January 2024 Summaries

In January, Redis released several new features and enhancements across its platforms. Key updates include JSON support in Active-Active deployment for Redis Cloud, multiple mTLS certificates for less downtime during maintenance windows, a new availability zone labeling feature, database tagging, a Billing Admin role, a new selector on the databases page, Active-Active support for Redis Cloud customers using AWS Transit Gateway in public preview, and integration with Confluent. Additionally, Redis Insight introduced a dedicated developer enablement area, is now available on Docker, and version 7.4.2 brought enhancements focused on improving the developer experience (DX).

The latest version of Redis OM .NET now supports Redis vector search and integrates with embedding generation APIs from OpenAI, Azure OpenAI, Hugging Face, and ML.NET, enabling users to efficiently store and query vector data without requiring extensive knowledge of vector construction. This allows for semantic searches and caching capabilities, making it easier to implement AI-powered applications that rely on vector databases. With the introduction of intuitive interfaces for vector search and semantic caching, users can easily build and query vector indexes, convert unstructured data into vectors, and construct queries using simple lines of code.

VectorFlow is an open-source vector embedding pipeline that helps preprocess data and ingest it into a vector database, alleviating the stress and time required for manually performing data chunking and vectorization tasks. It can be used with multiple embedding models, including third-party models like OpenAI embeddings, and can execute any Hugging Face Sentence Transformer model without requiring an API key. VectorFlow can run locally or use its free cloud offering, and it is compatible with various file formats such as images and text. By combining VectorFlow with Redis, a vector database, users can create a pipeline that ingests data into the Redis vector database, segments the data, converts it to vector embeddings using a chosen model, and stores them in the Redis database. The process involves setting up VectorFlow locally, connecting to the Redis database, establishing an index name and schema, defining metadata for the embeddings, downloading data, adding it to a dictionary, making a POST request to the embed endpoint, and checking the status of the task. With this pipeline, users can successfully ingest their embeddings into Redis and use them for various tasks.

Redis Cloud has introduced its Essentials plan, starting at $5 per month, making it more accessible for users to power various use cases like caching, session storage, and AI applications. The fixed monthly pricing model simplifies budgeting for projects of all sizes. In addition to cost-saving benefits, Redis Cloud offers unique functionalities such as JSON, Search, and exclusive Vector database capabilities. For those with more advanced requirements, the Pro plan provides 99.999% availability and round-the-clock support. With its commitment to providing innovative features and evolving solutions for developers' needs, Redis Cloud aims to change the game in the cloud service industry.

In-Memory Data Grids (IMDGs) such as GemFire have been popular for high-performance, low-latency data access since the mid-2000s. However, with the evolution of cloud-native, microservice, and serverless architectures, Redis Enterprise has emerged as a more adaptable and cost-effective alternative. This is due to its simplicity, versatility, and unparalleled performance. Redis Connect, a data replication and change data capture (CDC) framework, can facilitate the migration from GemFire to Redis by simplifying one-time migrations as well as continuous replication. Large financial institutions have successfully migrated from GemFire to Redis Enterprise using Redis Connect, resulting in improved performance, ease of use, and optimized total cost of ownership. When considering a migration from GemFire to Redis Enterprise, key factors include deployment options (open source vs. commercial), compatibility with major cloud service providers, terminology differences, and the pros and cons of offline or online live cutover. Additionally, Redis Enterprise offers advanced features such as auto tiering, active-active geo-replication, and native support for AI applications, making it a strategic enhancement for modernizing infrastructure and unlocking the potential of real-time data processing and management.

In the latest "What's New in Two" episode, Redis enthusiasts can learn about new updates from December. Firstly, a private preview for AWS Transit Gateway for Redis Enterprise has been announced, allowing users to connect with virtual private clouds (VPCs) on any cloud. Secondly, Redis Cloud now offers low-cost Fixed Plans, making it the most affordable per gig compared to other providers. These new plans are packed with functionality and are cost-effective for those looking to start using Redis in a cloud environment.

This guide demonstrates how to use Redis PubSub channels with Bytewax for real-time data processing, creating a custom input source and aggregating clickstream data efficiently. The example uses Python scripts as the producer and listener, utilizing the redis-py library for basic operations. A custom input source is created using Bytawax, which reads from a Redis pubsub channel and processes messages in a more streamlined and pythonic way compared to the original listener script. The resulting pipeline can be used to process various types of data streams with minimal changes or modifications. Bytewax's flexibility allows for adaptive data processing requirements, making it suitable for diverse use cases. The example showcases how to create a windowed calculation using Bytawax, highlighting its capabilities and encouraging users to experiment and extend the pipeline logic.