February 2022 Summaries

ScyllaDB is transitioning its database drivers to use Async Rust, aiming to unify all drivers under a single core for improved performance, safety, and maintainability. This transition began with the ScyllaDB Rust driver, which outperformed the reference C++ driver in benchmarks, prompting the idea to standardize on Rust to reduce redundancy in driver logic across multiple languages. By concentrating efforts on a central Rust implementation, ScyllaDB anticipates easier maintenance, fewer bugs, and enhanced performance, as Rust's asynchronous model and the chosen Tokio runtime offer high concurrency and low latency. Despite the advantages, challenges include the potential for widespread impact from bugs in the core and the complexity of language bindings, but the effort promises significant long-term benefits for users and developers.

The 2022 ScyllaDB Innovation Awards recognize organizations that have excelled in developing data-intensive applications using ScyllaDB, highlighting innovative use cases across various fields such as AI/ML, cybersecurity, and biomedical research. Notable winners include Palo Alto Networks for their high-performance database solution for network events, Instacart for their rapid implementation of a unified feature store for machine learning, and Janssen Pharmaceutical Companies for their AI-driven biomedical knowledge graph. IBM was recognized for smart growth by enhancing storage capacity without service disruption, while Happn achieved significant cost savings by migrating from Cassandra to ScyllaDB. China Mobile was acknowledged for their contributions to the Alternator DynamoDB API, and Meraj Rasool from SkyElectric was named ScyllaDB University Valedictorian for his dedication and application of knowledge from the courses. These achievements underscore the critical role of data in driving organizational success and innovation, as emphasized by ScyllaDB CEO Dor Laor.

Catalytic is an ORM designed for ScyllaDB and Cassandra, built using Rust, which aims to provide a fast and safe method for interacting with non-relational databases by leveraging Rust's features like compile-time checked queries and automatic serialization. Developed by Jasper Visser, Catalytic addresses common issues faced with manual coding such as run-time query errors and incorrect table-to-struct mapping by generating Rust structs from database metadata, ensuring seamless integration with ScyllaDB's capabilities. The project benefits from ScyllaDB's strengths, including its impressive scalability and support for materialized views, while utilizing the asynchronous capabilities of Rust's Tokio framework. Though not yet complete, Catalytic encourages contributions from developers to expand its type support and enhance its functionality.

Apache Cassandra, once a popular choice for handling massive amounts of data due to its horizontal scalability, high availability, and seamless replication, has been phased out by several companies due to various operational challenges. Despite its initial appeal, companies like Comcast, Fanatics, Expedia, Rakuten, and others faced issues such as node sprawl, garbage collection pauses, high latency, and infrastructure costs, prompting them to switch to alternatives like ScyllaDB. These organizations reported significant improvements in performance, cost efficiency, and system reliability after migration. For instance, Comcast experienced a tenfold improvement in latency and a drastic reduction in nodes, while Fanatics and Expedia saw major cost savings and smoother performance. The switch allowed these companies to address the limitations of Cassandra, such as JVM-related issues and unpredictable latencies, providing a more stable and efficient database environment tailored to their specific needs.

ScyllaDB Open Source 4.6 is a significant upgrade to the NoSQL database, introducing ARM support, improved node operations, and enhanced service-level properties for better session management. The release also brings advancements in reverse query performance, index caching, and installation support, alongside deprecations for older operating systems. Notable features include the ability to handle ARM architecture with EC2 ARM base AMI, the enablement of Repair Base Node Operations by default for replace-node tasks, and the introduction of service levels for more granular control over session properties such as timeouts and workload types. The update also enhances SSTable index caching, improves reverse query efficiency, and integrates several stability and performance improvements, including optimized read paths and off-strategy compaction for various operations. Additionally, the release addresses numerous bug fixes, security enhancements, and infrastructure updates, further solidifying ScyllaDB's capabilities in handling large-scale data operations efficiently and reliably.

The ScyllaDB Summit Day 2 offered a wide array of technical discussions and insights, with a focus on innovations in NoSQL databases and event-driven architectures. Key highlights included demonstrations of running ScyllaDB across multiple cloud platforms, discussions on Kubernetes enhancements, and explorations of ScyllaDB's Raft implementation for schema consistency. The summit also featured presentations on optimizing data pipelines for telecoms, benchmarking at petabyte scale, and the integration of ScyllaDB with Apache Kafka for high-velocity data sharing. Attendees were encouraged to engage with on-demand session recordings and participate in further learning opportunities through ScyllaDB University and other workshops. Additionally, the event fostered lively discussions in the Speakers Lounge and on social media, emphasizing the community's enthusiasm and collaborative spirit.

ScyllaDB Summit 2022 showcased the latest innovations and strategies in handling emerging data challenges through a virtual conference attended by engineers, architects, and developers worldwide. The event highlighted the capabilities of ScyllaDB, a fast and scalable NoSQL database, with discussions on its improvements such as the new Raft-based consensus algorithms for immediate consistency and various optimizations for enhanced performance. Keynote speakers, including ScyllaDB co-founder Dor Laor, addressed the demands of data-intensive applications and the importance of databases that can handle ultra-high performance and low latency. The summit also featured use cases from prominent companies like Rakuten and Palo Alto Networks, demonstrating how they leveraged ScyllaDB for tasks such as migrating data processing pipelines and correlating records in real time. Additionally, the event explored recent advancements in distributed databases, event streaming, and the broader data ecosystem, while looking ahead to the future of ScyllaDB and its role in the evolving tech landscape.