The Cost of CAD on the Cloud

To say that CAD vendors are far more excited about CAD on the cloud than their users would be an understatement. In fact, the more excitement the vendors show, the more skeptical their users get. Frankly, I am not surprised. The image of large corporations finding new ways to squeeze more from customers is quite easy to imagine in the capitalist world that we live in.

The many technical and logistical challenges to putting CAD on the cloud are known far too well. But probably the main challenge for CAD vendors would be to get users to adopt it. And for that the CAD vendors would need to give their users a very good reason to move to the cloud. I believe the most critical reason is cost.

During my interview with SolidWorks CEO Jeff Ray at SolidWorks world 2010, he said, “Why should you need to buy a power plant to power your house when you can simply plug your appliances into a socket which is connected to a power plant located elsewhere. You should only pay for what you use.” I found this statement quite interesting because Jeff is implying that his customers are currently paying for a whole lot of stuff that they do not use or need. This is precisely what a lot of his customers have been complaining about for a long time now.

In fact, I totally agree with his Jeff’s line of thinking. That’s why at SYCODE we have split our products down to a very low level. We could have easily had a single SolidWorks Data Exchange add-in product that imports and exports all the file formats that we support. Instead we have split our data exchange technolgies into 28 SolidWorks add-ins, each one of which imports or exports a single file format. We believe that a customer should pay for and get only that part of a technology that he needs. Of course, if a customer wants more than one technology, then we offer him a package where he ends up paying way lesser than the cost of the individual products put together. So if CAD on the cloud is actually going to achieve this, it will indeed be a very good thing.

Another important factor is the cost of hardware. And this is really a matter of implementation. At SolidWorks World 2010 we were shown SolidWorks V6 running on a netbook. In my discussion with Autodesk CEO Carl Bass he suggested making use of the computational power of the device to offer a better user experience or even to make CAD on the cloud feasible. I find this interesting as well because this means that a part of the application will be installed locally and the other part will be in the server. So I imagine for a better user experience, the user would need to use a more powerful device. The same thing goes for increased usability. For example you will obviously not be able to do the same things running Inventor on an iPhone as opposed to running it on a workstation notebook. So if CAD on the cloud is implemented in such a way that user still finds the need of spending a great deal of money on more powerful hardware, then the benefit of cloud computing as far as hardware cost is concerned will go down the drain.

Right now the information being put out by CAD vendors getting all excited about the cloud is extremely sketchy. Nobody is giving any specific information about the real cost savings for the user. I think this is because they have not yet reached a stage where they can figure it out themselves. Of course, the CAD vendors may have “guestimated” a few numbers before embarking on something as large as cloud computing. But the fact that they are not going public with their numbers goes to show that they are not confident enough on their accuracy. Frankly, I am not surprised because everything seems to be up in the clouds right now (pun intended).

This is just the beginning. The technical and logistical problems will be sorted out over time. No doubt about that. But I believe that the fundamental reason why users will or will not adopt CAD on the cloud will continue to be the same – money. The direct and indirect cost of CAD on the cloud is going to decide whether this concept works or fails. Or rather, whether it is adopted or not. All other factors like ease of use, ease of data sharing, increased storage, greater automation, better mobility and all the other advantages that we keep hearing about cloud computing will amount to nothing if CAD vendors are not able to deliver significant cost savings to their users.

Let me emphasize this. The cost savings has to be significant, not marginal. Users do not even use the latest version of a product until a few service packs have been released. That is how skeptical or resistant to change they are. Most of my customers are using a version of their CAD system two or three years old. As the wise man says, “If it ain’t broke, don’t fix it“.

  • Cloud CAE is going to be awesome. I can't wait. But, I think I have to (until broadband speeds in the US get fast enough to handle it). There are a lot of costs besides just software cost in total cost of ownership. As you mentioned, hardware is one of them. Also, the extra IT personal who are there just to make sure local installs go smoothly and fix things when they are broken.

    As a longtime cloud CRM user, I can say that there are no serious worries about security… surely customer data for many businesses is just as critical as IP. (Maybe more?)

    Named users seems like a great model to me. Totally scalable. Everyone is always on the same version, avoiding headaches on that end.

    Up to date, more robust technical support since organizations only need to focus on the current “release”.

    One of the bigger keys to this will be “who owns my data?” If you are using Salesforce.com and want to switch to ZohoCRM or Netsuite, you can do it. It's a bit of a pain in the butt, but the data isn't locked away in proprietary formats like most CAD data is today.

  • jgerth

    Who owns my data is definitely one of the bigger questions. The other one un-addressed to date by any vendor is 'who owns the liability' for uptime, security, confidentiality. None of the vendors has any track record to date is providing a reasonable guarantee of confidentiality nor do they have a substantive track record in security. If Freewheel or Dragonfly are examples of what the software vendors are thinking security-wise, the idea of putting any company data on the cloud in verboten,

  • I think of the combination of locally installed software and cloud computing the other way. Having lived through the advent of the move from IBM mainframes to Apollo workstations circa 1989), we workstation users would cringe at applications that were not ported to the Apollo – instead running in a VT100 terminal emulation window on the workstation. So when this “software as a service” paradigm surfaced a few years back, I blogged that this approach was turning my “Pentium CPU” (that dates the post) into a VT100 terminal.

    So we are very conscious of not wasting people's hardware investment and having it go down the drain. On the other hand if my shiny new 2010 PC could make a call to the cloud and have a bank of servers attack my problem to provide me with an answer in a matter of seconds instead of hours, then maybe software plus services is a WIN-WIN. Think of photorealistic rendering (like Project Showroom) or sustainability analysis (like Green Building Studio) as examples that would outstrip even my most well-equipped desktop PC. There is room in the computing world for both.

  • jgerth, never say never but I would say use of the cloud for our type of data needs to be considered extremely carefully. You raise the issues of vendors ‘track record’ and this is something they are all going to have to address.

    I have nothing against the various ways the ‘cloud’ can and is used but I do believe vendors must step up to the plate and demonstrate they can be honest, open and transparent about their operations, ownership, fault and loss acceptance. If vendors continue along the same path of ‘no responsibility’ as seen in their current EULA and T&C’s then using the cloud, for many, is a no brainer, its should not be used, as you say verboten!

    My experience with Autodesk, their EULA and their appalling response is a testament to why ALL users should treat any comments about the benefits, or Autodesk Labs Scott so called “WIN-WIN”, with a large dose of skepticism.

    As I said in Matt Lombard’s blog, in a discussion relating to Solidworks in the ‘cloud’; if vendors want users to consider or move in this direction it is time for them to publish ‘papers’ with details of how they would like to proceed; addressing ALL the issues including details of just what their responsibilities and obligations are to be. A good job for Autodesk’s Legal Lab 😉

    This would allow users to discuss the real issues and understand, more fully, the problems of cost, data ownership, access and possession and just how honest and responsible vendors may be, in general operation and, in times of crisis.

  • Kevin Quigley

    I'm waiting to see how all this pans out with SolidWorks – still waiting for that call to try it out 🙂

    But Paul's last point is, for me, the most crucial – times of crisis. If your data is all held on a remote server run by your CAD vendor it is not ownership that is the concern – it is access. What happens if your CAD vendor comes bust? How can you access the data then? Even if you could access it (from say local backups – if this will be possible), can you then use it, if the application is all cloud based?

    For this reason alone I think there needs to be some ability to run this system locally. But then this opens another can of worms. Is the cloud app running on a platform we have access to? Is it actually RUNNING on Windows or Linux or something else?

    There is talk of large corporations running their own internal clouds so does that mean I can run my own cloud as well?

    Too many questions….not enough answers.

  • jgerth

    all due respect Scott, but comparing a main-frame to workstation transition to a PC to cloud is not even apples to oranges — it's more of a pomegranate to pumice comparison.

    Since the same organization owned the mainframes and workstations, the data and quality of service was under the control of that organization. The proposed cloud structure requires incredible amounts of faith on customer's part, coupled with (to date) a total disclaimer of liability and responsibility on the vendor's part. Any CIO who entered into those terms should be terminated instantly, and the contract repudiated, since said CIO was obviously not sane.

    That aside, I do participate in 'cloud-like' activities daily, donating cpu cycles via BOINC to various philanthropic activities. The same benefit to CAD computation you describe could be accomplished cloudlessly with parallelizatin for code, and distributed computing within the confines of a single organization. While the CIO is playing solitaire, his quad core processor could readily be used via an internal BOINC mechanism for stormwater modeling or CFD. Heck, renderfarms have been around for at least one decade, and no cloud needed.

  • Scott, I can understand Autodesk's position in wanting to keep some installed app on top of the client end of the SaaS terminal but that is a major backwards step regardless of the PC's power – that in itself is the point.

    The promise that SaaS brings is “access and collaboration anywhere, anytime” and frees users from the horrific shackles of per machine installs and licensing. The moment you bring a geofixed dependency into the equation, don't call it SaaS because its not – call it buzzsaw and be fair.

    I'll say this for the record, at Aftercad we instituted a policy of using Google Docs instead of MS Office so that we could truly get the hang of collaboration and be done with “email Armageddon” – trying to figure out with Doc or spreadsheet was the most recent. To a person, none of us will ever go back to those neanderthal days of using Office and I gotta tell everyone here – that one time you need to access something at the last minute from home, the office, on holiday to save/close/rescue a deal and there is it exactly where you left it in the browser – its that experience that will make you realize SaaS is soooooooo much more civilized than the old way.

  • facebook-576497142 [inmate of…..? 😉 ]

    I understand and would agree with you on principal and, this obviously works for you and yours.

    Could I ask tho' from a business perpsective what checks and validation procedures does your company have in place and use to ensure the 'service' you expect, and beleive you have, is what is availble at all times?

  • Paul,

    we ran the service as Aftercad Online during the beta phase and Renderjam.com for the release. We managed to keep it up well within a 99% SLA. Never lost anything either which puts us ahead of Microsoft at least. It was a 24×7 commitment which we took seriously but for the most part, with modern equipment and data centers it was entirely manageable with the appropriate planning.

    Having said all that, we just handed it all off to the Open Design Alliance and each member will have to run things in their own way but we'll be there to guide them through the experience as required.

  • Kevin,

    In the Renderjam.com (now ODA property) service, the user had the ability to access a full web based file system where they could upload and download files as they pleased. There is even a function to zip entire folders of content and download it it right to the desktop. In this way people where able to backup content as they saw fit with out any interference from us.

    As far as major vendors going bust and taking your CAD data to the digital grave with them, that's unlikely to happen with Autodesk or solidworks and you are probably not going to use a fly-by-night service to host your IP sensitive CAD files on.

    Really this comes down to a gut thing – running the numbers, is the cloud app provider more likely to lose my data or is my desktop/laptop more likely to crash/get stolen/die? This is of course why everything is going cloud and SaaS.

  • facebook-576497142

    Thanks for the response. Whilst access is of importance; for me it is a requirement to be able to demonstrate security. There is much said about this but very little substance and your comment to Kevin Quigley, “Really this comes down to a gut thing – running the numbers, is the cloud app provider more likely to lose my data or is my desktop/laptop more likely to crash/get stolen/die?”, raises the eyebrows a little.

    ‘Gut feelings’ are not a good way to test security; proof of process and validation of what is claimed is necessary, as a starting point. What did you do, or have, in place for this type of scrutiny to be done by the customer, using the system, and or independently?

  • Kevin Quigley

    Whilst it is unlikely that the main companies will fold, it can happen and it does happen. Probably more likely is that the service provider of the infrastructure required will fold – it has already happened to me. My web host went bust, taking the website and all content, domain and email down for 3 weeks and a lot of trouble getting it all back again.

    You also seem to imply that we will choose who to host our files with? That's not the impression I'm left with after seeing and reading all the SolidWorks SAAS stuff. The impression I get is that everything will be done by SolidWorks (which to be frank, I would be happier with).

    If we are expected to arrange cloud hosting as well this would worry me.

    There is a certain arrogance in the IT industry that assumes that vendors are impervious to failures – none more so than the hosting business in my experience. Regardless of the “stats” the reality is – and it IS reality – systems fail, people make mistakes.

    I use a lot of online services for business admin and email. They do work well, but they also do go down occasionally as well. The issue then is access to the data rather than ownership. Access is the critical issue. Until we see tangible evidence from the vendors about how and when we can access our files, what procedures are in place for times when servers or infrastructure fail, and how we can work anytime, anyplace – as we can now without the cloud – many will remain sceptical.

  • Jonathan_Scott

    Deelip, I agree with you that the economics of “CAD-on-the-cloud” will dictate its adoption (or lack thereof). I think the major PLM vendors will figure this part out. However, I think FUD (Fear, Uncertainty, and Doubt) about intellectual property security on the cloud will trump even a great financial proposition for the cloud if we're not careful (as is evidenced by a number of the comments to your blog post). I hear a lot of people in the CAD community saying things like, “no way would I put my sensitive design data on a cloud.” But I think there are already answers out there to the security question, like private clouds and community clouds. I believe it was Jon Hirschtick who pointed out that one of the biggest threats to IP security today is people walking out the door with design content, and I agree with him. My more complete thoughts on this subject, the basics of private and community clouds, and how PLM will benefit from cloud computing can be found at http://www.razorleaf.com/2010/04/cloud-y-with-a

  • Patrick

    Perhaps we keep data local, and use the cloud as a tool, this sorts most data ownership issues as well as transferring the burden of install, configure and maintain to the cloud vendor, every one accesses world class, and we all get to sell our deliverable’s.

  • Fjaviemx

    I think there’s no question about it, if it’s convenient for software vendors that’ll be it. It is not coincidence that some CAD software deliverately prevent’s its users from saving/exchangin files in a determined format just for the sake of “bussiness srategy”. And I also support that could be effective to tackle software piracy. My only concern is bandwidth… that means the current “cumunications gap” between countries with low bandwith and a higher one, will indeed turn into a Technological gap..that would be unaceptable and vendors would strugle to sell their acces in places with no appropiate means.