AppEngine pricing looks similar to Amazon EC2 pricing for single core instance
AppEngine - $0.08 / hour
Amazon EC2 - Small (Default) $0.080 per Hour
But it seems that I have to pay for time when my script is waiting for IO. I expected to pay for actual CPU cycles, not for waiting. When my script is waiting for DataStore or URL fetch then other scripts may execute on same core. In my test case 97-99% of time is script waiting for IO. So do I have to pay for 70-100 times more then I expected or would pay if I run on EC2 or a dedicated server?
I thought AppEngine is good for apps that scale but if I will have a high traffic then it would be a Kill Bill. Low hosting costs and a scalable platform are important. Are my estimations of AppEngine pricing comparing to Amazon EC2 correct?
You can turn on multithreadind for java or python 2.7 applications. In multithreading mode every instance can work with 8-10 requests at same time.
Python:
https://developers.google.com/appengine/docs/python/config/appconfig#Using_Concurrent_Requests
Java:
https://developers.google.com/appengine/docs/java/config/appconfig#Using_Concurrent_Requests