Making ScyllaDB a Monstrous Database: Introducing Project Circe

ScyllaDB, a mature NoSQL database, is undergoing a significant transformation through Project Circe, which aims to enhance its consistency, elasticity, and ease of use. This initiative plans to introduce substantial features monthly throughout 2021, including the adoption of the Raft consensus protocol to improve data manipulation and ensure transactional consistency during schema and topology changes. The project addresses the complexities of eventual consistency and aims to make maintenance operations smoother with innovations like Repair Based Node Operations (RBNO). Furthermore, ScyllaDB is focusing on improving elasticity by enabling the addition, removal, and replacement of multiple nodes simultaneously, thereby simplifying cluster management and performance. Additionally, performance enhancements include uninterrupted operation under high loads, improved large partition performance, a new IO scheduler, and the use of coroutines to optimize asynchronous operations. The project also emphasizes stability improvements, such as transactional schema changes and workload shedding algorithms, while expanding deployment options with a Kubernetes Operator and extending ScyllaDB Cloud to platforms beyond AWS.