December 2022 Summaries

The ScyllaDB Innovation Awards recognize exceptional achievements by users of ScyllaDB, a highly efficient NoSQL database, across various categories celebrating technical prowess, business impact, community contributions, and more. Open to all users of ScyllaDB Cloud, Enterprise, and Open Source, the awards feature seven categories highlighting accomplishments such as groundbreaking data-intensive applications, significant business transformations, innovative technical challenges, and notable security projects. Previous winners include notable organizations like Palo Alto Networks, Instacart, and IBM, each of whom utilized ScyllaDB to enhance performance and innovation in their respective fields. Winners receive recognition through various channels, including a ScyllaDB Summit keynote and social media, and the awards aim to honor those who have pushed the boundaries of what is possible with ScyllaDB. Nominations for the 2023 awards close on January 13, with winners announced at the ScyllaDB Summit in February.

In 2022, ScyllaDB's blog highlighted a range of topics focusing on advancements and best practices in NoSQL databases, with particular emphasis on performance improvements and architectural innovations. Among the most read articles were those detailing how companies like Palo Alto Networks and Disney+ Hotstar replaced their existing systems with ScyllaDB for enhanced stream processing and data architecture scaling, respectively. Other popular posts explored technical enhancements such as implementing a new IO scheduler for mixed workloads and porting database drivers to Async Rust for superior performance. Benchmarking comparisons between ScyllaDB and Apache Cassandra, as well as discussions on using WebAssembly for user-defined functions, also captured readers' interest. The blog post encourages community engagement for future topics and invites readers to the virtual ScyllaDB Summit 2023 to explore further advancements in distributed databases.

In an effort to reduce database costs and improve performance, several companies have successfully transitioned from traditional database systems to ScyllaDB, achieving significant savings and enhancements. Comcast replaced 962 Cassandra nodes with 78 ScyllaDB nodes, eliminating the need for 60 cache servers and cutting annual infrastructure costs by $2.5 million. Rakuten experienced a 2.5x reduction in infrastructure costs by reducing their node count from 24 to 6 and improving data processing efficiency with ScyllaDB's Incremental Compaction Strategy. Expedia achieved a 35% cost savings by replacing Redis and Cassandra with ScyllaDB, which provided consistent low latency and higher throughput. iFood moved from DynamoDB to ScyllaDB Cloud, reducing their database expenses from $54,000 to $6,000 while meeting their throughput requirements. These case studies highlight how employing a more efficient database solution like ScyllaDB can lead to substantial cost reductions and performance gains, particularly by leveraging its architecture to maximize hardware utility and reduce administrative overhead.

The High Performance NoSQL Masterclass, hosted by ScyllaDB and Pythian, focuses on the decision-making factors for selecting NoSQL databases, particularly advocating for wide column NoSQL databases like Apache Cassandra and ScyllaDB. The sessions explore various NoSQL models, emphasizing the advantages of wide column databases for scalability and high performance applications. Allan Mason's segment guides attendees on effectively modeling data and queries using Cassandra Query Language (CQL) with a query-first design approach, contrasting it with traditional schema-first SQL design. Felipe Cardeneti Mendes discusses scaling strategies, workload prioritization, and deployment considerations to enhance performance for different use cases. The Masterclass format combines expert video presentations, interactive panel discussions, and a test to ensure engagement and comprehension, with the entire event available on demand for those who could not attend live.

ScyllaDB Open Source 5.1 introduces several enhancements and features aimed at improving the performance, stability, and capabilities of the NoSQL database. This minor release builds on the major 5.0 release, resolving over 100 issues and adding notable features like the distributed SELECT COUNT (*) with a new "super coordinator" role to optimize query speed across large clusters. It also introduces per-partition rate limits to manage read and write operations, helping to prevent hot partition issues. The Load and Stream feature facilitates easy data restoration without concern for original cluster topology, while the elimination of exceptions from the read and write paths enhances performance by reducing the overhead of timeout handling. The release also includes updates to the Raft consensus algorithm for improved schema management, a CQL extension to prune ghost rows from materialized views, and a new synchronous mode for materialized views to ensure consistency. Additionally, this version supports experimental features like Alternator TTL for data expiration in its DynamoDB-compatible API and WebAssembly-based user-defined functions and aggregates, providing a foundation for future enhancements.