Cloud Computing a Hot Topic at Design Automation Conference

Maybe it was just me, but the energy level at the 47th Design Automation Conference was higher than it’s been in a number of years.  It could have been a combination of things - general signs of an exit from The Great Recession; the M&A activity in the industry; a new market approach from Cadence got everyone a-buzz; and everywhere I went I heard discussions about Cloud Computing (and it wasn’t just because I was stood there ;-) ).

Bernie Meyerson, vice president of innovation at IBM, made a keynote speech on Wednesday that was wide ranging, but spent a great deal of time asserting that Cloud Computing is the future of IC Design.  Richard Goering wrote this up nicely on his blog here.  You should take a look.  The full keynote video is here, pick up what Dr. Meyerson had to say about the Cloud at around minute 38:00.

A crucial point to emphasize from Dr. Meyerson’s speech is the real issue at the heart of the computing challenge facing all industries, not just electronic design - IT resource overhead.  While it is tempting, as engineers, to focus on the technology of cloud computing (performance, upload time, latency, security, etc.) it is this business aspect of the equation that is the driver.  ie. The Total Cost of Ownership of data centers is out of control - driving a company’s balance sheet in the wrong direction.

If you missed the panel “Does IC Design Have a Future in the Clouds?“, don’t worry.  It was videoed and should be coming online soon.  For me, it was great fun to participate with a tremendous amount of interaction with the audience.  To the point where the Chair (Raul Camposano) had to cut off questions from the floor.  (He almost cut of Harry “The ASIC Guy“, but he was saved by the crowd.)  Richard Goering mentioned on his blog that he will put up a post about the panel soon, but in the meantime, you can catch a write-up of some of the highlights over at EETimes by Nicolas Mokhoff here.

If you were at the show, what were your impressions?  If you weren’t able to make it - did the various online channels get you what you needed?
Connect with me:

Posted under Xuropa, industry

This post was written by James Colgan on June 18, 2010

Tags: , , ,

Cloud Computing and Electronic Design Presentation: EDPS 2010

Last week I had the honor and the pleasure of presenting Cloud Computing and Electronic Design at the Electronic Design Processes Symposium in Monterey.

It was interesting to find that many of the members of the audience have used or are using Cloud Computing in some capacity within their work.  Not surprisingly, most implementations are still ad hoc or experimental.  However, there were a significant number of discussions leading to more formal adoption of Cloud Computing.  We’re definitely moving…

Here is the presentation I made.  Is your company using or looking at using Cloud Computing?

Posted under Xuropa

This post was written by James Colgan on April 12, 2010

Tags: , , , , ,

Cloud Computing Presentation at EDP Symposium

Monterey Bay by The City of Monterey

Monterey Bay by "The City of Monterey"

For the first time, Cloud Computing is to be a topic of discussion at the Electronic Design Processes (EDP) Symposium Workshop this year, and I have the honor of presenting the subject.  It promises to be an interesting discussion as remote “utility” or “grid” computing may well have been tried in the nineties by some of the audience members.

Sponsored by The Design Automation Technical Committee (DATC) and the IEEE Computer Society, the event will be held in Monterey from April 8-9, 2010.  Cloud Computing is included in the “Brave New World” session starting at 4:30pm on the 8th.

Event Information

Registration

Posted under industry

EDA: Alive

//www.ecoscapes.net)

Courtesy EcoScapes (http://www.ecoscapes.net)

Last year was my first time attending DVCon. So, that was the first time I got to witness the infamous EDA Troublemakers Panel. Last year Peggy Aycinena channeled her inner John Cooley with a provocative panel discussion entitled “EDA: Dead or Alive”. Of course, the EDA world looked very bleak back then, just months after the financial meltdown, Cadence financial woes, EDA layoffs left and right, VC funds shutting down … a forest fire of bad news sweeping the industry.

But just as the wildflowers bloom after the wildfire, optimism is returning to the EDA industry. (Gosh, that sounds really corny, huh). So it’s not too surprising that two EDA veterans will be at DVCon this year heralding EDA’s return. Fellow EDA blogger Paul McLellan and EDA Hall-of-Famer Jim Hogan have become the Captain and Tennille of EDA. They spoke at ICCAD on the future of chip design. And now they are giving another talk during DVCon entitled “So you want to start up an EDA company? Here’s how…”.

I don’t know what these guys are going to say, but if I were them, here are a few things that I’d point out that make for some optimism for starting up an EDA company now:

  • There are lots of good EDA people looking for work, so this is a good time to partner with them or to get them on board for a lower salary than you’d have to pay otherwise.
  • Small companies are able to use technology to their benefit to lower their cost of development. Development hardware is cheap and even rent-able through cloud computing providers. Open source software tools abound. The biggest cost will be your time.
  • The cost of sales is dropping due to new technology. WebEx. GoToMeeting. Skype. Xuropa. These tools and others enable you to reach out to customers globally while never leaving your office.
  • Then again, who needs an office? Working from home has never been easier.
  • And who cares where you are located? It’s easy enough to collaborate with your virtual team globally. (Xuropa has people in the Bay Area, the LA Area, and France).
  • The lack of VC money can also be a good thing. You retain ownership and control of the company and don’t have a “timetable” to “cash out”.
  • The lack of VC funding also forces you to think about getting cash flow positive as quickly as possible. Rather than wasting time and money designing the perfect product, you’re forced to start selling it early which leads to better feedback and a better product.

The idea that the downturn is an opportunity is not new. In fact, Sean Murphy has been spearheading the whole concept of bootstrapping for some time now. Check out his site for some great advice. Also, Dharmesh Shah, founder of Hubspot, writes a great blog on this subject as well.

I’m still not sure if I’ll make it to DVCon this year, but if I do, I will definitely attend this session. If not, I’m sure someone from Xuropa will be there. Just like my EDA SaaS and Cloud Computing Roundtable last year, it’s not part of the regular program. It will be held Tuesday evening Feb 23rd from 6:30 – 7:30 in the Oak Ballroom at the Doubletree Hotel (where DVCon is going on). This will be right after the DVCon reception, so grab a beer or a glass of wine and head on over.

I hope to see you there.

Posted under Xuropa, business, industry, marketing

This post was written by harrygries on February 8, 2010

Tags: , , , , , , ,

Figure Skating, Scott Hamilton, and EDA Demos

Scott Hamilton

Scott Hamilton

With the 2010 Winter Olympics approaching, there has been some talk recently about those sports that Americans talk about once every 4 years. One of them is figure skating.

Now, I’m not much of a figure skating fan myself, but my mother-in-law is. And all this talk reminded me of the time we took her to an ice skating show at the LA Forum. It was not just any old show … it was Scott Hamilton’s return to figure skating after a fight with testicular cancer. Scott was as close to a superstar as there was in men’s figure skating, having invented the back-flip and won the Gold medal in the 1984 Olympics. So seeing him back on the ice was a feel good story.

The show featured a lot of other well known figure skaters, including Dorothy Hamill and Katerina Witt, but the climax was towards the end when Scott Hamilton took the ice. Cheers went through the crowd as he performed move after move, if not flawlessly, at least without major error. Then came the moment of truth. Scott would now perform his famous back-flip.

The lights dimmed. A lone spotlight. A drum roll. Scott Hamilton took a deep breath, began his run, turned backwards, took off ……. and fell :-(

Obviously, it was disappointing. This show was being recorded and going to be shown a week later on Network TV, so unfortunately it would not be as great a moment as we’d anticipated.

When the show ended we started to gather our things when an announcement came over the PA that went something like this: “we’d like to do some extra takes of certain parts of the show. If you would like to stay, please come over to the west side of the forum and fill in the seats over there.”

What happened next I should have anticipated, but for some reason I was naive. The show producers proceeded to repeat parts of the performance where skaters had fallen or stumbled.  And of course, that included Scott Hamilton’s back-flip. Time after time the lights dimmed, the spotlight fell, and Scott skated and tried his back-flip. Finally, I think it was after 4 or 5 tries, Scott Hamilton nailed it and a roar went through the crowd.

I made a point of watching the performance when it aired on TV a week or so later. And sure enough, nobody fell or even slipped up. And of course, Scott Hamilton successfully executed his back-flip on his first try, to the cheers of a huge crowd.

Having been in the EDA business for many years, I know that a lot of EDA tool demos are a lot like Scott Hamilton’s return performance. If features don’t quite work, the demo avoids those features. Or if the feature is critical, then that is the one that gets fixed while other not-so-critical features may be left broken. It’s part of the smoke and mirrors that is the least well kept secret of EDA tool demos. The customer knows that the EDA company is avoiding the holes in its product.

But what if you don’t have big holes in your product? What if your tool really can nail that back-flip on the first pass? What if your tool really is a “game changer”? Won’t they believe you if you show them? After all, seeing is believing!

Unfortunately, customers have become so cynical and jaded about EDA tools and EDA salespeople that they hardly ever believe what they see anymore. I know. I’m also one of those customers. Read ESNUG or any one of the many EDA forums and you’ll know that I’m right.

But what if they could try the tool out themselves? No strings attached. That’s what we are trying to do with the Xuropa labs. To give your jaded, non-trusting, cynical customer a chance to try your fabulous tools himself.

“Seeing is believing” is no longer good enough. “Doing is believing” is the new reality.

If you’d like to see what a Xuropa Online Lab is like, you can try it out here. We’ve got some tools loaded and you can play with them as long as you like. No strings attached.

Oh, and to be fair to Scott Hamilton, here’s a video of him nailing that same back-flip just a few weeks ago. In this case, seeing is believing.

Posted under Xuropa, business, industry, marketing

This post was written by harrygries on January 17, 2010

Tags: , , , ,

Your Own Private E Ticket Ride

We took my daughter to Disneyland this past Friday for her birthday. I know that she was playing hookie but please don’t call the cops. Her birthday was actually on Saturday but we decided to go on Friday because it would be less crowded and she’d get to go on a lot more rides. We had a great day and arrived home around 11pm dog tired and with sore feet.

On the way home, as the kids slept in the backseat, Joyce and I figured out that we were able to go on 12 attractions in about 7 hours at the park. Not bad. If we had gone on the weekend, no way we’d have been able to do as much, not to mention the frustration of waiting in long slow moving lines. Of course, Disneyland doesn’t want to frustrate its visitors or have them wait in long lines, but it can’t just add more rides and attractions on busy days and get rid of them on slow days.

Back in the 1970s, Disneyland tried to manage this issue by providing each visitor with a coupon book that contained a limited number of A, B, C, D, and E tickets, each type of ticket good for a different class of ride. The E Ticket was the most valuable, good for the most popular rides and attractions. In this way, Disney knew exactly how many visits there would be each day to each class of attraction.

Disnet E Tickey Circa 1970s

Disnet E Tickey Circa 1970s

Which got me thinking that most corporate design centers are managed like Disneyland used to do it, IT departments handing out A - E tickets to its visitors via LSF in order to prioritize the use of limited CPUs and licenses. Like Disney, they can’t just easily drive up a truck of CPUs and licenses on busy days and send them back on slow days. In fact, it’s even harder for IT departments because you never know when a busy day or hour might come or what “attraction” will be the most popular.

But some of that may be changing.

Licensing has become much more flexible over the last few years. Cadence and Synopsys offer short-term licenses that allow customers to peak their license usage almost on demand for periods as short as a week. Cadence offers a license for mixed-mode simulation that allows customers to trade several A tickets for their basic analog simulator for an E ticket for their more advanced and expensive mixed-mode simulator. And I’m sure there are several other examples of which I am not aware.

Hardware is a lot more difficult. Whereas a peak license can be downloaded over the internet, hardware can’t be shipped and installed overnight. Or can it?

Cloud computing has been all the rage the last year or so and as it goes with rages there has been a lot of hype, misconceptions, and promises. And as is always the case, there emerges the first practical application of the over-hyped technology. We’re seeing the industry start to focus on so-called hybrid cloud environments that allow companies with captive private data centers to virtually provision additional peak resources from public clouds like Amazon EC2. There is a HUGE potential cost savings for companies operating large data centers. The Economist quotes a study by McKinsey that estimates “on average only 6% of data center capacity is used” since data centers are scaled for near peak capacity needs, not average needs. That’s like building Disneyland to handle the busiest day of the year and letting the rides sit idle the rest of the year. Good for the visitors but hardly cost effective.

With a hybrid cloud, data centers can scale to meet the needs of the users. For IT and CAD groups, adding additional CPUs and licenses is kinda like adding additional Splash Mountain ride capacity on a hot day.  For designers, it’s kinda like having your own private E Ticket ride whenever you want it. And for EDA companies, it’s a way to keep the IT, CAD, and design guys from getting frustrated.

No waiting in LSF queues.

No getting cut off in line by the VIP project.

Ride as much as you like whenever you like.

Surely, The Happiest Place on Earth!

Posted under Xuropa, industry

This post was written by harrygries on September 20, 2009

Tags: ,

Experience and Education in EDA

When I was in 11th grade, my Social Studies teacher, Mr. Lewis, took us on a field trip from Brooklyn to Lower Manhattan. That was the first time I ate “real” Chinese food and it was also my first time in a “real” bookstore, The Strand.

Back then, circa 1980, it was very common for people to go browsing in used bookstores, and I caught the bug real bad. Not only was it an affordable way for a kid to buy books, but there was also something exciting and addicting about the serendipity of the process, not the least of which was the intoxicating fragrance of those weathered and worn volumes.  It was like a treasure hunt.

I never went in looking for any one book in particular, but I always walked out with half a dozen or more books on totally different subjects. In fact, if you were to look at my bookshelf today, over 80% of the books were purchased as “previously owned”. That’s one of the great things about books … even when they are used, they still retain 100% of their value!

John Dewey on Education

One of the books I’ve picked up over the years is “John Dewey on Education”, which is a compendium of excerpts of his various writings. For those of you who are unfamiliar, John Dewey was perhaps the most influential educational reformer of the late 19th and early 20th century, best known for his writings on the role of experience in education. In fact, several excerpts in this book are from Dewey’s other book, “Experience and Education“.

In this book, Dewey proposes 2 important ideas. First, that real learning only comes by connecting ideas with experience. Second, that every learning experience influences future learning experiences. For example, consider a child that touches a hot stove. Through that experience, he first learns that stoves are hot and not to be touched. A valuable lesson, indeed. Second, this experience teaches him to be cautious about touching objects he has not touched before. In that sense, it actually dampens his enthusiasm and curiosity for future experimentation, which can be harmful to future learning experiences. Not so good. As can be easily reasoned, the best learning happens through experiences that teach and also encourage future learning.

__________

Last week I came across a blog post by Jim Lipman entitled “So this is where Engineers get their Information“. It seems that Jim attended a seminar by eg3.com on how designers search on the internet. One portion of the blog post caught my attention:

In the “What do Designers Want” category, ‘hands-on’ items such as demos, software and evaluation kits rank very highly. On the flip side, vendor articles and webinars ranked moderately low and podcasts very low in response to a “What Information do you seek” question (the webinar response probably due to the same perception that I have that many are sales or marketing tools).

In other words, designers want to be educated through actual experience with the tools, not by listening to webinars or presentations or even podcasts. If you have spent any time around a teenager, you know that this tendency towards interactive experiences will only increase. No longer will designers be satisfied being passive receivers of information. Rather they will prefer, no, demand to be active participants.

Second, their experience accessing these forms of online learning will have an effect on their future learning habits. If they have to log in through several screens and enter personal information just to get access to a Webinar, fagetaboutit! They won’t come back. But if the process is simple and easy and the experience is smooth, they will come back.

That’s what we are trying to create at Xuropa. An environment where designers can learn by experience, by using your tools hands on and where they will come back again. We’re trying to make the process painless, just a simple password login and there they are using your tools in seconds. It’s not just another form of WebEx or VNC or any other screen sharing program as some have suggested, but a way for the designers to drive the tools themselves and learn first hand.

We’re just getting rolling and we’ve got a lot to learn by the experience as well. The process, although smooth, can still be improved. And we are working hard to get more valuable tools in place for designers to try out. If you are a designer reading this, go ahead and try out one of our labs and let us know what you liked and how we can improve. We really want to make this painless and rewarding for you. If you are in the EDA world, consider putting your tool in the hands of a customer and providing a learning experience whereby they fall in love with your tool.

It’s all about the Experience.

Posted under Xuropa, industry, marketing

This post was written by harrygries on September 2, 2009

Tags: , , , ,

MaaS: Music-as-a-Service

My favorite source of online music is Pandora and their Music
Genome Project.  If you’re not familiar with them, it’s a web based service that streams high quality music to your browser through a flash applet.  Just the idea of working out my musical “genetic” make-up was cool to me, but it was the exploratory nature of it that really got me hooked.

Essentially, you first of all seed your stream with some of your favorite tracks, musicians, or bands.  From that Pandora’s algorithm starts to stream to you similar music with the bands you selected sprinkled in.  The genetic part of it is when you see common threads across music that you may not have listened to or heard of otherwise…but you like.  For example, seeding with 80’s alternative music (New Order, Joy Division, Spear of Destiny, etc.) got me to modern era Killers, The Strokes, and a few other bands that I either hadn’t heard of, or only peripherally.  But you can tell their “musical genetic makeup” is very similar…and I liked it!

So - a cool app that many people love, but their business model was broken.  There they were delivering tons of value to users, but they were trying to monetize that value through advertising.  Of course this wasn’t working for them, especially as they have to pay royalties on the music they stream.  So today I noticed that they have moved over to a subscription, or “freemium” model.  This makes complete sense.  You get 35 hours free per month, and then you have to pay $36 per year for better audio quality and no advertising (there are other premium features, but who really cares about “skins”?)    There we have it - Music-as-a-Service!  MaaS!!

Now that their platform is built, their greatest costs are licensing royalties that I’m assuming they pay after the fact.  So charging “power users” up-front will put them in a good position to fund the company and the royalties.  I hope they make it (there were rumours that they were struggling hard and may even shut down).

Now, why $36?

SaaS (Software-as-a-Service) is all about charging for value as that value is consumed…or as near to this as is reasonable to expect, ie. monthly or annually (with discount).  And here lie the challenges for companies looking to move to a SaaS model:

  1. Is there a free (or relatively cheap) version of the product with a reduced feature-set that provides some value without giving up the farm?
  2. Can the lower priced version of the software be delivered and supported in a low cost manner?
  3. Is there a clear and natural path to upgrade to the premium product that will make up the core of your business?
  4. What are the features of your software that provide 80% of the value…and what is their value to the end-user?
  5. What is that value, what is it worth, and how should it be charged?

Within the electronic design tool industry segment, business models and use models are essentially orthogonal today.  SaaS provides a way to bring these models more in line. 

But as I’ve argued before, unlike the current license-based model, it’s not a simple equation for electronic design tools with a one-size-fits all approach.  Not only do different tools deliver different levels of value, but also the value is consumed in different ways - ie. use models differ depending upon the tool being used and where in the chip design process the tool is being used (verification and simulation tools are a good example of this variance over time).

Fortunately, a SaaS model not only provides a low cost delivery and support vehicle, it also provides you with visibility into the use of your product and the mechanisms to adapt relatively quickly to market conditions and learnings gained. 

There-in lies the opportunity to better serve your customers, and your investors…or maybe GET customers and investors!

Posted under business, industry

This post was written by James Colgan on August 25, 2009

Tags: , , , ,

Blooming of an EDA Revolution

(Photo courtesy Yankee November)

Hi! This is Harry Gries (aka the ASIC guy). And this is my first blog post at Xuropa.

It’s been almost a year since I made some bold predictions on my other blog concerning the impact of Cloud Computing and Software-as-a-Service on the EDA industry. So, I thought it would be interesting to revisit that blog post and see where we are today, this being my first official post on the Xuropa Blog.

The gist of my original post was that a revolution does not happen overnight, but often depends on the confluence of critical technologies. As Ron Ploof had pointed out in his post on the Birth of a New Media Revolution, social/new media required easy-to-use publishing tools (e.g. Wordpress), simple syndication/distribution (e.g. RSS), and low-cost bandwidth. Once those were in place, new media hit the tipping point.

My similar prediction concerning an EDA evolution was as follows:

The pieces are coming together for a revolution in EDA. Like most revolutions, it is starting small, hardly noticed by the big guys on the block. In the next 5 years, it will change our industry forever by leveling the playing field, allowing smaller EDA companies to compete with larger ones, giving customers greater flexibility on how and when they access tools and which vendor’s tools they use.

And the three barriers that I predicted would need to come down were:

  1. The high cost of sales, marketing, and support.
  2. Licensing models that lock-in customers.
  3. Lack of comprehensive standards for tool interoperability.

So, where do we stand with respect to these barriers as compared to one year ago?

  1. A funny thing happened since last Labor Day. The world’s economy hit a bit of a bump in the road. Reducing the cost of sales is no longer a good idea or a key initiative … it’s a matter of survival. And “doing more with less” is the mantra. In the world of EDA, we’ve seen companies slash their AE headcount and marketing budgets. New media, which was pretty much just Synopsys last year, is now a part of almost every company’s marketing plan. And Xuropa, in particular, has seen growth from just a stealthy startup to having a key customer in Cadence and dozens of other EDA companies looking to adopt on-line labs as part of the sales process as a way of doing more with less.
  2. Licensing models continue to become more flexible. Synopsys introduced e-licensing, similar to Cadence’s eDACard, enabling peak licenses for periods as short as a week. Cadence launched their Hosted Design Solutions which constitutes a SaaS-like option for small companies looking to turn capital expenditures (CapEx) into Operating Expenses (OpEx). Semiconductor companies want to take advantage of the cost savings and flexibility of cloud computing and are asking their EDA suppliers for flexible licensing models to support them. And the idea of SaaS as a way to differentiate a software offering is being conceived by EDA execs, not just Xuropa and the crazy bloggers. Look for that to happen in the next year.
  3. On the standards front, we’ve had the merger of Accellera and Spirit. We’ve also had a new flow offering from Synopsys, called Lynx, which could form the basis for an industry standard flow. And we’ve had other flow offerings from independent consultants like Steve Golson of Trilobyte with his flowmaker flow. Synopsys’ VMM now run’s on Mentor’s and Cadence’s simulators and vice versa for OVM. There’s talk of burying the UPF/CPF hatchet. And the world is awash in peace, love, and interoperability.

As I said last year, I’m optimistic and I’m also a realist. It will take several years and this revolution is still in its childhood. But there has been a lot of growing up in the last year. I can’t wait to write this post again next year when the revolution will be in its adolescence.

harry

Posted under business, industry

This post was written by harrygries on August 19, 2009

Tags: , ,

Cadence Using Xuropa Labs at DAC

I’m sure there are very few people in EDA marketing departments with the time to read blogs this week, never mind write one - about 96 hours from now the show floor of DAC will open in San Francisco and all their work for the past six months will be front and center.  (No one really thinks about DAC seriously until January is out of the way.)

Anyway, the pressure is off Xuropa.  Being online and available globally 24/7 definitely has its advantages ;)

We won’t have a booth, but we will be there walking the floor talking to vendors and users about what we could do to help their business or experience on the Xuropa Platform.  If you see me walking the floor I’d love to hear your feedback - it’s a great opportunity to meet people that have a passion for EDA, SaaS, and the future of the business of the semiconductor design industry.

If we don’t bump into each other on the exhibition floor (or at the Denali Party), stop by the ChipEstimate booth (#1100) at the below times.  The Cadence Verification IP team will be presenting their products and providing attendees the ability to use the products for themselves on the show floor - using their Xuropa Online Labs.

They haven’t had to worry about installing PC’s and testing them just before the exhibit opens.   Everything is all ready to go and sat up on the Xuropa Platform.  They’ll be using the ChipEstimate PCs with just a browser on the day.  Of course, once the show is over the demos will still be available 24/7 for anyone in the world to continue to go into the Lab and trial Cadence products - a beautiful thing.

So, here are the times.  Susan Peterson will be presenting and I’ll be stood in the wings to answer any questions and get feedback during the demo sessions.

  • Monday at 2pm
  • Tuesday at 2pm
  • Wednesday at 1pm
  • Thursday at 10:30am

Enjoy the show, and hope to see you there!

- James

Posted under Xuropa, business, industry, marketing

This post was written by James Colgan on July 23, 2009

Tags: , , , ,