In this interview John
Bruggeman, chief marketing officer at Cadence, answers questions about EDA360 and its potential impact on the
electronics industry. He identifies key takeaways from the EDA360 vision paper
and discusses the motivations and market conditions that are leading to this
new approach to electronic design.
Q: Cadence has done
something very unusual - it has released a visionary document that looks at
broad trends in the electronics industry and then outlines a new approach to
EDA to respond to these challenges. Why?
A: We are undergoing a disruptive transformation in which
semiconductor companies must provide more than just hardware. They now have to
provide some or all of the software stack in order to build application-ready
platforms. At the same time, they're anticipating potential $100 million SoC [system
on chip] development costs. In this environment, the problems our customers are
facing are significant, and working together to solve those problems will have
a huge impact. With the vision paper, we're making a call to the industry to
address these challenges.
Q: What key takeaways
do you hope people will come away with from the vision paper?
A: We want readers to understand the transformation. Systems
companies are now achieving their differentiation and their value from offering
the latest, greatest applications, or "apps." They are demanding that their
semiconductor suppliers build integrated hardware/software platforms ready for
applications development. This is true not just with cell phones, but
everywhere. This leads us to three important points.
First, the EDA and semiconductor industries have until now
focused on design creation. With $100 million development costs on the horizon,
there will be far fewer creators. Many designers will become integrators who
make heavy use of pre-designed IP [intellectual property] to build SoCs and systems.
EDA tools so far have only addressed creators, and this must change.
Second, EDA until now has primarily focused on helping
creators overcome the productivity gap. This work must continue, but what
integrators are most concerned about is a profitability gap. Closing this gap
requires new tools and approaches that can reduce design costs and bring in
Third, design going forward will be driven by apps. People
will start with the applications and then build, or source, highly optimized
hardware/software platforms. The traditional approach, in which hardware is
built first and software and applications are tacked on later, has become too
inefficient and costly. Thus traditional design tools and methodologies must
evolve, and EDA360 will accomplish that.
Q: People have been
talking about top-down, system-level design forever. What's new today?
A: A couple of things. One is the economy. We think we have
stripped out the inefficiencies in the supply chain, but we haven't - we just
pushed the problem down. The second thing is the iPhone. Expectations about
apps are fundamentally different than they were five years ago. I recently
bought a barbeque and there are apps on the barbeque!
Three years ago a "smart phone" had email, a calendar, and
voice. That's all. What was the imperative to do system-level design? Some
smart guy at Berkeley may have thought it was a cool model, but there was no
business imperative. Now we have a business imperative. On the upside, the apps
drive the revenue and the differentiation. On the downside, there's the design
While I used the iPhone as an example, I want to clarify
that I'm not just talking about phones or consumer electronics. Apps are in
anything that uses a processor. At Wind River I worked with a commercial
airplane that had 400 apps. How do you design a control system that makes them
all work in a safe and reliable way? Those apps are critical, and approaches
used until now are hard, costly and inefficient.
Q: Both EDA and
embedded software companies have had a hard time realizing value for their
technology. Will EDA360 help?
A: I think so. Look at it this way. Today, a semiconductor
company creates IP for $1, and then spends $3 integrating it. With EDA360 the
idea is that you will spend only $2 to create and integrate IP. Now the total
cost to the semiconductor company is $2 rather than $4. The question is whether
the EDA vendor can articulate enough value to claim some of that savings.
Everybody should win in this equation.
Q: Who is going to
provide and support EDA360 solutions?
A: No one company can do this alone. It's going to be solved
by a community of companies that step up to solve this problem together. This
includes EDA companies, processor companies, IP companies, OS companies,
Cadence will become a leading EDA360 company. I hope and
expect that my competitors - the traditional EDA players - will see the light
and become EDA360 companies as well.
Q: What is Cadence
doing today to support EDA360?
A: We just announced a collaboration with Wind
River and we announced the Cadence Verification Computing
Platform, called Palladium XP. These announcements fit into the EDA360 System
Realization vision. In coming weeks and months we will make more announcements.
Q: Does existing EDA
technology fit into EDA360?
A: Absolutely. EDA360 is a broader view of what EDA should
be. Without the traditional technology we built as a foundation, we wouldn't
have a leg to stand on. And EDA360 needs to be tightly integrated with what we
had before. All of the design, verification, and implementation work we did
needs to be very tightly tied to System Realization, SoC Realization, and
Silicon Realization, because without design, verification, and implementation,
you can't realize the final product.
Q: The EDA360 vision
paper talks a lot about integrators. Will EDA360 still serve creators?
A: Yes! The creator's job is getting harder, and EDA360 must
continue to serve the creators. In many projects there are both creation and
integration elements. EDA360 has to serve both.
Q: How will EDA360
change the electronics industry?
A: It will truly enable electronics to become what it can
be. There is a mass of potential today, and people are just starting to scratch
the surface. As more and more people think from an applications-driven,
top-down point of view, the world of possibilities is immense.
Larry -- the IP stack concept described in the vision paper is not about "hiding" software IP; it is about providing bare metal software IP that SoC integrators would otherwise have to develop or source on their own. As for OS integration, many systems houses are demanding HW/SW platforms with software stacks including the OS. Somebody has to integrate the SW with the HW, and EDA360 will help do that.
I do NOT agree with you. If you want to integrate WR OS with Cadence design, I don't care.
Just admit it. If you want silicon houses to be able to hide their SW IP, so what (many companies will fight that from SW QC aspect). TI wanted to put a DSP in everything, we see how far that has gotten. The industry wanted to get rid of real design engg in the 80's, new engg have to get a calulator to add 2+2.
Sean -- EDA360 is a high-level vision for the EDA industry, not a Cadence product roadmap. It is based on the challenges that customers are experiencing. The first step was to put the vision out there and to provoke some thought and discussion. Over time we expect to have considerable feedback, as well as customer experience with the first steps Cadence takes towards EDA360. We will present as much of that as we can.
One of the things that Lip Bu Tan stressed at the EDA CEO dinner earlier this year was how much he had been out listening to customers and incorporating their feedback into Cadence strategy. What I am puzzled by in this interview is no mention of any customers by Bruggeman. Which Cadence customers are driving this vision? Can we hear more from them and less "marketing vision"?