Jump to Content
Google Cloud

Introducing Cloud Memorystore: A fully managed in-memory data store service for Redis

May 9, 2018
Gopal Ashok

Product Manager, Memorystore

At Redisconf 2018 in San Francisco last month, we announced the public beta of Cloud Memorystore for Redis, a fully-managed in-memory data store service. Today, the public beta is available for everyone to try. Cloud Memorystore provides a scalable, more secure and highly available Redis service fully managed by Google. It’s fully compatible with open source Redis, letting you migrate your applications to Google Cloud with zero code changes.

As more and more applications need to process data in real-time, you may want a caching layer in your infrastructure to reduce latency for your applications. Redis delivers fast in-memory caching, support for powerful data structures and features like persistence, replication and pub-sub. For example, data structures like sorted sets make it easy to maintain counters and are widely used to implement gaming leaderboards. Whether it’s simple session caching, developing games played by millions of users or building fast analytical pipelines, developers want to leverage the power of Redis without having to worry about VMs, patches, upgrades, firewall rules, etc.

Early adopters of Cloud Memorystore have been using the service for the last few months and they are thrilled with the service.

At Descartes Labs, we have long been fans of Redis and its high performance. We have used Redis on everything from storing asynchronous task queues for tens of thousands of CPUs to a centralized persisted key-value pair store for the feature vectors output by our ML models. Cloud Memorystore provides an agile, scalable, no-operations Redis instance that we can instantly provision and scale without administration burdens.

Tim Kelton, CoFounder and Cloud Architect, Descartes Labs

Cloud Memorystore has provided us with a highly reliable Redis service and has been powering our critical applications. We have been using Cloud Memorystore as an early adopter and we are impressed with the reliability and performance of the service. Google has helped us forget about our Redis instances with Cloud Memorystore and now we can focus more time on building our applications.

Cristian Bîrzan, Software Developer, Ship.Cars

https://storage.googleapis.com/gweb-cloudblog-publish/images/google-cloud-memorystore-feature-summaryl4bz.max-700x700.PNG

Simple and flexible provisioning

How you choose to deploy Cloud Memorystore for Redis depends on the availability and performance needs of your application. You can deploy Redis as a standalone instance or with a replica to provide high availability. But while replicating a Redis instance provides only data redundancy, you still need to do the heavy lifting of health checking, electing of a primary, client connections on failover, etc. The Cloud Memorystore service takes away all this complexity and makes it easy for you deploy a Redis instance that meets your application’s needs.

Cloud Memorystore provides two tiers of service, Basic and Standard, each with different availability characteristics. Regardless of the tier of service, you can provision a Redis instance as small as 1 GB up to 300 GB. With network throughput up to 12 Gbps, Cloud Memorystore supports applications with very high bandwidth needs.

Here is a summary of the capabilities of each tier:

https://storage.googleapis.com/gweb-cloudblog-publish/images/google-cloud-memorystore-summary-capabilitie.max-700x700.PNG

Provisioning a Cloud Memorystore instance is simple: just choose a tier, the size you need to support the instance availability and performance needs, and the region. Your Redis instance will be up and running within a few minutes.

https://storage.googleapis.com/gweb-cloudblog-publish/images/google-cloud-platform-memorystoreks2x.max-700x700.JPEG

“Lift and shift” applications

Once provisioned, using Cloud Memorystore is a breeze. You can connect to the Redis instance using any of the tools and libraries you commonly use in your environment. Cloud Memorystore clients makes use of IP addresses to connect to the instance. Applications always connect to one IP address and Cloud Memorystore ensures the traffic is directed to the primary in case there is a failover.

Other key features

Whether it’s provisioning, monitoring or scaling memory, Cloud Memorystore simplifies common management tasks.

Security

Open-source Redis has very minimal security, and as a developer or administrator, it can be challenging to ensure all Redis instances in your organization are protected. With Cloud Memorystore, Redis instances are deployed using a private IP address, which prevents the instance from being accessed from the internet. You can also use Cloud Identity & Access Management (IAM) roles to ensure granular access for managing the instance. Additionally, authorized networks ensure that the Redis instance is accessible only when connected to the authorized VPC network.

Stackdriver integration

Cloud Memorystore instances publish all the key metrics into Stackdriver, Google Cloud’s monitoring and management suite. You can monitor all of your instances from the Stackdriver dashboard, and use Stackdriver Logging to get more insights about the Redis instances

https://storage.googleapis.com/gweb-cloudblog-publish/images/google-cloud-platform-memtorystore-dashboard.max-700x700.PNG

Seamless memory scaling

When a mobile application goes viral, it may be necessary to provision a larger Redis instance to meet latency and throughput needs. With Cloud Memorystore you can scale up the instance with a few clicks, and the Standard High Availability tier lets you scale the instance with minimal disruption to the application.

On-demand pricing

Cloud Memorystore provides on-demand pricing with no upfront cost and has per second billing. Moreover, there is no charge for network traffic coming in and out of a Cloud Memorystore instance. For more information, refer to Cloud Memorystore pricing.

Coming soon to Cloud Memorystore

This Cloud Memorystore public beta release is just a starting point for us. Here is a preview of some of the features that are coming soon.


We are excited about what is upcoming for Cloud Memorystore and we would love to hear your feedback! If you have any requests or suggestions, please let us know through Issue Tracker. You can also join the conversation at Cloud Memorystore discussion group.

Sign up for a $300 credit to try Cloud Memorystore and the rest of GCP. Start with a small Redis instance for testing and development, and then when you’re ready, scale up to serve performance-intensive applications.

Want to learn more? Register for the upcoming webinar on Tuesday, June 26th 9:00 am PT to hear all about Cloud Memorystore for Redis.

Posted in