April 2020 Summaries

The newly released Redis Enterprise 6.0 improves Redis security and operational safety with the addition of access control lists (ACLs) and role-based access control (RBAC), allowing for fine-grained control over user access to Redis deployments. These new capabilities bring major security benefits, making it easier to scale Redis usage within organizations. The release also incorporates many changes from open source Redis 6 and provides support for Redis Streams on Active-Active databases, enabling high-speed, in-memory streaming database functionality. With RBAC, users can create roles with specific permissions, allowing for a complete security-management solution for clusters. This enables scaling of Redis deployments while minimizing overhead involved in managing complex access control configurations.

Redis 6 has been made generally available and introduces several significant changes and features, including access control lists (ACLs) that allow for more fine-grained control over user permissions, improved security with SSL encryption, threaded I/O to boost performance, and a new client-side caching layer. The Redis protocol has also evolved with the introduction of RESP3, which offers richer results and paves the way for thinner client libraries. Additionally, Redis 6 includes a cluster proxy to simplify access to Redis clusters in smaller language communities, faster key expiration cycles, and a new command family for finding non-contiguous sequences among strings, known as the longest common subsequence (LCS) commands. These changes aim to make Redis easier to use, more secure, and more performant, opening up vast new possibilities for the Redis community.

RedisEdge is a multi-model database designed for the Internet of Things (IoT) edge, offering low latency and high performance in constrained compute environments. It bundles Redis with additional modules such as RedisAI and RedisTimeSeries to provide a purpose-built solution for IoT edge applications. Partnered with EdgeX Foundry and Microsoft Azure IoT Edge, RedisEdge simplifies data management for developers by providing a flexible multi-model database that supports various data structures, including time-series and graph models. With its compact footprint and support for 10 native Redis data structures, RedisEdge enables businesses to focus on insights instead of data management, removing the complexities of polyglot persistence architectures.

Redis Streams are a data structure in Redis that can efficiently handle time-directed and immutable data streams, such as those generated by IoT devices, making them suitable for the initial ingestion of large amounts of data. Redis Streams offer flexible data modeling options, allowing multiple instances of the same field to be present in an entry, which can help reduce storage space. The use of Redis Streams can lead to significant savings in overhead when dealing with large datasets, as demonstrated by a hypothetical example of 1 million entries with 20 fields each. The key benefits of using Redis Streams include constant-time access to entries and the ability to compress index IDs using a radix tree data structure. Additionally, Redis Streams allow for space-saving mechanisms such as storing only unique field names.

Redis' geospatial data structures are useful for working with location-based data. The key data structure is the Geo Set, which holds a named set of locations on the globe and can be easily updated using the GEOADD command. Other commands include GEOPOS for retrieving coordinates, GEODIST for determining distances between sightings, and GEORADIUS/GEORADIUSBYMEMBER for finding members within a specified radius. Geo Sets are implemented as Sorted Sets in Redis, allowing the use of additional Sorted Set commands such as ZSCORE, ZRANGE, and ZREM. These features can be used to track various types of location-based data, from mobile app users to fleet trucks or animals in an environmental study.

The artificial intelligence (AI) industry has experienced significant growth since 2016, driven by advancements in GPU technology and the need for faster model training. The focus has shifted towards deploying AI models to production and managing the entire AI lifecycle. A critical step in this process is AI serving, which involves deploying a task usually performed by an AI inference engine. To achieve fast end-to-end inferencing/serving, several challenges must be addressed, including optimizing AI processing, running the AI inference platform where data lives, and using a purpose-built serverless platform. By overcoming these challenges, businesses can benefit from running AI on dedicated inference chipsets and ensure a seamless user experience despite the potential slowness in the AI inference engine.

Redis Enterprise Software 5.6.0 is now generally available, featuring several key enhancements including support for HyperLogLog in Active-Active Redis databases, Redis on Flash (RoF) databases with Redis modules, and improved performance with the Redis OSS cluster API. The new version provides flexibility in customizing installation paths and user/group settings, as well as supporting remote users. Additionally, it adds support for RedisJSON, a module that implements the JSON Data Interchange Standard as a native data type, allowing users to store, update, and fetch JSON values from Redis keys. These changes aim to provide significant benefits to a wide range of Redis users, including improved performance, cost savings, and enhanced capabilities.

Redis University has launched a new YouTube channel called Redis Explained, offering short instructional videos focused on teaching developers about Redis data structures and popular modules. The series has received positive feedback from viewers and will continue to be updated regularly with more content. Redis University, established two years ago, now hosts five online courses and offers Redis Certification, providing various educational resources for users. The channel aims to cater to both beginners and experienced developers who need a quick refresher or want to learn about specific topics. Viewers can expect to find videos covering topics such as Sorted Sets Explained and Redis Hashes Explained in the future.

While microservices architectures often rely on Kubernetes for orchestration, having knowledge of the platform can still prove useful even in non-microservices scenarios due to its ability to maintain clusters and handle scaling requirements. This is particularly true when dealing with stateless applications, distributed caching, and sessions, where automation is crucial. Additionally, containers solve problems related to application portability and resource management, which may be addressed by newer technologies like WebAssembly. However, even if a new orchestration technology emerges, Kubernetes' fundamental building blocks remain universally useful, making it a worthwhile investment for medium-to-big organizations looking to streamline service deployment and orchestration.

The Jaccard similarity is a simple calculation that determines how similar two sets are, created by botanist Paul Jaccard in 1901. It's calculated as the ratio of the cardinality of the intersection to the cardinality of the union of the two sets. Redis, with its Set data structure, can be used to calculate this similarity using set operations such as union and intersection. However, storing the results of these operations can lead to performance issues, especially if the resulting sets are large. To avoid this, alternative approaches can be taken, but they come with tradeoffs. The Jaccard similarity can be calculated using Redis Set commands or Lua scripts, allowing for more control over the calculation and its performance implications.

The concept of a "day of deluge" refers to the maximum load that an organization's software systems experience, often coinciding with significant events or peak usage periods. This phenomenon can be caused by various factors such as natural disasters, holidays, or widespread adoption of remote work due to pandemics. The author, who worked at a financial services company, experienced literal catastrophes during hurricanes like Sandy and Katrina, while others face more predictable but still overwhelming loads on their systems. To address this issue, developers and architects must consider techniques such as scaling up hardware, scaling out hardware, statelessness, minimizing state, caching, queuing, microservices, autoscaling, and distributed data types like conflict-free replicated data types (CRDTs). These strategies can help organizations achieve near-linear scalability, reducing the impact of diminishing returns and plateaus.

RedisConf 2020 Takeaway is a virtual conference that brings together experts and contributors to share knowledge on Redis, data platforms, and related topics. The main conference will take place on May 12 with live keynotes, while online training sessions will be held on May 13. This year's event features an A-list group of speakers from various industries, including Microsoft, HashiCorp, Twilio, Zenerchi, DoorDash, and Gretel.ai. The lineup includes renowned experts such as Julia Liuson, Chris Richardson, Nicole Mitchell, Scott Demming, Carlos Reyes, Zohaib Ansari, and John McCallister, who will share their insights on topics like microservices architectures, securing networks with service meshes, machine learning fundamentals, creating models of human physiology, improving cache speed at scale, and joinability analysis. The conference offers a unique opportunity for developers and cloud professionals to learn from valuable experiences, share ideas, and get hands-on training in the comfort of their own homes. Registration is free, and attendees can expect an engaging and informative experience with RedisConf 2020 Takeaway.

RedisGraph 2.0 has introduced significant performance improvements, including latency improvements up to 6x and throughput improvements up to 5x. These enhancements are attributed to the incorporation of the latest version of SuiteSparse:GraphBLAS, which enables shared-memory parallelism via OpenMP. Additionally, RedisGraph 2.0 supports more Cypher features, allowing it to approach full support for the Linked Data Benchmark Council (LDBC) benchmarks and further optimize its performance.

RedisGraph is a noSQL database that enables users to store and manage highly connected data using graph technology. Originally released in 2018, RedisGraph has since gained significant popularity with over 1,000 GitHub stars and more than 100,000 Docker pulls. The latest version, RedisGraph 2.0, offers performance enhancements and new features such as full-text search on property values, support for graph-aided search, and the ability to return nodes, relations, and other data types directly. Additionally, RedisGraph 2.0 integrates with RedisInsight, Linkurious, and Graphileon, allowing users to visualize and analyze their data more effectively.

Redis University is pleased to announce the immediate availability of free, self-paced, on-demand online courses. The company has created these courses to help people make the most of their time at home during the current shelter-in-place period. Redis University offers a range of five courses that cover various aspects of Redis and its data structures, including an introduction to Redis, Redis Streams, full-text indexing, Java programming with Redis, and best practices for using Redis in Node.js applications. The courses are designed to be completed at the learner's own pace and will remain available until May 12, 2020. With this new self-paced approach, Redis University aims to make it easier and more convenient for developers to level up their Redis skills.

The discovery of over 8,000 unsecured Redis servers across various clouds highlights a persistent issue with internet-connected devices being vulnerable to cyber threats. This number has been declining since the introduction of protected mode in newer versions of Redis, which prevents external access to the server. However, many exposed servers still use outdated versions, indicating intentional disabling of security protections or human error. The upcoming version 6.0 of open source Redis will introduce Transport Layer Security encryption and a new Access Control List mechanism to improve security. Ultimately, users should carefully review their Redis configuration settings, especially when using open source software, and consider opting for the production-proven and security-hardened Redis Enterprise.

Redis is being used to power an action platformer game called Super Redis Brothers, which features a character named Redisman who collects coins and uses them to gain power-ups. The coin count of the character is stored in a Sorted Set data structure, allowing for efficient retrieval and update of scores. The use of Sorted Sets enables the implementation of leaderboards and player-matching systems, as well as tracking of individual players' scores. To manage potential issues with hot keys, strategies such as using multiple clusters or implementing a round-robin write approach can be employed. Redisman's power-ups, such as freezing enemies or reducing coin count to zero, are implemented using commands like ZINCRBY and ZADD, which allow for atomic updates of the score. The use of Lua scripts enables simple and efficient management of complex logic, while Sorted Sets provide a near-perfect data structure for leaderboards.