November 2019 Summaries

In November 2019, ScyllaDB organized a two-day retreat in Sonoma following a week of intensive training and their summit, aimed at fostering team bonding and relaxation. The company, which had experienced significant growth, brought together employees from 13 countries, promoting camaraderie through activities like hiking, a bicycle-building challenge for local children, and an Amazing Race-style competition. These events allowed team members to engage in diverse activities, enjoy local cuisine and wine, and explore historical sites in Sonoma. The retreat emphasized teamwork, creativity, and community involvement while providing a platform for employees to form lasting friendships and enhance collaboration. Feedback from participants was overwhelmingly positive, highlighting the success of the event in strengthening the ScyllaDB community and setting a high standard for future gatherings.

Modern distributed databases, particularly those supporting microservices architectures, must balance concurrency and connection management to optimize performance while minimizing timeouts and overloads. Overloading a database server can increase latency and lead to timeouts, which are exacerbated by improper retry policies. Effective concurrency management involves correctly configuring client-side parallelism, connection pools, and timeout settings. Connection pooling reuses established client/server connections to enhance efficiency, while thread pools mitigate the overhead of creating and destroying threads. Properly sizing connection and thread pools based on server capacity is crucial, as excessive connections or in-flight requests can saturate the server, degrading performance. Tools like ScyllaDB demonstrate how connections and server attributes interact, using a shared-nothing model to maximize throughput. The importance of aligning client-side and server-side timeout settings is emphasized, as mismatches can lead to cascading performance issues. The article highlights the necessity of careful concurrency control, offering insights into optimal configurations for scalable and efficient database performance.

ScyllaDB Summit 2019 showcased significant advancements in ScyllaDB's capabilities and its impact on the Big Data industry, with keynotes and sessions highlighting the database's scalability, cost efficiency, and performance improvements over competitors like Cassandra and DynamoDB. The event featured discussions on ScyllaDB's innovative features such as its Seastar engine, Change Data Capture (CDC), Lightweight Transactions (LWT), and User Defined Functions (UDFs), as well as successful case studies from companies like Comcast and Kiwi.com, which demonstrated substantial improvements in read/write speeds and cost savings. Participants also explored the integration of ScyllaDB with modern technologies like Kubernetes and streaming platforms, emphasizing its adaptability for diverse applications across industries. ScyllaDB's focus on simplifying software adoption and enhancing operational efficiency through its cloud services and Alternator API was a prominent theme, as was its commitment to ongoing development and user engagement.

At the 2019 ScyllaDB Summit, various companies were honored with ScyllaDB User Awards for their innovative and efficient use of the ScyllaDB database across different categories. Tubi was recognized for integrating ScyllaDB with Spark to handle large-scale personalization of user homepages, while Grab leveraged ScyllaDB with Kafka to enhance its microservices architecture. FireEye utilized ScyllaDB in combination with JanusGraph for a high-performance graph database, and Dynamic Yield improved GDPR compliance by integrating ScyllaDB with Apache Flink and Spark. Comcast achieved significant infrastructure optimization by replacing numerous Cassandra nodes with ScyllaDB, leading to substantial cost savings. Opera drastically reduced latency in its real-time data synchronization after transitioning to ScyllaDB, and Kiwi.com managed to efficiently balance analytics and operational needs on a single platform. Yannis Zarkadas was acknowledged for his contribution to the ScyllaDB community, particularly with the Kubernetes Operator, and Numberly showcased a dramatic reduction in processing time through a sophisticated integration of ScyllaDB with multiple technologies. The awards celebrated the diverse and impactful ways in which ScyllaDB has been implemented, underscoring its value within the open-source NoSQL database community.

In an interview with SmartDeployAI's Timo Mechler and Charles Adetiloye, the focus is on their ScyllaDB Summit 2019 presentation about simplifying the creation of machine learning workflow pipelines for IoT applications using Kubernetes and ScyllaDB. They emphasize the strategic use of these technologies to democratize artificial intelligence by making it accessible without specialized DevOps skills or large budgets. Their approach treats pipeline components like LEGO blocks, promoting simplicity and scalability. SmartDeployAI evolved from their consulting experiences, recognizing the need for infrastructure that supports AI and ML at scale. Both Mechler and Adetiloye highlight that their goal is to help clients efficiently reach their objectives with minimal resources, and they share personal insights into their interests outside of work.