|App Engine graduates and charges apply|
|Written by Lucy Black|
|Thursday, 10 November 2011|
Google App Engine 1.6.0 is now out of preview and has the status of a fully supported Google product with a charging structure according to on-demand frontend/backend instances.
App Engine had been "experimental" for three and a half years and started out with the philosophy that it should be "easy to use, easy to scale, and free to get started". As a result it was widely adopted - according to its own blog it achieved 100 billion+ monthly hits, 300,000+ active apps, and 100,000+ developers using it every month.
Many of those developers are now disgruntled by billing on its new pricing structure that has just come into effect, along with revised term of service and a service level agreement, Google seems to be trying to appease developers with "business friendly" terms, including a 3 year deprecation policy (no pulling out the rug for commercial products in the way they have been doing for Labs ones) and 99.95% SLA for-paid applications on the High Replication Datastore. But this hardly makes up for having to pay for something that's been free for such a long time and then changing the basis for charging.
Several other changes were also included in this week's announcement. One was that the App Engine SDK now finally supports the Python 2.7 runtime. Also with regard to Python, Google is releasing the full MapReduce framework comprising the Map, Shuffle, and Reduce phases.
For Java, the Memcache API now supports asynchronous calls. Additionally, putIfUntouched() and getIdentifiable() now support batch operations. Facilities for capability testing has been added as have datastore callbacks.
With regard to charging, developers can now use the Blobstore API without signing up for billing. For billing enabled apps, Google is providing two additional scheduler controls, Min Idle Instances and Max Pending Latency.
Also new is an experimental High Replication Datastore Migration tool for taking data from Master/Slave to High Replication Datastore, and switch an application's serving to the new HRD application.
|Last Updated ( Thursday, 10 November 2011 )|