October 2017 Summaries

With Redis Enterprise, users can now extend their RAM-based storage into flash memory, enabling hybrid storage that moves data between fast RAM and flash as needed. This allows for high-performance storage without requiring changes to code. To test the performance of Redis Enterprise on flash, users should use i3.2xlarge AWS instances with NVMe SSD drives, and a memtier_benchmark setup that simulates realistic access patterns. The benchmark shows an average throughput of over 100,000 ops/sec at sub-millisecond latency, making it suitable for many real-world database use cases. Using flash as a RAM extension provides significant savings compared to running a RAM-only version with three times the infrastructure cost. Users can start a free trial and learn more about Redis Enterprise Flash architecture and performance characteristics.

Bee-Queue v1.0 is a fast, lightweight, robust Redis-backed job queue for Node.js that has been revived to make it the fastest and most robust Redis-based distributed queue in the Node ecosystem. The project was initially built with short, real-time jobs in mind and can scale by simply running more workers. Bee-Queue's release comes after Mixmax decided not to use a different queue like RabbitMQ due to its ease of use and performance benefits, instead opting for a custom Redis-based solution that has resulted in significant improvements over their previous queue Bull. The new queue is now hosted in a new Github organization and is maintained by Mixmax alongside the original author Lewis Ellis, with the goal of keeping it small with performance and stability as top priorities.

Redis Enterprise Flash technology combines RAM and flash storage in Redis to store large data sets at a lower cost than traditional options. This approach allows for more efficient use of resources, reducing infrastructure costs while maintaining performance advantages. The architecture is based on identical nodes deployed within a data center or across availability zones, with a management path handling proxy and cluster manager, and a data access path composed of master and slave Redis shards. Each database has a quota of RAM that can be extended to local flash drives using Redis Enterprise Flash, allowing for more flexible storage options. The technology uses in-memory replication to maintain high availability and scalability, while also providing durability through disk-based or replication-based options. By leveraging smart data placement and efficient IO patterns, Redis Enterprise Flash reduces the need for disk writes and optimizes performance.

Redis Enterprise 5.0 introduces a new flexible, multi-master replication technology built for WAN, utilizing conflict-free replicated data types (CRDTs) to simplify development of active-active systems and automatically resolve conflicting writes. CRDTs allow for intelligent handling of conflicting writes in geo-distributed deployments, enabling the creation of scale-minimized geo-distributed Redis clusters with active-active access. The new feature is designed to facilitate the development of active-active systems that can handle conflicting writes, providing a mechanism to easily develop such systems that can intelligently handle conflicting writes. A hands-on walkthrough demonstrates how to experiment with CRDTs in Redis Enterprise 5.0, including setting up a geo-distributed deployment and testing connectivity to the database using redis-cli. The new feature simplifies development of concurrent distributed writes to data by automatically resolving conflicts, allowing for accurate final values in distributed systems.