AWS: us-east DC problems
Amazon Web Services is currently experiencing a problem with datacenters on the east coast of North America.
And on the net (bloggers) the question of the failure of cloud computing immediately arose.
Is the Internet at risk of crashing? Foursquare, Reddit and Quora fell with Amazon
Here is an example above, where we talk about crashes. Amazon’s transparency, which keeps us updated in real-time on the status of its services around the world, speaks of latency problems on EC2 and various errors on EBS.
I am very surprised to hear about the failure of cloud computing, hardware problems happen to everyone from small to large, amazon offers certified services and if the SLA is not guaranteed it repays the lack, but at the same time amazon recommends using its services well, if I entrust one of my sites to a single EC2 instance, this service does not guarantee the HA, but it has its own SLA, if I want the HA with AWS I design the infrastructure with at least two EC2 in two different datacenters with an LB to manage the load, if I keep my essential data on EBS, maybe I have not understood how AWS services should be used.
If it happened to a classic hosting, no one would have noticed it and maybe the provider would not even have noticed, or would not have made the necessary technicians available to restore everything.
A few weeks ago google lost thousands of gmail accounts and worked day and night to restore them even taking them from backup tapes (I didn’t imagine they were still used), obviously it was not kept on all of them as many accounts were free.
Cloud computing is a convenient model to make even huge portals without having to build a datacenter, see for example the farmville game that uses about 12000 EC2, I bet it has not been affected by the problems as they will have distributed the load well geographically. No one defends us from a disaster, only the replication of services in locations far from our primary site, in the same way any provider is subject to hw problems or even disaster. When Amazon offers high HA availability services (such as EC2 replicated real-time as in vmware’s fault-tolerance) and problems arise here too, then it will be to be condemned.
