March 2023 Summaries

Redis Cloud has been updated with Redis Open Source 7.0 compatible features, including Redis functions and changes to Pub/Sub ACLs, in a limited release. The update is available on AWS and GCP for Fixed and Free subscriptions, and supports Redis functions, which offer several advantages over scripting with Lua, such as persistence and durability. The default ACL access to Pub/Sub channels has also been updated to be more restrictive by default, but the permissive approach remains supported for ease of upgrade. Developers can now create new Fixed subscriptions that support Redis 7.0 preview, and all databases created will be compatible with its features and capabilities. The rollout and availability of Redis 7.0 commands, functions, and capabilities will be expanded to new regions, and triggered functions are planned to be included as a preview within Redis Cloud later this year.

The upcoming changes to Redis Cloud's security practices involve replacing existing TLS certificates with new short-lived certificates issued by the publicly trusted GlobalSign Certificate Authority. This change aims to improve security, as the current self-signed certificates are not verified by a third-party authority. The transition will affect different tiers of Redis Cloud subscriptions at varying dates, depending on when their subscriptions were provisioned. Affected users should ensure that their Redis clients trust the new GlobalSign certificates and update their client's certificate store if necessary. Users can inspect the provided public certificate bundle to verify the certificates' authenticity and obtain any necessary updates for their Redis clients.

The financial services industry is struggling to keep up with the increasing number of fraudulent transactions and evolving tactics of cybercriminals. As digital banking usage increases, online fraud has also risen, with 51% of businesses experiencing fraud in the last two years, resulting in an estimated $42 billion loss. To address this issue, banks and financial services organizations are turning to digital identities, combining document verification, biometric records, and behavioral patterns to create a dynamic and complex identity for each user. However, updating information quickly enough to stay ahead of criminals without hindering the genuine user's experience remains a challenge. Machine learning algorithms and artificial intelligence predictive models can evolve and learn as they analyze and detect payment fraud based on historical and real-time transactional information, but successful implementation depends on accurate models, performance, and resiliency of underlying machine learning operations database. The industry needs to adapt to the digital age and move away from legacy relational database management systems to support modern AI/ML-based fraud detection and dynamic digital identities.

Redis Enterprise 6.4.2-30 introduces security features such as extended client certificate validation and pub/sub access management, while also enhancing the Kubernetes-based release with improvements for managing Redis Enterprise inside Kubernetes clusters. A public preview of Active-Active database deployments with Redis Enterprise for Kubernetes is now available, allowing users to experiment with this feature before its official release. This feature simplifies the process of deploying a Redis Enterprise cluster across multiple regions or data centers, enabling declarative configuration in a few lines of YAML code. The new Active-Active controller provides automated high availability and scalability for the database layer, making it easier to deploy and manage Redis Enterprise clusters.

The Redis Stack's introduction of t-digest, a new probabilistic data structure, offers an efficient way to estimate percentiles in large datasets. T-digest can be used to answer common questions about data streams and large datasets, including estimating the 50th, 90th, and 99th percentile. The data structure provides sub-millisecond latency, sub-linear memory requirements, and high accuracy, making it suitable for streaming data and huge datasets. With t-digest, users can query statistics such as percentiles, ranks, and trimmed means, allowing them to efficiently analyze and process large amounts of data. Additionally, the Redis Stack's support for t-digest extends its existing functionality, including RedisInsight, search and query, JSON, time series, and probabilistic data structures.

### Developers are expected to level up their baked-in security measures to prevent financial institution losses due to fraud. This is a tall order, but tools and software can assist developers in this task. Few developers realize the extent of the urgency of finding workable practices in Fintech, where fraud has always been a huge problem for banks and other financial institutions. Synthetic fraud, which involves building a false identity from one or two stolen data points, costs U.S. banks upwards of $20 billion annually. Protections built on static checklists and data are doomed to failure in today's fast-moving world, and designing more fluid user identification and authentication methods that can be checked in real-time is a better strategy. Developers need an idea of where to look for emerging or evolving patterns and should rely on AI and machine learning to detect patterns and anomalies indicating fraud. They also need to consider geographic biases, racial biases, gender biases, and other factors that may affect their systems' ability to identify fraudulent activity. By deploying AI to look for identity theft and synthetic identity account creation, and adding more layers of user authentication, developers can improve their organization's fraud detection capabilities.

Microservices architecture is a game-changer for organizations, allowing them to decouple business domains into their own respective services while maintaining open communication through APIs. This approach enables faster time to market as each service has its own release cycles. However, concerns about complexity, eventual consistency, and latency can deter companies from adopting microservices. To address these challenges, efficient use of a data platform for caching and a lightweight message broker for inter-service communication is essential. This enhances developer productivity, provides faster time to market, and streamlines the architecture of your application. Additionally, implementing Command Query Responsibility Segregation (CQRS) pattern can improve cross-domain data access with sub-millisecond query latency.

Redis OSS's logical databases provide a simplified approach to managing Redis instances, but they may become insufficient when functional and operational needs change. Four signs indicate that a single Redis instance is no longer sufficient: throughput-intensive use cases, O(n) Redis commands, centralized microservices' databases failing at once due to a bug, and unique configuration requirements. Dedicated instances can address these issues by allowing for better performance, resilience, scaling, and monitoring of each database without compromising on shared configurations or data consistency. Migrating logical indexes to dedicated databases requires manual effort, but automation tools are available to simplify the process. Redis technical teams offer assistance with planning migrations.

Redis Enterprise Cloud has released its Terraform resources for managing multi-region Active-Active Redis databases, making it easier to work with the popular infrastructure automation tool. The new Active-Active Terraform resources allow users to create, update and delete databases in any region at a push of a button, as well as define different configurations in each region. This feature is particularly valuable for applications requiring disaster recovery, geographically redundant applications, or serving data closer to users' physical locations.

This tutorial demonstrates how to optimize a brokerage application using Redis's JSON data structure and enhanced query capabilities. A brokerage's success is tied to its investors' engagement, and creating better investor engagement leads to more assets under management, trades, and fees/commissions. The application faces scalability and performance challenges when trading activity surges, requiring high availability, low latency, strong consistency, scalability, security, and consistent performance. Redis Enterprise supports JSON, graph, and time series data structures, offering features like Active-Active Geo-Distribution to meet demanding needs. A sample implementation models brokerage entities using JSON, storing and retrieving securities portfolios with Redis indexing. The application can retrieve security lots owned by an account, filter views, and calculate total quantities or average cost prices of securities. Redis Enterprise provides full support for JSON, real-time indexing, fast access, and atomic operations for JSON values, outperforming competitors in metrics like latency and read-write throughput.