On Wed, Aug 8, 2012 at 9:24 AM, vlad <[email protected]> wrote: > Jeff, what storage subsystem did you use on your dedicated host? If you kept > all data in RAM then it is just a sleuth of hand. Overall I am not surprised > about lame performance of backends and pull queues. I am a bit surprised on > the cost of doing it in Datastore. Good work!
Why do you call it sleuth of hand? Storing the data in RAM represents the actual business need. We're doing not only an apples-to-apples comparison but an "actual implementation of business case to actual implementation of business case". Persistent RAM is one of the principal values of a Backend, and it turns out that they aren't very useful for that purpose. Not only is QPS low, but if you saw my post in Brandon's thread, latency is erratic and horrible. Just to give some latency numbers: Urlfetches to the rackspace VPS are consistently 30-50ms. Not great, but tolerable. Urlfetches to a backend average about the same - they skew more towards 15-25ms, but there are a large number of 250ms+ outliers that push up the arithmetic mean. No-op fetches F1 to B1: https://img.skitch.com/20120808-nn749683wqdg5516fy732k71iw.jpg No-op fetches F1 to B8: https://img.skitch.com/20120808-gerejtgfd958i691hsiw897wjw.jpg The test F1 to Rackspace: https://img.skitch.com/20120808-npj3kutc2p511bw5gujxkm5kg4.jpg BTW I picked Rackspace for this test because I happened to already have a VPS there. I would love to see latency numbers for Linode, EC2, and Google Compute Engine. Does anyone want to run the test? I'm still waiting for an invite to GCE. Jeff -- 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.
