Google App Engine developers can now access dedicated cache up to 20GB of memory and split the application in modules providing stateless and secure services.
Dedicated Memcache
Normally, applications running on GAE use a shared memory cache with an upper limit of only 1MB, but that capacity is not guaranteed and it is not covered by an SLA. The advantage is that it is free. Google now introduces a cache service that can be used to provision up to 20GB of dedicated cache with a performance of up to 10k ops/sec/GB for $0.12/GB/hour.
The dedicated cache service is currently available only in US in preview mode and without an SLA. An SLA is to be expected to be provided when the service becomes generally available.
Modules
Google App Engine 1.8.2 introduces the option to divide a larger application into separate modules that can share stateless services in a secure way. Module functionality is available for all languages supported.
When running, each module may have multiple versions and multiple instances per version. Each instance is accessed through a unique URI, service requests from clients being dispatched to the appropriate instance based on URL addressing conventions and a dispatch file. Instances may be created when needed depending on the scaling type, the following being available:
- Manual – the module runs continuously.
- Basic – an instance is initiated when a request comes in and it is terminated when the application is idle.
- Automatic – the general scaling type used by GAE; instances are started and stopped based on usage and configuration.
Push-to-Deploy
Introduced with GAE 1.8.1, Push-to-Deploy now offers a simpler way to deploy Python and PHP applications stored in a Git repository. When a branch is pushed to the master, the code is automatically deployed on GAE. Complete details for setting up and using this feature can be found on this documentation page.
Other
The Google Plug-in for Eclipse now fully supports the Web Tools Platform and EAR files, enabling developers to use all sorts of plug-ins for Eclipse.
The PHP runtime is better integrated with Cloud Storage.
The Python interpreter has been updated to 2.7.5.
There are other various small improvements and bug fixes across all languages supported.