February 2026 Summaries

Modernizing COBOL and other legacy systems involves more than just rewriting code; it requires addressing the orchestration and governance of the entire system to ensure reliability and controlled evolution. The challenges often lie in the fragile dependencies, manual processes, and hidden knowledge that these systems rely on, creating resistance to change. Tools like Kestra provide a solution by enabling orchestration, which makes systems observable, repeatable, and governable, allowing for careful modernization without disrupting existing operations. While AI can assist in tasks like code generation and documentation, it is not a replacement for operational discipline. Instead, AI should be integrated into orchestrated workflows to enhance productivity while maintaining control and reliability. Kestra enables the coexistence of legacy and modern workloads, supporting gradual modernization through controlled, repeatable steps without forcing risky, large-scale replacements.

This crash course provides a comprehensive guide on setting up and managing Apache Iceberg on AWS, focusing on creating, querying, and modifying Iceberg tables using Amazon Athena, S3, and AWS Glue. Apache Iceberg is highlighted as an open table format that acts as a metadata layer, enabling reliable transactions, schema evolution, and efficient data management at a petabyte scale. The tutorial walks through creating an Iceberg table, inserting and modifying data, and optimizing data storage to address common challenges like the "Small Files Problem" using SQL statements like OPTIMIZE and VACUUM. It also covers data ingestion methods, including row-by-row inserts and bulk ingestion, using Python scripts and AWS services. Additionally, the course explores scheduling and event-driven data pipelines with Kestra, enabling automation and orchestration of data workflows, while emphasizing the separation of business logic from orchestration. The tutorial concludes with insights on integrating Iceberg with AWS services for scalable data lake management and offers resources for further exploration and community engagement.

MotherDuck is the managed cloud version of DuckDB, offering enhanced features like serverless infrastructure, managed storage, data sharing, and an interactive SQL IDE, making it a collaborative analytics platform. DuckDB, known as "SQLite for Analytics," is an open-source, embedded OLAP database that allows for fast querying of large datasets directly from a user's environment without additional setup. While DuckDB is excellent for local data processing, it lacks high concurrency and horizontal scaling, which MotherDuck addresses by facilitating hybrid execution between cloud and local data. MotherDuck also simplifies data sharing, allowing for easy database collaboration with features such as secrets management and a built-in SQL IDE. DuckDB Labs, a research and consulting firm, supports the open-source development of DuckDB and collaborates with MotherDuck to integrate their services. Users are advised to start with DuckDB for local data processing and transition to MotherDuck when advanced features like hybrid execution and data sharing become necessary. The integration of DuckDB and MotherDuck into ETL pipelines, through tools like Kestra, enables complex data transformations and reporting, demonstrating the versatility and scalability of these platforms.