It's pretty wild that this stuff happens. Similar to today's nasty outage, Google has had some massive problems with its app engine datastore...
I'm curious if anyone has any predictions about what the landscape will be like in a few years? Will these be solved problems? Will cloud services lose favor? Will everything just be designed more conservatively? Will engineers finally learn to read the RTFSLA?
The benefits of the cloud is just too great, we won't go back. Except in a few years, when something goes down, instead of being some random site who's down, it's going to be the 20,000 sites that are hosted on that hardware.
Now, different clod providers "speak" different languages now. But I can see in 5 or so years that the cloud will speak a similar set of languages. One could use a storage cloud from this provider and a CPU cloud from another provider.
I could eventually see, with help from functional languages like Lisp or Erlang, a intra-company cloud running on and between networks. CPU could be bought from 3 providers, and storage could be bought from 4 providers, with GPU acceleration clusters when big data needs crunched quickly.
Or right now, companies can make their own clouds via Eucalyptus. Don't want Amazon to hold your keys? Load balance between Your cloud and Amazon's.
I'm curious if anyone has any predictions about what the landscape will be like in a few years? Will these be solved problems? Will cloud services lose favor? Will everything just be designed more conservatively? Will engineers finally learn to read the RTFSLA?