Does the upcoming changes to AppEngine billing mean asynchronous batch operations will be much cheaper than with the current billing model?

In the pursuit of efficient batch operating on AppEngine, I've been experimenting with async calls and have ended up with a request signature that looks like... /task/batch_remove_model 200 402ms 43582cpu_ms 42859api_cpu_ms.

I understand that under the current billing model I'd pay for the 43582ms of cpu time, but how would that same resource be billed under the upcoming 'instance based' billing model? At that point am I simply paying per datastore operation?

Thanks for your response开发者_如何学JAVAs.

You will pay for the 402ms of instance time used to server the request, plus you'll pay for each of the datastore operations. Datastore operations include entity reads and entity and index writes (not RPC calls), so the cost will be proportional (probably higher) to what you're paying now.

For more details see the pricing FAQ.