Let's talk about the Cloud topic that CAD vendors won't discuss

The corporate reaction to CAD on the cloud ranges from "it's stupid" (PTC) to "it's gonna make your desktop useage painful" (DS SolidWorks) to "try it now for free" (Autodesk). But none will discuss price.

I figure CAD vendors haven't talked about cloud pricing, because they are desperately trying to figure out how to make it expensive. They have the rest of 2010 and a few months into 2011 to set their prices. Next year, 2011, they have to deliver -- or else find a different "future" technology with which their marketing departments can distract us.  (Hint: CAD-friendly Android devices.)

Pricing the Cloud 

Let's take a look at what Amazon charges for its cloud service. I'm using them as a measuring stick since (1) they just lowered their prices and (2) at least two CAD-related vendors have indicated that they would either use Amazon or else would have pricing similar to that of Amazon.

Amazon has an array of unit prices that depends on (a) how much storage space you use, (b) the region of the world in which you are located, (c) how much data you transfer, and (d) the number of times you access your data. (See aws.amazon.com/s3.) I'll take 100GB storage in the USA, of which I upload 1GB a day, and I access 5,000 times a day. (There are 220 working days a year, at least in North America!)

Heh: Amazon lets you store as little as a 1 byte!

• Data storage rate is 14 cents per GB/day = $168/year for 100GB
• Data transfer rate is 10 cents per GB = $22/year
• Data access rate is 10 cents per 1,000 = $110/year 

Total cost for a modest CAD design office = $300 a year to store and access 100GB of drawings and other data with 99.999999999% durability and 99.99% availability.

The $300 is the base cost; on top of this, the CAD vendors add their software licensing prices.

The problem with the cloud is that nobody significant is doing it in the CAD world. There is no ComputerVision to under-price. CAD vendors who were used to saying, "Buy us because we cost 20% as much as the incumbent" are now the incumbents who have to set the pricing for others to undermine.

 

Comments

[Chris Wade]

I think the Freemium model is what will emerge here, as it has with most cloud based computing. There will be a basic set of features for free with additional features costing money (probably a few tiers along the way). It is just what makes the most sense.

But they need to do a better job at integrating multiple users on one account.

[Steve Johnson]

Where's that deadline come from? It won't happen. Everybody's in bait mode at the moment, and next year is way too early to do the switch. People aren't locked in tight enough yet. Those vendors who switch while the others are still baiting will lose their customers. Those who keep baiting and waiting will lose their money.

Hmm, maybe this cloud thing isn't quite as attractive to vendors as we've all been assuming. What if they held a cloud war and nobody came?

[Ralf Steck]

Hi Ralph,

you compare apples and pears, as we say in Germany. The Amazon service you mention just provides disk space to save data to (from an application running on your local computer), but the cloud paradigm means that the application runs on the cloud servers and the local machine is just a dumb terminal.

Greetings from Ralf

(I had a rather bumpy flight back from Boston)

[Dave Ault]

There is just no viable solution on cloud for the vast majority of cad users. The PR has been deafening and the proof of concept and definition of liabilities and responsibilities totaly lacking and these add up to bad news for anyone offering or forcing it as a replacement for local installs on ever increasingly more powerfull and cheaper workstations.

The latest parasolid kernal allows for multithreading according to Siemens and thus dies the only possible advantage accruing to the cloud in the form of faster processing. More dot com type hype looking for a problem to solve that we "just can't live without" fails once again to deliver a real useable worthwhile product.

[Aakelley]

I still contend that the missing ingredient for CAD in the cloud to get off the ground is the user business model (distinct from the vendor business model - there are plenty of those, as you point out in your post all depending on some underlying realities of cloud economics). The user business model has to emerge where CAD in the cloud just makes more sense that an installed application. It could be that users need to access the application from many different machines. It could be that they need to connect with users that don't have normal access to CAD (of course ease of use becomes a critical issue here). It could be that the social production model takes hold and CAD users come together to complete a specific project then move on to form a new group to get something else done. Whatever it is, there has to be a fundamental change in users' needs or behavior before CAD gets serious in the cloud.

[Sid]

(Ralf Steck has also mentioned this above in the comments)

This is just the price of the platform, not for the whole CAD. Add the cost of access to Internet as well (probably substantial, as you would need a connection that allows you to work on 2-3 GB of data per day)

In addition to this, you will need to add the cost of IP, ie the cost incurred in providing the functionality to the end user.

Things like viewing can be free, core design, interoperability & other apps. can be charged on a tiered basis.

But the whole industry is reluctant to move in this direction simply because it will shift the focus from the CAD vendors to the users, and a per use model will expose a lot of things. The foundations of this industry will be called into questioning (things like price vs. benefit, channel cost, support cost and the latest "future")

just my 2 cents...

[Bill Fane]

I agree with "Aakelley"; the cloud will only work when/if it is to the advantage of the user. I can see a mixed model; a company would buy a number of stand-alone (or network-licensed) copies for their core work, and then buy additional metered cloud time to handle temporary peaks.

[Ken Elliott]

The cell phone system is pretty close to what we can expect from a cloud app. We pay $1200 a year, and get performance much worse than our land lines. We sacrifice a lot for portability - the killer feature.

I think most CAD vendors are playing a wait-and-see game. Let other's pioneer, then copy whoever has success.

The first good cloud app will be metered license servers. Everyone already has network license servers, so all you need is the ability to have it "borrow" a license for peak usage and pay accordingly.

I also think specialties like mold flow analysis, FEA, CFD and such will appear as cloud apps. The cost of such software is expensive, and so is the person running it. The ability to push the model to the cloud, run in on a powerful machine, and get results fast make this attractive. Naturally, vendors will realize they can raise the price of software to drive more cloud business. And the IT department would rather not spend their budget on high performance workstations anyway. I'm sure you see human-operated offerings based in India and China first.

There's also the "lego" CAD model. That's where you use a web-app to simply connect a bunch of components together to create your device. In the cloud, the vendor already has models of all those components, and you can pretty easily 'snap' them together. Simulation, CFD, and such might work pretty good, since the cloud vendor has stored pre-calculated data on all these parts, they get paid by many for doing the work once. Plus, they become the gateway to the store.

But I doubt you see any CAD vendor you know of doing this. A startup will drive this, and get bought out.