Cloud computing (and my company was early in this market) is not software as a service, which could be done on a web site, but computing as a service. The idea, for big companies, is that you don’t have to buy servers for times of maximum use, but can go into the cloud when you need more compute power than you have, and for anything in fact. It’s also good for disaster recovery - if your computer center is wiped out, everything is on the cloud.
There are just as good ways to protect against single processors going down - most high end servers have all sorts of failover capability and redundancy. Nothing new in this. Ditto for storage. Doesn’t help much if your computer center floods, and you are not big enough to be able to afford off-site backup of computing capability.
You can put your applications on the cloud or use software as a service, but that is not fundamental to the concept.
It is also not the same as mainframe computing, except in the sense you don’t do it locally. I used mainframes for many years, and your data and computing is in a very specific location. On the cloud you don’t know or care where your jobs are being run, or the exact physical disk where your data resides.
BINGO. I’ve been in the industry since 1983 and in my view this sums it up perfectly. In the modern age, marketers and management love to toss around buzzwords and cliches to impress. Don’t buy the hype - nothing’s really new.
“It just works” is one of the major arguments against cloud computing. Of all the electronic devices I have, all of them are more reliable than my network connection. It’s bad enough (though unavoidable) that my web browser stops working when the network is down, but if my word processor and spreadsheet and C compiler and all the rest also stopped working, that would just be unacceptable.
Now, the cloud can make sense for some purposes, such as data storage (especially if a large number of people all want access to the same data). But computing in the cloud only makes sense if processor power is expensive, which it’s not.
When I worked for a PC-centric company, we had a lot of shared drives, and I was as nearly dead in the water when the network went down as I am now with a thin client. It is different if you have minimal interaction with others.
And how secure is your data, if you don’t really know where in the world it is? What if whoever’s got it pulls the plug, or turns round and says you can’t have access to it unless you pay us off?
Nothing about cloud computing changes this; the same questions have existed for decades. When you outsource to a large data center, you’ve got to have a certain amount of trust/verification, whether it’s one mainframe or a virtualized server farm.