July 2017 Summaries

The text discusses the launch of new Redis meetups in Barcelona, France and Toronto. In Barcelona, Carlos I. Peña started a local Redis Meetup to bring together enthusiasts, while in France, François Cerbelle launched a set of meetups for French Redis Geeks across the country. In Canada, the Toronto Redis Meetup was launched to support the growing Redis community in Toronto. These meetups are looking for speakers and hosts, and interested individuals can join or volunteer by contacting the respective organizers.

The use of machine learning in real-time environments is a complex problem that has not received enough attention from framework developers. Most frameworks support the "learning" part of machine learning but lack support for the "predict" side. Redis-ML, a module built into Redis 4.0.0, aims to address this gap by providing a native datatype for linear regression and other machine learning algorithms. The Redis-ML module uses the new Modules API to add standard machine learning commands, making it easier to implement real-time predictions in applications. By using Redis as a predictive engine, developers can build fast and reliable applications that can make predictions based on data stored in Redis.

We, at Redis, are happy to announce the expansion of our Redise Cloud services to two additional regions on Google Cloud Platform. The new regions include Eastern US (us-east1), Central US (us-central1) and Northeast Asia (asia-northeast1). Our Redise Cloud service is a fully managed database-as-a-service solution that provides stable high performance, effortless scaling, robust high availability features, including persistence and automatic failover. Users can sign up for our free tier plan to get started with this service, which offers no data loss and is fully compatible with existing Redis applications.

Redis 4.0.0 GA has been released, featuring a number of new features such as Modules API, improved replication mechanism (PSYNC2), new caching policies, asynchronous deletion operations, microcontroller support, Redis Cluster improvements, and memory management improvements. The Modules API enables developers to extend Redis with new data types, operations, and processes. Additionally, the 4.0.0 release introduces least-frequently-used (LFU) maxmemory policies for improved caching performance and a new MEMORY command for memory introspection.

Redis Enterprise cluster achieved sub-millisecond latency for durable 'write' operations, but not for single durable operation under sub-millisecond latency on state-of-the-art cloud storage infrastructure like AWS io1 EBS. Despite the performance limitations of Redis in an ACID configuration, it can still achieve over 650K ACID ops/sec with high-end persistent storage devices like Dell-EMC VMAX. The use of a Dynamic Virtual Matrix interface and FlashBoost technology enables VMAX to scale from one engine to eight and deliver sub-millisecond latency for reads and writes. A single Redise cluster node can support over 650K ACID operations per second while maintaining sub-millisecond database latency with high-end persistent storage devices like Dell-EMC VMAX, which far surpasses the performance of state-of-the-art cloud storage infrastructure.