We have applications that do large batches of work - mapper-driven,
fantasm-driven, other custom fan-out driven. It is a very, very
powerful feature of App Engine to be able to scale out massively to
handle these large jobs in a single spike of work. No other platform
provides this capability and
Wow. So many fundamental design assumptions are being turned on their
heads with the new incentive model!!!
It is unfortunate that Google failed to make the 100% granular cost
model work. The promise that made Appengine attractive was: You build
an app (adhering to our limitations). We will
One more point. We all expected prices to increase. What was a
surprise is for the incentive model to flip as much as it did.
Maybe a compromise pricing model is not based on # of instances, but
on sum of response times? This would eliminate the customer paying
for inefficiencies in the
+1
Interesting suggestion on pricing as a sum of response times, which is what
a user should be worried about, not tinkering with the scheduler.
On Fri, May 20, 2011 at 7:41 PM, johnP j...@thinkwave.com wrote:
One more point. We all expected prices to increase. What was a
surprise is for
On Fri, May 20, 2011 at 3:11 PM, johnP j...@thinkwave.com wrote:
One more point. We all expected prices to increase. What was a
surprise is for the incentive model to flip as much as it did.
I expected prices to decrease. After 3 years of Moore's Law, why would
it cost more?
--
You
Very interesting question:
Is MapReduce still a flexible solution on AppEngine under the new
pricing model ?
My answer: probably not, new pricing model makes mapreduce operations
a no - no. Price will be prohibitive for such operation especially
ones that depend on many instances to run a job
