March 2020 Summaries

Here's a neutral and interesting summary of the text in one paragraph: The author explores how graph databases like RedisGraph can be used to model complex relationships in data, using the example of Dungeons & Dragons dungeons. Graphs are composed of nodes with labels and attributes, and edges that represent relationships between them. This allows for efficient querying of the data, such as finding all monsters in a room or all rooms guarded by a monster. Graph databases have several strengths, including their ability to model complex relationships without a schema, making it easier to work with them once they are in production. The author demonstrates how to create and query nodes and edges using Cypher, the query language used by RedisGraph, and encourages readers to explore the capabilities of graph databases further.

Redis Enterprise has been integrated into Google Cloud Console, allowing users to enjoy unified billing and use their Google Cloud credits to pay for Redis Enterprise services. To enable the Redis Enterprise API in a Google Cloud project, navigate to the sidebar of your Cloud Console, click on Redis Enterprise, set up billing information, and enable the API. Once done, you can create Redis Enterprise databases. The integration also provides access to sample applications for testing purposes.

A large online game like World of EverCraft can benefit from using a Bloom filter to efficiently check for the existence of usernames without having to query the entire database, instead relying on probabilistic data structures that offer faster and smaller data structures in exchange for some uncertainty. By controlling the error rate and bit width, users can mitigate this uncertainty and gain performance and compactness at the cost of limited accuracy. The Bloom filter implementation is surprisingly easy to understand and implement, with an array of bits used to store indices generated by hashing functions applied to user input, allowing for fast lookups and insertions while minimizing memory usage.

The Redis community is invited to participate in the Redisdiscover Redis Competition, which aims to showcase examples of how developers are leveraging Redis beyond its traditional caching capabilities. The competition encourages submissions of projects that demonstrate Redis' versatility and potential, with a focus on "rediscovering" Redis and unlocking its full capabilities. Participants can submit videos or other forms of entries, with the top three submissions receiving prizes including project feature in the keynote at RedisConf Takeaway and a Valve Index VR Kit. The competition runs until May 8, 2020, and all valid submissions will result in a $100 donation to the COVID-19 Solidarity Response Fund. Winners will be announced on May 12 during the RedisConf keynote.

The author of the article discusses seven "worst practices" when using Redis, a popular in-memory data store. The first practice is not setting a password on the Redis instance, which makes it vulnerable to unauthorized access. Another worst practice is using the KEYS command, which can be slow and clog the server as the number of keys increases. Numbered databases are also criticized for being isolated but not well-supported throughout the ecosystem. Hashes with many fields and values can cause performance issues, similar to lists with a large number of elements. Redis connections should remain open across multiple commands, rather than opening and closing them per command. Creating hot keys that access the same data repeatedly can lead to performance issues and make the cluster less efficient. Finally, using Redis as a primary database requires additional features such as high availability and durability, which need to be configured properly. The author encourages readers to avoid these worst practices and follow best practices for optimal performance and security.

To address the coronavirus (COVID-19) pandemic, Redis has reiteratated its commitment to customer health and safety while ensuring reliable service and support. The company has developed a business continuity plan over the past eight years, with measures in place to ensure seamless delivery of service and support. With five offices on three continents and employees working remotely, Redis is taking steps to maintain connectivity and productivity through virtual collaboration tools. The company's managed database-as-a-service offering, Redis Enterprise Cloud, provides high levels of resiliency and availability, ensuring 99.999% uptime and uninterrupted operations. To contain the virus, Redis has moved its annual conference to a virtual event, providing an opportunity for the community to connect remotely.

The latest version of RediSearch, 1.6, introduces significant improvements in performance compared to its predecessor, 1.4. This is achieved through refactoring the original code and adding new features such as aliasing, a low-level API, and improved query validation. Additionally, Fork garbage collection has been made the default module. The performance boost leads to better user experience with applications becoming more reactive than ever before during search operations. To measure this improvement, a full-text search benchmark (FTSB) was created which allows developers, system architects, and DevOps practitioners to find the best search engine for their workloads. RediSearch 1.6 increased simple full-text search throughput by up to 63% while cutting latency (q50) by up to 64%. For aggregate queries, throughput increased from 15% to 64%. The FTSB also helped in monitoring performance between different versions of RediSearch and eliminating bottlenecks.

RediSearch 1.6 offers significant performance improvements, including up to a 73% boost and reduced latency of up to 64%, making it faster than its predecessor, RediSearch 1.4. The new version introduces aliasing, a low-level API that allows other Redis modules to use RediSearch's capabilities, enhancing the flexibility and power of development with RediSearch. Aliasing enables transparent redirection of application queries from one index to another without downtime. Additionally, RediSearch 1.6 includes improved query validation, better garbage collection, and support for polygon search, aiming to further improve developer experience and performance. The new version is also the fastest RediSearch yet, reducing latencies and peaks across all types of queries.

RedisTimeSeries 1.2 has been benchmarked using the Time Series Benchmark Suite (TSBS) against its predecessor, version 1.0.3, to demonstrate its performance advantages. The benchmarks showed that RedisTimeSeries 1.2 outperforms version 1.0.3 in terms of throughput and query latency, with minimal degradation even when cardinality increases or compression is introduced, consuming only 5% more CPU cycles. The results also indicate that RedisTimeSeries can handle large datasets efficiently, reducing memory consumption by up to 94%, making it the best option for working with time-series data in Redis.

The RedisTimeSeries module has reached general availability with version 1.2, which introduces powerful new features such as compression that reduces memory requirements up to 90% and improves performance by up to 70%. The compression technique used is based on double-delta and XOR encoding of timestamps and values respectively, resulting in a significant reduction in memory usage. The API has been made more consistent and intuitive, with redundant response data removed to further improve performance. Other enhancements include the ability to model counters for time windows of fixed sizes and the introduction of low-level APIs for full-text search queries and cross-time series aggregation.

A e-commerce website aims to provide a fast and easy-to-use purchase history for its customers, but faces challenges in scaling its index due to the large number of users and frequent updates. To address this, the website considers using a dynamic search index that is populated with user-specific data when the user logs in and expires after a certain time. This approach allows for efficient searching within each user's purchase history without maintaining a massive global index. The Redis Search and Query feature can be leveraged to create such indexes, which are ephemeral and can be easily managed and scaled.

A Brazilian software agency, Z3 Works, is using a microservices architecture to design and build software projects across various industries, including home-security companies and retailers. For example, their team built an application for a home-security company that collects data from video cameras and doorbells, providing real-time notifications to clients. The microservices approach allows the team to work on small components, update code frequently, and deploy applications effectively without crashing everything else. Redis is used as a caching layer to speed up the application's performance and as a message broker for image requests, ensuring that the system can process large amounts of data in real-time. By using a microservices architecture with Redis, Z3 Works can scale the client's application as needed, making it more effective in the long run, and avoid bottlenecks. The agency is now transitioning another retailer to a similar approach, combining multiple services for improved performance and scalability.

The virtual event, RedisConf 2020 Takeaway, will deliver inspiring content, including keynotes, training, and breakout sessions, to the Redis community. Due to concerns surrounding COVID-19, the in-person conference has been transformed into a virtual event. Registration fees for the original event will be refunded automatically, and attendees will be registered for the virtual event. The exact details of the virtual experience are still being worked out, but more information will be provided on the FAQ page.

The concept of "multi-cloud" is often misunderstood and misused, with many companies adopting a multi-cloud strategy without fully understanding its implications. A true multi-cloud implementation involves running different workloads in different clouds to take advantage of specific services and features from each provider. This approach can make it easier for companies to maintain relationships with multiple cloud providers and move quickly in any direction at any time. However, it also comes with increased costs and management complexity, as companies must deal with multiple vendors and technologies. The rise of microservices architectures and containers is helping to solve some of these portability issues and accelerate application deployment. Additionally, the concept of hybrid cloud, which combines public and private clouds or on-premises data centers, is often related but distinct from multi-cloud. As cloud vendors continue to differentiate themselves, multi-cloud approaches are expected to become even more widespread and complex.