Here are my thoughts on what were for me, the top three sessions at DVCon.
“Techononmics of Verification” by Aart de Geus.
If you have ever attended one of Aart’s talks, you already know he is a fanastic speaker. Through amusing graphics and wry humor he showed how global practices and financial instruments created a “vicious system” with way too much feedback. Now the output is headed for the negative rail and we are all looking at a “standard of living reset”. It all seems so clear in retrospect.
Then there was the panel session titled “EDA: Dead or Alive?” moderated by Peggy Aycinena. This was a panel of seven EDA CEOs and VPs. Peggy did a excellent job moderating the group and forcing upon them a level of discipline they should all take back to their offices.
Given that each of the panelists had a personal stake in the EDA industry, it was not surprising that they all felt it would survive the current economic downturn. If they didn’t, they would not have been there. Still, not all was unbridled optimism. On the positive side they were saying things like “the electronics industry is growing so EDA will grow”. On the negative side I heard “There are no cash cows”, “zero visibility”, and “death is part of life”. The later is true only until it is your death.
The conclusion I came away with was that EDA will survive, but many EDA companies will not.
Finally, there was a very interesting panel on “SaaS and Cloud Computing for EDA”. This panel was assembled by Harry Gries (the ASIC Guy). SaaS is Software as a Service and refers to a computing/business model where the software runs on the providers computers and customers interact through a web browser or other network interface. This is a model that has worked well in a few other industries but has failed when previously tried in EDA.
The technology is well suited to batch mode applications or applications that require primarily text input. Interactive graphic tools such as schematic capture or PCB layout suffer from mouse to cursor latency. Applications that are highly compute intensive and can be spread across many CPUs could receive the most benefit because adding short term compute power in the “cloud” is both cheap and easy.
The biggest draw backs to EDA SaaS are vendor licensing policies and data security. There are technical fixes for the later but, the former requires a new mindset on the part of some of our favorite EDA partners.
Thanks for the DVCon report. I didn’t make the trip this year.
>>Interactive graphic tools such as schematic capture or PCB layout suffer from mouse to cursor latency.
Are you thinking that the mouse position is sent to the server in real-time? I could see latency being an issue for VPN-based SaaS for schematic capture and PCB layout. However, given what Google maps accomplishes with just a web browser, I think it’s possible now to have such a tool on the web. The key is to use the browser’s javascript engine to the max. This requires a completely different client / server architecture that EDA tools are generally not constructed for.
I would like to see a Google maps style web based simulator / waveform viewer… I’d use it! It’s gotta be simple and convenient though. A web API would be needed to replace the existing TCL environment. This might make it possible to use many different programming languages rather than getting stuck with TCL. Hey boss, I’ll send you the link to my regression simulations…
“EDA will survive, but many EDA companies will not.”
I agree with your assessment, but I don’t think it’s a new situation. What’s been interesting about the last five or so years is that even venture backed firms have ceased to exist without anyone buying the assets.
It’s an industry characterized by a lot of experimentation and innovation: many startups fail within a few years. It’s not a new condition.
W. Edwards Deming had two quotes that are think are relevant to the current challenges EDA firms face:
“Learning is not compulsory…neither is survival.”
“No one has to change. Survival is optional.”
Hello Rick,
>>Interactive graphic tools such as schematic capture or PCB layout suffer from mouse to cursor latency.
We started testing Virtuoso from Cadence Design Systems in a Xuropa Online Lab (www.xuropa.com) last week and didn’t see any of the latency issues you mention. The use model was a workshop where one person was driving the tool remotely while two other “instructors” were observing and coaching remotely (three different locations in total). The observations were of “impressive performance” with no latency experienced.
Unfortunately this lab is closed to the public at this time. However you could go to one of the Incisive-based verification labs and ask for access to try that out.
[Side note: spot on Sean!]
– James