April 2023 Summaries

The blog series by Felipe Cardeneti Mendes focuses on avoiding common mistakes when using ScyllaDB, beginning with selecting suitable infrastructure for optimal performance. ScyllaDB, known for high throughput and low latency, is a distributed database that requires specific considerations for infrastructure to achieve its full potential. Users often make mistakes by applying practices from other databases, which may not align with ScyllaDB's architecture. Key recommendations include starting with smaller, under-provisioned environments, ensuring cluster configurations are balanced, and considering networking and memory allocation carefully. The series emphasizes the importance of understanding specific use case needs and making informed infrastructure choices to avoid pitfalls and maximize the database's capabilities. Future posts will delve into storage considerations and data modeling mistakes.

The tutorial outlines a step-by-step process for using Spring Boot with ScyllaDB to manage time series data, specifically for a sample stock application. It demonstrates how to build APIs for creating, retrieving, and deleting stock records with dates and prices, leveraging ScyllaDB for data storage. Throughout the guide, Postman is utilized to test API functionalities, while a ScyllaDB cluster is set up using Docker or ScyllaDB Cloud. The tutorial includes setting up a data model with TimeWindowCompactionStrategy for efficient compaction and data retention. It highlights the resources available through ScyllaDB University, offering further insights and a comprehensive learning path. Developers can gain hands-on experience by interacting with the application and exploring the underlying data management strategies.

Numberly's Omnichannel Delivery team replaced their messaging infrastructure, previously reliant on Kafka, with a Rust-based ScyllaDB shard-aware application to streamline message processing and improve reliability. The new system consolidates message handling into a single platform, allowing for deterministic workload distribution and reduced dependency on multiple technologies, enhancing reliability and idempotence. The change required significant architectural adjustments, including the use of ScyllaDB's shard-per-core architecture to align data processing with CPU cores, effectively turning ScyllaDB into a distributed cache and ordered-data bus. The team chose Rust for its efficient, low-level capabilities, contributing to the ScyllaDB Rust driver to facilitate this transition. The implementation faced challenges, such as ensuring idempotence and optimizing memory usage, but ultimately led to a highly efficient and scalable system that earned the ScyllaDB Innovation Award for Top Technical Achievement.

Since the inaugural ScyllaDB Summit in 2016, both the ScyllaDB database and its community have significantly evolved, with various tech talks highlighting user achievements and migrations to ScyllaDB from other databases like MongoDB, Cassandra, and DynamoDB. Notable discussions included ZeroFlucs' innovative approach for low-latency replication in sports betting, Discord's transition from Cassandra to ScyllaDB for message storage, and Epic Games' use of ScyllaDB as a binary cache to enhance game asset distribution. Other mentions were Numberly's replacement of Kafka and Redis with ScyllaDB for cost-effective operations, ShareChat's scaling with reduced latency, SecurityScorecard's migration to ScyllaDB for improved cybersecurity risk assessment, Strava's performance enhancements, and Level Infinite's solution for event dispatching in games. These talks collectively underscore ScyllaDB's capacity to handle diverse, high-performance applications across various sectors, emphasizing its scalability, cost-effectiveness, and flexibility in meeting specific user needs.

ScyllaDB, known for its high-speed NoSQL database, organized a free virtual event, "ScyllaDB Labs – Introduction to NoSQL," specifically tailored for audiences in Asia-friendly time zones. Held on April 19, the event aimed to provide an immersive learning experience for participants from regions like India, Singapore, Malaysia, and Australia by offering a combination of theoretical insights and practical labs. Participants were introduced to NoSQL concepts and strategies used by major companies such as Disney+ Hotstar and Tencent Games, and had the opportunity to engage in hands-on labs to deepen their understanding of ScyllaDB's architecture, including data modeling, cluster setup, and SQL to NoSQL migration practices. The event was designed to be interactive, with ScyllaDB experts available to assist attendees in real-time as they explored topics like replication factors, consistency levels, and materialized views. Guy Shtub, head of ScyllaDB University, emphasized the experiential nature of this event, highlighting its focus on practical application over traditional lecture formats to ensure a comprehensive learning experience.

TRACTIAN embarked on a significant transition from MongoDB to ScyllaDB to accommodate a tenfold increase in data throughput and enhance their real-time machine learning environment. The decision was driven by ScyllaDB's scalability, high performance, and ability to handle large volumes of data, which were vital for TRACTIAN's needs. A comprehensive comparison of ScyllaDB with MongoDB and PostgreSQL highlighted ScyllaDB's superior performance, especially in terms of latency and throughput, although the migration process posed challenges including data modeling adjustments and infrastructure updates. The transition required meticulous planning and execution, with separate teams handling the old and new architectures to ensure continuous operation without outages. This process underscored the importance of selecting a database based on specific use cases, being prepared for mid-project changes, and focusing on the future potential of the product rather than the convenience of the current engineering setup.