One concern with any SaaS (Software as a Service) product is the potential for downtime. Is this really an issue? After all, big Cloud vendors have multiple server farms as part of their huge infrastructure investment. This provides redundancy to keep things going even in the event of a major local disaster or two. Cloud vendors have a lot of experience handling things such as power outages, hackers, denial-of-service attacks and the like. Amazon, the vendor currently used by Autodesk, promises an annual uptime of 99.95%. That’s got to be good enough, surely?
Maybe not. The Amazon cloud service has had some noticeable failures, in some cases affecting customers for several days. Amazon may promise a certain average uptime figure, but it provides only credits if it fails to meet its targets. Amazon has been known to be slippery about using fine print to avoid paying those credits, which in any case would go to Autodesk. Joe Drafter, who relies on a Cloud application to do his work and who suffers a significant loss of income and business reputation from a 4-day outage, probably shouldn’t hold his breath while waiting for a big fat compensation check to turn up.
But is a Cloud solution really going to be less reliable than what you have now? Nothing’s 100% reliable, including a standalone PC, so what’s the problem? The problem is that with the Cloud, the potential for downtime is in addition to that you currently experience. Here’s a non-exhaustive list of the sort of things that could stop you producing a design using traditional software:
- Power failure at your office
- Your hardware fails
- Your operating system fails
- Your CAD software has problems bad enough to prevent you working
Here’s an equivalent similarly non-exhaustive list for a SaaS CAD application:
- Power failure at your office
- Your hardware fails
- Your operating system fails
- Your browser or thin client software fails
- Your modem fails
- Your Internet service provider has an outage
- Internet connectivity infrastructure failure
- Cloud vendor infrastructure disaster
- Cloud-based CAD software is down for maintenance
- Cloud-based CAD software has problems bad enough to prevent you working
Each of these items may represent a relatively small risk, but the additional potential for disaster adds up and is real.
There’s another aspect to this issue that makes it significant, and that’s the psychological one. People hate feeling powerless when faced with a problem. If your hard drive crashes, even if you don’t have IT people to look after it, you can hop in your car, buy another drive and start working towards getting your problem fixed. If Amazon has a Cloud outage, there’s absolutely nothing you can do about it but wait for an unknown amount of time. Even if you were Amazon’s direct customer and not a sub-customer through Autodesk, you could expect to have a very frustrating time even trying to find out what’s going on. I’ve been in that situation when my old web hosting company went through a massive and protracted meltdown, and it’s horrible.
What do you think? If everything else about the Cloud was great, would worries about downtime prevent you from considering a SaaS-only solution? Is it non-negotiable for you to be able to keep working even when “the Internet is broken”?
They say that Data doesn’t exist unless you have it saved in at least three places. But if you have your stuff backed up all over – how do you track which version is current?
I don’t see Cloud storage as a replacement for local storage and backup. However, using the cloud to sync documents could be a really convenient way of providing access to the right version of the right document as you need it.
Paul, thanks for your comment. I intend to cover data storage and security in a future post.
i’ve had to explain cloud computing to a lot of people this year, and i’ve found that a great analogy (when trying to explain the risks of cloud computing to the non-technical) is the difference between owning and driving your own car to get to work every day versus relying on a taxi service. There are pros, cons and risks inherent in both solutions. The important thing (IMHO) is making sure the customer understands what those risks are so that there are no surprises and several backup plans in place if the chosen solution has a failure. For example, if the customer chooses to go with a taxi service instead of owning and driving a car: the taxi may cause delays for pick ups and drop offs, the driver may be a dangerous driver, the taxi may be dirty, unreliable and poorly maintained, the driver may not known how to navigate well or take extra long routes to get to places to increase the fares, the taxi may not have enough room for the passenger’s baggage, the taxi service might not be available 24/7, etc etc etc. it pays to walk down every possible scenario that could cause failure, delays and ultimately a loss of profit.
Steve,
I think you should be commended for taking the discussion about the cloud in this direction. Cloud vendors are reaping a huge amount of free (through the cloud) business consultation via. you and others.
Reality is, the advantage and disadvantages, of the cloud, for a users, are as unique as each of their businesses. Dave ea is correct about people knowing the possible problems but defining what they are is not as straight forward as it would appear. I don’t think using analogies (comparing the cloud) is necessarily the correct way to explain, as it raises other complications.
The problem with promises about down time/losses caused by vendors is, it would appear, are going to be (are currently) covered by vendors “contractual” terms and, software vendors are un-regulated businesses. Issues of variation/late delivery/losses etc. are handled – contractually – on a daily basis in businesses projects/employment and that means cloud vendors need to be prepared to negotiate their T&Cs. It’s not rocket science: cloud vendors just have to accept they are a small part of a wider business world and understand the need to play by sensible rules and accept their responsibilities.
At the end of the day, if only considering down time, in this post, I would have no great concerns about cloud use if losses, no matter the size, can be contractually and reliably recouped.