Cloud and Subscription
So there are some advantages to some ways of looking at the cloud. But there are also an entirely new set of issues that it raises, and none more interesting than subscription.
On the good side, it means you never have to apply another service pack. That will become an automated function on the server farm. But, will the whole system be available during the update? Do we have to talk about multiple server farms so that the cloud itself is redundant, and one can be updated while another continues to run? My guess here is that this is an area ripe for dissatisfaction. Whenever you update something, you’re going to make someone unhappy.
Also good is that every license would be a floating license. Well, that’s good until we find out that cloud-capable licenses are not interchangeable with serial numbered installed software. There may be some sort of a price differential, probably finely tuned to make it marginally worth while to use the advantages of a cloud-based license.
Licensing terms? Do you have to pay extra for the “distributed” computing aspect of the cloud? Are there levels to how many CPUs you can occupy at any time? an average value? a total time limit?
There has been an assumption that tech support disappears on the cloud, but it doesn’t. Bugs still exist. You still have to report them. But aside from training, what is the VAR there for?
What happens to your access to the software, and indeed your access to your own data when you drop subscription? Other pay-as-you-go software models have not been successful, but it doesn’t make sense for SW to allow you access to the software if you have not paid, because they now have to pay for the server farm.
And then you need to use old version for some reason. Is SW going to serve up multiple redundant clouds for each of the last 4 versions? No. This is another source of lack of control, and a source of potential frustration with this scheme. I don’t believe that this is a system that will find universal acceptance or something that solves all problems. It does look like a solution to a few problems, while it causes others, so you just have to make the call if swapping the old problems for new problems is going to be worth the additional cost.
Additional cost??? You betcha. They can’t set up (or rent) all of those “thousands of servers” for nothing. If you aren’t required to buy your own workstation, you’re still gonna pay for that, because no one is going to give it away. Cloud licenses are going to cost more per year than the licenses we have now. Driving that recurring income is what this is all about.
Up front costs could be handled in different ways. The $3995 purchase price doesn’t seem to have a place in the cloud, but I guarantee you they will take advantage of the fact that you expect to pay something for the software other than maintenance. In the same way that a car lease still expects you to make a big down payment, I think this will work that way. Leasing as a finance option is a great option to avoid if you can, but I think we can expect this excuse to rewrite the sales and subscription/support rules to be used as an opportunity to make more rather than less money. There is clearly the option to offer customers stuff that they find more cost effective, but I’ll be willing to bet that it will be priced so that in the end you wind up paying more rather than less.
Then what happens when they change vendors for the cloud service? Do you ever get to keep your real data?
So we may dream of ultimate flexibility and control, but I doubt we will see that dream satisfied. Cloud is going to solve some problems, but it will also create some problems, and the best you can do is to make the call between which gives you the best net functionality for the best price. It will be interesting to see how this winds up playing out over the course of the next few years.
I had a friend who used to say “the water is always cold”. By that I think she meant that starry-eyed dreaming about what could be will always be interrupted by reality. It seems easy for some people to get caught up in the “wow, this is gonna be great!” enthusiasm, but any deal that is that good for you would probably be not so good for SolidWorks. I guarantee they will look out for their best interests first, and while you may get some stuff you want, there will be a “cold water” trade off, and you will have to pay more than you do now to get it.
Matt,
I have been reading all the postings about SW in the cloud stuff with a lot of interest. I have not see any of what was shown at SWW and hence I am in a way commenting blind here plus I have been away from SW for more than 2.5 years. But I still read about the product and company that I spend 11 years in. Since leaving SW, I have been working for companies making SaaS-based products and hence I can reflect on my experience to comment on how things work in the cloud and what issues SW will have to iron out.
I believe the whole challenge with SW in the cloud will do with performance and nothing else. Everything else such as upgrading servers, load balancing, multi-tenancy, how to charge for subscription etc. should all be easy to figure out given how many companies (large and small) make their living out of this. Take the examples of Salesforce, Freshbooks, Mozy and a zillion others out there.
I used to think about this particular way of computing when Alibre first came out. I was not convinced for one simple reason – customers (including you) have never been satisfied with SolidWorks performance (open, save, regeneration) when everything was running on your own machine. I read that Jeff Ray himself rated SolidWorks performance as C+ during this SWW. Opening assemblies over WAN or your own LAN was a big no-no because of influence of network traffic on open times. So the main hurdle could be the responsiveness of opening large files over the Internet. Downloading even a 10 MB image file over the web even over a fast connection is still slow.
I am not saying any of this is not doable, but will require significant advances and for all you know, SW may have figured out some of this (again, I have not seen any of what was shown other than the images I have seen in the various blogs).
Here are the significant advances that will need to be made
1) Reduced file sizes
2) Open/save times
3) Regeneration – how much of this can be done on the client vs. the server
What is not clear to me is how much is done on the client and how much on the server? I would think you would need some client that can take advantage of the graphics processor of your machine, RAM etc.
I don’t personally think versioning will be an issue because in a true SaaS application, there is only one version – the version that is on the server. So if they upgrade, everyone upgrades.
Having said this, I don’t think SolidWorks has a choice but to try to make advances here. It is only a few years back where there were enough naysayers who said SaaS will not work, but a large portion of the software industry has adopted it and successful with the model. SW has already said that both desktop and Saas versions will co-exist – it has to because SaaS version still has a long way to go.
If SW can pull this off, then it will be a disruptive innovation and given where SW is in its life cycle, it may be time for one because the desktop version is probably getting bloated with every release. But I think we have a few years to go before it truly becomes a reality, I hope I am wrong.
Gopal
This is in my honest opinion one reason other software companies are formed. If you don’t think you’re getting what you want out of existing 3D packages then design your own. Handpick some of your smartest friends, code writers or whatever you need and get busy. But if you can’t then start shopping for another package.
When a 3D software provider start losing market shares then maybe they’ll listen and ask why? Are we providing a product that customers want? If not, why not? Have we become so smart that we will tell the customer what they need and want? A big piece of humble pie is good for the big players in the marketplace! Look at the big three automakers. You build crap no one buys.
Bottom line is I hope SWX is not getting too high and mighty and not listening to what the customer needs.
Matt, the speculation and questions you raise here are legitimate areas of consideration, all who go ‘wow’ about the ‘cloud’ need to consider. At the end of the day decisions about control and access to our data must be made based on ‘reality’.
Many of the concerns I have raised to do with Subscription T&Cs and EULA etc. – even tho’ I started this 5 year ago -, and the appalling way that has been handled by vendors, have relevance to this discussion and, MUST be addressed properly before using services in the ‘cloud’.
I have absolutely no objection to the principals and various ways ‘clouds’ could be utilized, just as I have NO objection to Autodesk’s audit clause. What I object to is the fact, I as a user, are being expected to give a vendor access to my systems without the ability to oversee or validate what the vendor is doing or has done! It is ridiculous that one company should expect that level of access and at the same time not expect to be checked?
The relevance caution and or the outright rejection of ‘cloud’ computing is obvious if looked at from my perspective. If users are not allowed, or given the opportunity, to validate vendors actions when – the vendor is – accessing the users own system it is fair to ‘assume’ the users will have an equivalent lack of control, in the ‘cloud’ and therefore no sensible reason to use it.
For a user, or me, to even consider using the ‘cloud’ I would want answers to, not only, the questions I have already asked but others as well and an ability to ensure ‘total and absolute’ control over my data.
We have witnessed a vendors response to my requests at a much more benign level and they have not been able to address, indeed they will not even discuss them, so not only are they going to need to ‘educate’ users, about the ‘cloud’, as you put it, they are going to need to become much, much more transparent in answering and addressing users concerns in the areas of EULA, T&C’s, access, and accountability. Maybe then, the ‘cloud’, might loose its ‘WOW’ and be seen and used appropriately.
I for one would jump on the cloud; I currently use online backup so I am not paranoid of data loss. I find a bigger hit to productivity is the endless SW crashes on my system; if the cloud truly made these go away or diminish in number seriously, that would be enough to satisfy me.
Productivity is what I want, and the biggest issue I have is SW performance/crash issues on both my work and home computer. If the cloud addresses these issues, then it is a move in the right direction in my book.
Of course, choice should always be maintained (cloud or local version).
Matt,
This cloud deal is nothing but increased cash flow for companies. And I believe that SWX will “own” the data and hold users hostage to recurring costs. There will be no recourse for users of the cloud to be made whole by SWX failures such as data loss and corporate espionage and in the fine print SWX will exonerate themselves from any of these potential, no wait that should be inevitable failures on their end. NSA and the Military state without reservation that the only network that can be made secure is the one you have in house that NEVER goes online. It boggles the mind that any cad company could think that the internet is reliably usefull for any serious cad work.
Numerous reasons for the fact that the cloud will be still born have been in many posts today. Productivity is what I want to and I know that when I hand 90% of my tools and 100% of my proprietary information over to someone else to manage I am doomed. I can imagine when and not if being saddled with a bad release that shuts aspects of my business down for perhaps a month or longer while my reliable “cloud gurus” fix the problem and I have to wait. I wonder if the cloud these execs are refering to somtimes is the cloud of herbal smoke in the boardroom that considers this stuff.
I really get hopping mad when I see cad outfits create window dressing with my yearly fees and then leave serious unresolved problems affecting user productivity day in and out.
Cloud Cloud go away come again never.
Far to many pitfalls to bother with any kind of consideration let a lone any REAL consideration.
This cloud crap is another area in which software developers waste time (and our money). Instead of giving us the things we ask for to get our jobs done how we wish them to be done (like.. with less time spent on crashing and work arounds or hell… new features that are useful).
All this development time to try and big note themselves on news websites with whatever buzz word crap they can think of next instead of focusing on REAL WORLD feature-sets (and there depth) and stability/speed/UI.
If they put as much effort into overhauling the parts of SW that desperately need updating to the 21st century… imagine how excellent Toolbox would be. Or a modern globally available Equations/Parameters interface or equally Custom/System/File Property integration. *sigh*
I’ll take sunny day’s thanks!
PS. There is only one thing in all this Cloud talk that is of any earthly use and that was probably something not even related to it anyway (and SHOULD be a focus for implementation outside of anything cloud). And that’s something i read about better/more realistic flexible/soft component support. e.g. springs, o-rings, rubber, gaskets etc. A professional implementation of such a feature alone would give me so much more then Monkey Magic’s pet!
I’ll wait and see what announcements are made today regarding this. All I’ll say is that if the technology is so advanced now why don’t SolidWorks prove it. Send me the link and I’ll log in and run it on some of my files.
The issue that worries me is not so much the mechanics but the cost. You see I used to use Think3 software that you rented year by user. A good idea in theory. Problem was that Think3 did not develop it fast enough to meet my needs so I stopped using it after 3 years. All those files I had I cannot open because I stopped paying.
The other thing is what about disaster management? If I am using software that is hosted by another entity, what happens if that entity goes out of business – that can’t happen? Recent events prove otherwise and I wouldn’t count on a Government to bail out a CAD firm.
What happens if the server company goes under? Ever had that happen to you when a toolmaker goes pop and you have tools on their premises?
There needs to be a backup option. So what is it?
BTW all the cloud stuff is on YouTube:
http://www.youtube.com/watch?v=jN01m0w4qYA&feature=related
Thanks for the link Kevin. I wanted to see what was presented before I commented on this.
I see two potential issues for users in the cloud model. Bandwidth, and data possession.
Bandwidth is obvious, but may not be as bad as I think if we don’t have to wait for entire models to download. I’m eager to see how this works outside of tech demos.
Data possession is different from ownership. A lot of people are going to need permanent access, and quite a few others can’t let their data off-site. If SW is aware of this (how could they not be?) Then they will have to provide cloud server installations at the customer’s site (maybe with cloud backup). If so, then the whole thing could still be installed on one workstation. (With cloud access to shared models like hardware?)
The cost of using the cloud I expect to be much less than owning workstations. I rarely get to use more than about $250 worth of cores of my i7 965. The whole workstation cost over $2000. (I’m talking original cost, of course.) Cores in an 8-socket server somewhere are a whole lot cheaper, and I won’t have to pay for them when I’m not using them. And those cores might be more powerful than Core i7 can deliver. Maybe the servers will run on IBM Power servers, or clusters of PS3 Cell processors. I dunno. Maybe rendering and FEA will be computed on farms of GPUs. A user who can afford one workstation can’t afford to own that kind of hardware. Several thousand could.
And there’s the possibility that SW or Dassault may not own them either. You can lease space on server farms that are already established.
If SW can address bandwidth and data possession, I think they might have a winner. The demo showed several worthwhile things you just can’t do on a desktop workstation within the next say, decade.