Your interesting post has generated a lot of excitement within the SW forum.
My take on "cloud computing" is that it might just be our only hope of still using SolidWorks in Heaven. Assuming that you can't take your money to Heaven, you probably can't take your PC either. For my life, I will stick with my own PC, and pray that I can at least take my golf clubs to Heaven. I'm sure that Tiger Woods Golf will be on the "cloud" sometime soon.
So, Gerald, do you want to believe that SolidWorks will cease to exist after death without the cloud???
Oops. This space has been left blank intentionally.
"I pray......." After reading this interesting thread all the way down from where Gerald started it, I regret to say that your post gives me a bad taste. How about editing it in a way that it doesn't hurt other (religious) persons feelings...? I expect many other readers would appreciate that too.
Jan, i second that.
...send a complex analysis/simulation job to a supercomputer if it would return csv results within minutes...
Using a subcontracted supercomputer is an interesting idea. I like the way SolidWorks integrates a variety of tools - sheet metal, surfacing, mechanisms, animation, etc.
If the offloading on a pay-per-gigaflop basis turns out to be convenient, I think your suggestion may get legs.
My dreamy CAD deals with friction and flexibility - springy things that are currently too CPU intensive for the common workstation to handle. I played with a BOXX brand computer at SWW2010. I think it had 4 I7 CPU's - it was a screaming fast system. With that kind of power, the trick is to leverage multi-threading into giving the user a realtime experience.
That multi-threading technology could be leveraged into using a server farm as a peripheral. Again, I'd rather not obsess with the internal code or technology. I'd like you to help me describe and envision an ideal CAD environment.
My motives? Better tools, happier thoughts, people who mutally benefit from their association.
...SolidWorks will cease to exist after death without the cloud...
My prediction is that the trend towards specialization in the CAD industry will continue.
Some 2D applications (e.g. AutoCAD LT) persist because they satisfy the goals of the end users well. Some 3D modeling systems are fantastically easy to use, but ill suited to manufacturing (Google Sketchup comes to mind). There is a market for CAD tools that are tightly integrated with business administration, manufacturing, and marketing to support global enterprises - CATIA, perhaps.
For the indefinite future there will be a market for workstation-centric CAD that supports high security, small scale budgets, as well as the development of manufactured goods with limited support for collaboration. SolidWorks is currently in the mainstream of that market.
I'm worried about a "divide and conquer" strategy. The traditional application of this cliche is to split up the resources of the opponent so you can oppose the fragments from a position of superior strength. I sense that my favorite CAD vendor is dividing its resources in persuit of an ill defined market rather than sticking to its knitting.
I believe that "knitting" should be to focus on reducing the latency between inspiration, vision, design, engineering, and the ultimate fabrication of manufactured goods.
I'm not about to predict that SW will fail without the ability to use a server farm as a peripheral resource. It seems likely to me that, as a targeted product line, it will be segregated and spun off from its parent. Whether that split happens or not, I'm likely to continue as a itty bitty mainstream 3D CAD user.
A couple of articles on the topic are posted here http://www.dezignstuff.com/blog/?p=3103&cpage=1#comment-4247 and http://www.rickyjordan.com/2010/02/solidworks-in-the-cloud.html. I dont think anyone claims SolidWorks should ignore the cloud. The question is how SolidWorks should go about seeing if it has genuine benefit and introducing it in our context.
SolidWorks are already talking up the advantages of using the cloud, at this stage through third parties. Ricky Jordan's webpage lists a number of the supposed advantages. Claims include the cloud will allow the ability to work on various operating systems; enhancing direct editing and breaking down barriers between parts, assemblies and bodies; reduce tedious sketching; reduce overly complex models; improved search features; give life like performance of parts; increase stability; backup data in real time and allow access to Catia data.
Obviously working on the cloud gives the potential for more computing power for some users, greater access via the internet and the ability to have data backed up in real time. How much this will enhance the areas mentioned above remains doubtful. There is a strong implication being made by SolidWorks that these advantages are unique to working on the cloud. But it seems likely that most could be implemented on workstation software. Even at this early stage SolidWorks are adopting a hard sell approach to this development.
That leaves you feeling uneasy for several reasons. It seems these new and enhanced capabilites will only be available to users that migrate to the cloud and not who remain working on workstations. That is a concern. Many of them could be implemented in workstation software and should be where possible. Then you have to ask to what extent is this push being influenced by the revenue the former CEO of SolidWork stands to gain from this development?
I'd prefer to see SolidWorks taking a more reactive approach to this development. Try it out on a small scale to see if the advantages materialize. Wait until the market gains greater maturity and there is competition among a number of suppliers. Companies selling cloud services like CloudSwitch stand to gain a lot from the development. Their obvious interests should be leveraged so they carry as much of the cost of development as possible. You want us to use your services? Fine. Make it work. SolidWorks seems to be leading the charge and riding roughshod over the wishes and concerns of users. Claiming users want cutting edge technology, is an oversimplification. We want technology that delivers real advantages without unwanted side effects. Whether this development offers that remains to be seen.
I wonder about the bigger picture in the computing industry in general. The push for developments in computer hardware and software now often comes from areas outside of engineering that have a greater user base. What direction are these other areas taking? If everyone starts working on the cloud, will the userbase for cpu's be drastically reduced and will this inhibit its future development? If we try to get ahead of what is happening in industry then it will likely come at a greater cost to users and may not be sustainable.
A couple of things as a parting shot. Given the problems that we are having in getting software that works properly, this seems to be an unnecessary and costly distraction. Is CloudSwitch working with SolidWorks on this development and if so are they the only company? Why?
Just a couple of things. Several of the benefits you list under Ricky's article are not the result of being in the cloud, but the result of using SolidWorks 2011. They are two different things. The direct editing stuff doesn't have anything to do with the cloud. The way SW showed it, I don't think that was clear. I believe, although I don't know it for sure, that the functionality of locally installed SW and SW on the cloud will be exactly the same, possibly aside from some add-ins.
Also, I never said that Solidworks is working with CloudSwitch on putting SolidWorks in the cloud. I don't believe they are.
Thanks Matt. I've changed comments to reflect your points.
The claims made on Ricky's webpage are blatantly misleading in that they imply these benefits will result from moving to the cloud. If these developments will be included in software for workstations then that serves to highlight the error.
A bit off topic but could you be a bit more specific on that Boxx machine you've seen? Did it run SW on a 4 I7 CPU system?
The Boxx system is a single socket Core i7 system. It is not a four socket motherboard, just a single socket motherboard with a quad core cpu.
Core i7's are not designed for dual socket or greater motherboards. That is the domain of Intel's Xeon cpu's for this cpu architecture.
...Did it run SW on a 4 I7 CPU system?
My socks were blown off by running PV360 on it. I opened the chop saw tutorial/demo assembly in PV360, assigned a few transparent plastics, set the environment and did a final rendering in "better" mode.
There were at least 8 buckets dancing on the screen during the render. It few by so quickly that I didn't really try to count them. I had the impression that there were 16. Anna knows much more about this stuff than I do - I'm not at all certain what CPU set was on the MB. The sales guy did say it was over glocked or something.
The same task on my M6400 Core Duo / Nvidia 3700 laptop takes on the order of 8 minutes. This demo machine completed in just under 18 seconds. It had an Nvidia video GPU card, but I don't know the specs. other than dual screen in really high resolution. Windows7-64b for sure.
I also launched SW10 SP0 and set up a simple simulation on a cantilevered beam with CosmosExpress. The mesh and presentation of animations was virtually instantaneous; this wasn't a very elaborate test, but it clearly was multi-threading.
Spinning a backhoe with eDrawings was very light weight - absolutely no mouse latency. That's a single threaded process. We had around 9 apps running on the task bar. Win-Tab between them was very quick.
So, the short answer to your question is NO. SolidWorks2010 is not entirely a multithreading application; the modules in SW10 that are so really prove that parallel processing can accelerate some stuff pretty darned good.
Warning: Don't read this as an endorsement or recommendation - I'm not any kind of hardware expert. Do your own test drive.
My first visit to this thread and just finished it top to bottom.
I see a significant hurdle with the "Pay as you Go" paradigm - at least as it applies to my working environment. And that is "getting the money." We currently pay a fixed price once per year for our subscription. As budgets tighten I constantly have to re-justify THAT expenditure. If I had to go back to management monthly...
I can envision the discussion already: Why is this month so high? Did you really need to run that add-on? Well, how much will it be for next month? We don't have any budget left for that. You'll have to model with the basic service only. You'll need to (spend a lot of your valuable engineering time to) put together an accurate projection for service fees. Somebody forgot to close their connection over the weekend. We had to pay for 60 hours of idle connect time. Why do you need to upgrade CAD hardware. I thought we switched to the Cloud so we could use standard stuff! (Because our other engineering software products are NOT cloud-centric).
Could pay as you go save money? I can certainly appreciate the argument that it can. We use SW Premium for access to PDM and Routing. But routing is only used occasionally and COSMOS (included) rarely. So in the end the dollars sent to SW for only what we use would probably be less. But there is the additional cost of my time to manage the usage and campaign for the justifications continually.
I have a project to manage and don't want to add another layer of non-engineering workload.
Another thought regarding the security of files stored online... We often work on contracts involving sensitive government information. The language of the contract and adherence to standard secure protocols would absolutely prohibit the transmission, computation and storage outside of our infrastructure.
Thanks Anna. Just curious: looking in your benchmark excel tabel: is the one in row 1 such a BOXX system?