Google today announced the beta of Memorystore for Memcached, a new service that provides a fully managed in-memory datastore that is compatible with the open-source Memcached protocol. It will join Redis in the Memorystore family, which first launched in 2018.
As Gopal Ashok, Google’s product manager for Memorystore notes in today’s announcement, Redis remains a popular choice for use cases like session stores, gaming leaderboard, stream analytics, threat detection and API rate limiting, while Memcached is typically used as a caching layer for databases. Developers also regularly use Memcached as a session store and with this new service, developers can scale their clusters up to 5TB of memory per instance.
Since the service is fully compatible with Memcached, developers should be able to take any of their applications that use the protocol and migrate them over to Google Cloud and its Memorystore platform. As a fully managed service, Google will handle all of the routine tasks like monitoring and patching. Figuring out the right size of a cache remains a bit of an art, though, but Google Cloud argues that its detailed metrics will allow developers to easily scale their instances up and down as needed to optimize the service for their specific use cases. Those metrics, the company notes, are exposed in Cloud Monitoring, Google Cloud’s centralized monitoring dashboard, and the Cloud Console.
Currently, Memorystore for Memcached can be used for applications that run on Compute Engine, Google Kubernetes Engine (GKE), App Engine Flex, App Engine Standard and Cloud Functions.
It’s worth noting that Amazon, with ElastiCache for Memcached, and specialized startups like MemCachier. And Redis Labs, too, is offering a fully managed Memcached service that can run on AWS, Azure and Google Cloud.