First of all, I must say that I agree with almost all the negative
said about new pricing policy in this thread (maybe except the PHP
thing - I don't think PHP would attrat by a lot of mature paying
developers, I think it's the opposite). Even though I understand
Google makes money, not charity, I cannot find any logic in the new
pricing policy - no one single positive or logical point, only
negative (and it's not that the new pricing policy makes things a bit
more expensive... it's that the new pricing policy makes things
excessively expensive with no arguments for it).


The main points I want to pay attention to:

= 15 minutes startup/shutdown+start fee =
I can't believe instances' startup/shutdown is that long - instances
are not even fully virtualized OSes, I bet they are started and
shutdown in hundreds of milliseconds (or at least in seconds)! I don't
believe instance is reserved when it is shutdown and then started back
in less than 15 minutes. Even if my beliefs are wrong, I can't
influence these problems and don't understand why I have to pay for
the issues googlers didn't manage to resolve.

= instance hours vs ram hours =
Provided official arguments (i. e. "optimize memory usage") for
instance hours make no sense. If I am told ram and cpu are not billed
and the only thing that is billed is instance hours the first things I
will do is fill up instance memory with cache, refactor all the code
optimized for latency and CPU to be CPU and memory intensive instead
and... make a cron job to disturb my app on a regular basis not to
restart instance. Free quota will be still enough for small apps but
those will stably consume more resources (constantly memory at max for
instance, CPU intensity all day long instead of short bursts that must
be simple to shard with those big numbers google have).
If you introduced new CPU hours and new RAM mbs/hours quotas instead
of old CPU hours or new instance hours, it would make much more
sense.
By the way, I may be thinking too much of googlers development
potential, but really... isn't memory overselled (http://
en.wikipedia.org/wiki/Overselling) in appengine?

= 50k quota for datastore ops =
That's the killer (will kill appengine :) feature of new pricing
policy. Vendor lock-in, nosql instead of sql (I can't say it's
worse... it's unaccustomed)... with new pricing policy you go even
further - files instead of nosql. Yeap, that's really what datastore
will be used like when one need to decrease number of datastore ops -
nearly pure old file write/read. By the way, memory consumption (both
because of bigger entities and memcache usage intensity) and cpu
intensity will increase (maybe that's the point you made instance
hours so expensive? predictions?).


Conclusion. I was looking closely at appengine, getting used to it for
nearly a year, wrote some testing apps. Two weeks ago I started real
business project for my employer with billed estimates of several
hundred dollars a month (not that much though). Our bad we didn't see
this thread earlier. With estimates of several thousand dollars a
month based on new pricing policy (especially instance hours +
datastore ops) the project will surely migrate to another cloud
hosting. Thanks we are not yet very much vendor locked-in and have
time to rewrite code. By the way, I will continue hosting very small
utility projects on appengine for my sole purpose consuming no more
than free quota.


Googlers, it seems to me you want the pie you saw in the sky.


Thanks for your time spent reading.

-- 
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