#valkey7.2

Google is excited to launch Valkey 8.0 in preview form on Memorystore today, making Google Cloud the first significant cloud provider to provide Valkey 8.0 as a fully managed service. This strengthens Google Cloud’s dedication to open source, building on the release of Memorystore for Valkey 7.2 in August 2024 and giving you access to the newest and best features from the Valkey open-source community.

Major League Baseball (MLB) and other customers are examples of Google cloud dedication to customer service, as demonstrated by Valkey 8.0 on Memorystore. MLB, the most storied professional sports league, processes enormous volumes of data in real-time to give fans insights and statistics during games using Memorystore.

The release of Valkey 8.0

The open-source community banded together to establish Valkey a fully open-source alternative under the BSD 3-clause license, earlier this year after Redis Inc. altered the license of Redis OSS from the permissive BSD 3-Clause license to a restrictive Source Available License (RSAL). The open-source Valkey 8.0 was made available to the public in a matter of months by the Valkey community, demonstrating the strength of unrestricted innovation and open-source cooperation.

The latest version of the open-source NoSQL in-memory data store, Valkey 8.0, was released today, according to a statement from the Linux Foundation. With major upgrades intended to improve performance, dependability, and observability for all installations, Valkey 8.0 demonstrates quick innovation. By adding new features that elevate Valkey and building on the qualities of its past open source releases, this release strengthens the foundation.

The release’s main highlights are as follows:

Throughput on AWS r7g instances is improved up to 1.2 million requests per second with intelligent multi-core usage and asynchronous I/O threading, more than three times greater than the previous version.

Enhanced cluster scalability with replicated migration states and automated failover for additional shards

Dual-channel RDB and replica backlog streaming for faster replication

Granular visibility into performance and resource utilization is provided by extensive per-slot and per-client metrics, which include pubsub clients, rehash memory, event loop latency, and command-level heavy traffic recording.

Reduced memory overhead by up to 10% with improved key storage.

Valkey 8.0 may be downloaded right now at valkey.io, and those who already have it can update with a few quick keystrokes. Valkey can be installed using pre-made container images or constructed from source.

As a fully Google managed service, Memorystore for Valkey 8.0 offers better performance, increased dependability, and complete compatibility with Redis OSS.

Wth to recently introduced asynchronous I/O capabilities, the Valkey performance benchmarks have improved. By allowing the main thread and I/O threads to run simultaneously, the improved I/O threading system maximizes throughput by minimizing bottlenecks in handling incoming requests and enabling parallel processing of commands and I/O operations. Compared to Memorystore for Redis Cluster, Memorystore for Valkey 8.0 achieves up to a 2x Queries Per Second (QPS) at microsecond latency, enabling applications to handle higher throughput with clusters of same size. Because of this, Valkey 8.0 is an excellent option for real-time, high-throughput applications that need to deliver very rapid user experiences.

Valkey 8.0 has further optimizations in addition to the throughput gain that improve the service’s overall speed:

Set union procedures can be completed more quickly with the SUNION command.

To achieve faster execution speeds, changes have been made to the ZUNIONSTORE and SDIFF commands.

For expired keys, the DEL command prevents redundant deletions.

Responses from CLUSTER SLOTS are cached to improve throughput and lower latency during cluster activities.

Large data batches benefit from increased CRC64 performance, which is essential for RDB snapshot and slot migration scenarios.

Enhancements to key-memory efficiency are another feature of Valkey 8.0 that let you store more data without having to modify your application. Performance has improved and memory overhead has decreased because to the direct embedding of keys into the main dictionary. Furthermore, the main dictionary is divided into slots by the new per-slot dictionary, which reduces memory overhead by an additional 16 bytes per key-value pair without compromising performance.

Meanwhile, a number of Google-developed innovations that were later added to the project have increased the stability of Valkey 8.0 and greatly improved cluster resilience and availability:

Even in the early phases of scaling, automatic failover for empty shards contributes to high availability by facilitating the seamless failover of newly created, slotless shards.

Replicating slot migration states lowers the chance of data unavailability during failover events and allows new replicas to immediately inherit the right state. It also helps ensure that all CLUSTER SETSLOT commands are synchronized across replicas prior to execution on the main.

Furthermore, slot migration state recovery guarantees that, during a failover, the source and destination nodes are immediately updated, preserving precise request routing to the appropriate primary without the need for operator interaction.

These improvements increase the resilience of Valkey 8.0 clusters against failures during slot transfer, providing clients with the assurance that their data will remain accessible even during intricate scaling processes.

Compliant with Redis OSS version 7.2

Similar to Valkey 7.2, Redis OSS 7.2 APIs are fully backwards compatible with Valkey 8.0, facilitating a smooth transition away from Redis. Well-known Redis clients like Jedis, redis-py, node-redis, and go-redis are completely supported, negating the need for application code changes when transferring workloads to Valkey.

Because Valkey combines the flexibility of open-source software with the dependability of managed services, it offers you an ideal mix of control and simplicity when it comes to your Redis OSS workloads.

Start using Valkey 8.0 on Memorystore

Google cloud cordially encourage you to begin using Valkey 8.0 on Memorystore right now so you can see the aforementioned improvements. Memorystore’s Valkey 8.0 offers the performance, stability, and scalability that today’s high-demanding applications require with features like zero-downtime scaling, high availability, and RDB snapshot and AOF logging based persistence.

Begin by setting up a fully managed Valkey Cluster using the gcloud or Google Cloud console, and become a part of the expanding community that is influencing the direction of truly open-source data management.