I see the new pricing as very positive. It would get rid of spamming apps 
which consume the resources of the
ordinary apps with this 9$ barrier. For some apps prices would actually go 
down and once the multhreaded
python runtime is ready the price difference would be very minimal because 
Python uses less RAM than Java.

As I've understood frontend instances are actually similar to B1 backend 
instances.  In other words they have
this 128MB RAM limit... so Java apps which mostly consume more RAM and are 
slower to recover it would
scale less in multithreaded mode. The scheduler would not serve requests to 
instances close to 128 RAM...
so would spawn a new instance. So a typical singlethreaded Java app with 
latency of 200ms would have
up to 5 QPS. In multithreaded mode this number would rise up until the 
instance gets near 100% CPU use
for this VM or reaches the maximum RAM which could be tested with similar B1 
backend instance.

As for requests that have latency more than 1s these should be optimized or 
put in the backend because
they destroy frontend scalability.

That's the way I see it ... time will show whether things will be that way.

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en.

Reply via email to