
After selling Design Acceleration, I started looking for other EDA opportunities. That was back in 1999/2000 era.
WWJD: The golden years in EDA?
Dean Drako: Perhaps not the golden years, but yes, the more prolific years.
It was then I happened to encounter Shiv Sikand, who had written an open-source tool for doing design management in the Cadence environment, using the Cadence tools.
He was getting a lot of traction [with the tool] and they were asking him for a more commercial version, because they wanted something that was going to be supported. And they wanted someone working on it full-time.
So I spent about 6 months researching and looking at the market for design management, and then I started IC Manage – basically to commercialize a seriously strong design management system for chip design.
At the time, there was a lot of work and effort going into software management systems, but there wasn’t really anyone focused on strong, industrial strength, enterprise-ready management of design.
WWJD: What about Synchronicity, which was active in those years?
Dean Drako: Yes, Synchronicity was around at the time. I remember that I went to their booth at DAC in one of those years, but there were some fatal flaws in their [products]. Something was really wrong.
From a big picture point of view, I concluded it wasn’t going in the right architectural direction and there was a significant opportunity to develop [something better]. Then Synchronicity was sold into the PLM space.
[Synchronicity was acquired by MatrixOne in 2004, which was then acquired by Dassault in 2006.]
WWJD: So you took Shiv Sikand’s open-source entity and commercialized it?
Dean Drako: Kind of, but none of the code we used actually came from the open-source entity.
WWJD: Who wrote the code at IC Manage?
Dean Drako: I hired a team that was led by Shiv and, yeah, they started from scratch.
The implementation and architecture were slightly different. The open source was written in SKILL, a Cadence language, but customers wanted a real tool written in a real language like C or C++, so we had to embark on making something real.
WWJD: When did IC Manage publicly launch its first product
Dean Drako: It was around 2003 or 2004.
WWJD: Not the golden years in EDA.
Dean Drako: No, perhaps a little bit after the golden years.
WWJD: Who is your competition today?
Dean Drako: Well, we still get a tiny bit of competition from Synchronicity, which is now known as Dassault, especially a little bit in Europe and France. Although, none of the original Synchronicity engineers remain.
And, we still encounter some home-grown stuff from folks.
WWJD: Intuitively, it seem the big companies, those with deep pockets, would prefer to have their own internally developed design management systems.
Dean Drako: Yes, over the years they’ve put together some really great systems for design management.
But what happens is, there’s a really great engineer who puts it together for them and makes it all work – writes the code and maintains it – and everything’s all good until that engineer leaves. And then the company is kind of stuck, because they don’t have anyone to take it over and run with it.
So the company says, ‘Alright, what do we do now?’, and then they call us.
The other challenge: There have been a lot of acquisitions and mergers in the semiconductor space, continuing through to today. However, as the companies have consolidated, that internally grown management software can’t deal with all of the variations from the different companies [that have been acquired].
Originally, their internal management tool was alright for the company’s design methodology, but after an acquisition, it didn’t work so well.
But with our global management platform, our tool supports all the different design methodologies, so customers can merge in, and manage, all different kinds of styles.
WWJD: Sometimes I think Mike Fister [Cadence CEO 2004-2008] gets a bum rap for his leadership. Perhaps he was trying to expand the company’s portfolio to service the entire customer enterprise, not just the designers.
Dean Drako: Mike Fister had not yet taken over at Cadence when I was there, so I do not know what he was doing there, or if he got a bum rap. I cannot comment on that.
I do believe there was a period in Cadence’s history where they went heavily into design services, and that is an enterprise kind of effort, because they’re basically saying: ‘We’ll help you design the chips and the products. We’re a design services company for electronics.’
And that was the beginning of expanding the market into IP and design services.
Today, if you look [at the big EDA companies], I would expect that their software tools still make up the lion’s share of the revenue, design services are significant, and IP has continued to grow.
In fact, IC Manage GDP [Global Design Platform] is playing a big role in IP reuse, which is one of the other reasons that these huge global companies [need us]. We give them a platform that all of their design teams can share, internally and externally.
It’s a really difficult problem to make all the IP available to all the different groups [on a design team]. What we’ve found is, the organization is so large they don’t do design reuse because one group doesn’t know what another group in another country has developed, so the [first group] builds it again.
Then it gets really complicated, because you end up with some really interesting security issues. You start to have design groups in some countries where government restrictions exist on who can access the IP, and when and how.
So the problem of an IP management systems becomes a much bigger problem then when you think about it at first blush.
WWJD: Given that IP drives so much of design today, would you call it IP Manage if you started it the company today?
Dean Drako: [laughing] Yeah, I probably would, [but when you start a company, as we did] you’re always looking for names that haven’t been used yet, where there’s a domain name available.
WWJD: How do you handle IC Manage given your other responsibility as CEO of Eagle Eye Networks?
Dean Drako: If you investigated my background, you’ll see I was President and CEO of IC Manage and Barracuda Networks at the same time. I just work really hard, and don’t sleep much.
WWJD: Is there any caffeine in that equation?
Dean Drako: [chuckling] I’m almost 100-percent caffeine free. For me, caffeine makes the roller coaster ride too tumultuous.
WWJD: How are things going at IC Manage?
Dean Drako: Things are going really well.
We’re continuing to enhance the GDP and IP management system. We’ve been heavily investing [based] on our success, reinvesting that success back into the tools. And some of what we’re doing is expanding the EDA ecosystem, which I think is good.
We’ve also been working on Big Data analytics as applied to the design management process.
If you think about it, we have hundreds of engineers working on a design project, but getting all of them to simultaneously finish on the same day so you can tape out – there’s always what we call, the ‘long pole of the tent’. The one piece is not yet finished although everyone else is done.
This can cause project delays of up to 6 weeks, which is 6 weeks less time in your product window to make money. If you’re an iPhone or an Nvidia chip, that can be a really expensive proposition.
So by looking at all of the pieces going into a design, and tracking it with Big Analytics, we can reduce or eliminate the delay, so all the race cars cross the finish line at the same time. Our Envision Design Progress Analytics is getting quite some success with customers [interested] in a whole new tool in the productivity area.
WWJD: Why does everyone always talk about the ‘long pole of the tent’, and not the ‘weak link in the chain’?
Dean Drako: Probably because ‘weak link’ implies that the chain breaks.
WWJD: But the chain does break if you miss your product window because of delays in one part of the team.
Dean Drako: Yeah, but you rarely lose your product window, you just land in the window later. You could lose your product window if you really mess up, but generally you don’t.
WWJD: But the ‘long pole’ analogy also implies poor performance on the part of the team which represents that pole.
Dean Drako: No, not necessarily. It could just imply a resource allocation [issue]. The problem that team is working on is a lot harder than everybody else’s problem.
WWJD: So you’re trying to achieve load balancing?
Dean Drako: Yes, you could certainly call it that.
WWJD: Speaking of ecosystems, why do you participate on the board of ESD Alliance? Why belong to the organization?
Dean Drako: So you know, we used to call it EDAC, and I’m on the board because there are certain topics that the ESD Alliance does a phenomenal job on – things that benefit the industry.
One is educating us about export controls. Every once in a while, the federal government – trying to do the right thing – creates new rules about export, a new list of countries with bad actors we can’t sell to, or indirectly sell to.
But the rules aren’t always clear. So having one group, the ESD Alliance, that can educate makes for a lot of clarity and we can speak with a single voice.
I also believe that the ESD Alliance has proved extremely valuable in defining the platforms the tools will run on and support, forcing [the choices] to be a little narrower and making our job as software developers a little easier.
When I was running Design Acceleration, our tools had to run on 10 different platforms, with 5 versions of each platform. We were testing new versions and bug fixes on something like 50 different platforms. It was a crazy amount of work!
The ESD Alliance has supported and documented the official platforms for many years now, so the user community has narrowed the platforms [they request] and only use the ones sanctioned by the ESD Alliance. That has made it a much easier job for our software developers.
For example, at IC Manage we’re working really hard on a product called PeerCache, an EDA tool accelerator targeted at EDA design teams, closely coupled to design operating systems and coordinated with the customer’s IT department.
It just won’t be possible to develop this product on a whole lot of platforms. Yes, we’ll make the EDA tools run faster, but we won’t be able to do it on 30 different platforms. The ESD Alliance, with its officially supported platforms, allows us to focus on the [most important] platforms.
WWJD: Speaking of export regulations, I recently wrote about the Siemens turbines that were sold to Russia, modified, and installed in Crimea, putting Siemens in direct violation of current EU and U.S. sanctions. Mentor is now a part of Siemens, so their corporate persona is possibly implicated. What do you think about these things?
Dean Drako: It’s true, you often end up in places that you didn’t expect. At Barracuda Networks, for example, we were selling our products to customers in Saudi Arabia, Dubai and India, and that was all perfectly legal.
But someone in there was, in turn, re-selling to someone else, who sold to someone else, and our product ended up in a sanctioned country – somewhere like Iraq, Iran, or even Syria that were forbidden places to sell to at the time. I don’t remember the exact list, but the countries involved were problematic.
Of course, we could sit at Barracuda and say it was not our fault, which was the truth. But that was not good enough and there was a lot of panic around the situation. So much panic, in fact, that we remotely bricked up all the hardware we could that was [potentially] involved.
WWJD: Okay, so throw out a product idea that you’d like to see developed. And by the way, know that someone’s going to steal your idea.
Dean Drako: I love cars and I love driving. I look forward to self-driving cars, so I can be on the phone, doing work, sleeping. The self-driving car will add huge value to society, and we can kind of see it out there.
We’ll also have private Virtual Assistants fairly soon, so my calendar, my meetings, who I know – all that background stuff – will be handled by an Alexa or Siri that knows my voice and my personal world pretty well.
We’ll have 3D-augmented reality kinds of things that will be very useful. It’s already a little useful now.
And, TV as we know it goes away. You know how people build little theater rooms in their houses? Instead, they’ll have a Virtual Reality room. It’ll be screens on all 6 sides, so you get an immersive experience in ‘TV viewing’ rather than just viewing one screen on a wall.
Gloria Nichols: [also on the call] How about air taxis?
Dean Drako: Actually, there’s an energy problem with air travel. It’s takes up 10-to-a-100 times more energy than a wheel. A wheel is amazingly efficient, while holding something up in the air against gravity takes a lot of energy.
As we become more eco-concious, we’re not going to just fly everywhere. The wheel will still be important.
WWJD: Back to the technology, how will IC Manage deal with the post-Moore’s Law world?
Dean Drako: We’ve already dealt with it. We’re not about a particular generation of technology, or the next big chip.
Of course, we do enable the next super big chip, but what we’re really about is design assembly and making it possible for people to say: ‘I need a new chip or new system to do X. I need this piece and that piece, and now I’ll design some other little pieces to glue it all together.’
That’s what our IC Manage design platform will enable forever.
We’ll always have slightly newer chips. Right now, it’s WiFi and wireless chips. Or we’ve got a bunch of guys working on chips for self-driving cars, or speeding up chips, or machine learning, or speech, or voice recognition, or neural nets.
Are any of these guys making these chips from scratch? No, they’re assembling pieces that already exist and [targeting] them at new applications.
Yeah, there will be bigger, faster CPUs, and the nodes may not always be [shrinking] every 18 months, but you’re always going to have to make chips and systems and boards – both digital and analog – for new applications.
We’re not going to run out of things to do.
Apropos to Elon Musk, would it surprise you to know that Dean Drako also has a car company?
“Drako Motors Inc. provides a software platform for electric sports cars to get maximum performance, safety, and cyber security.
The lap-record setting Drako DriveOS operating system for drivetrains is a 4-wheel torque vectoring drive system with a single vehicle control unit (VCU), with proactive millisecond acceleration and deceleration of individual wheels.
The Drako DriveOS delivers unrivaled cornering performance and control, along with stability and traction in diverse road and weather conditions.”
Dean Drako has a BSEE from the University of Michigan, and an MSEE from UC Berkeley.