Vertica Deep Dive Comes to Boston: #OwnTheNew

This blog post was authored by Sarah Lemaire.

Two weeks after the successful Deep Dive in Los Angeles, Boston-area Vertica users had the unique opportunity to meet with Sumeet Keswani and Shrirang Kamat from the Vertica Customer Experience Team. These two Vertica experts offered attendees tips for managing their Vertica environment for maximum performance. These attendees ranged from brand-new Vertica users to users with many years of experience. Many of them came with their co-workers.

After a delicious breakfast (the chia seed pudding was fabulous!), Sumeet started off with an in-depth discussion of the new features we’ve added since 8.0 was released. He described how these improvements—all available in Vertica 8.1 or 9.0—can improve the efficiency of your Vertica database.

• Kafka integration improvements
• Wide columns for more efficient buffering
• Flattened tables, with normalized and denormalized columns
• Hierarchical partition management
• UUID data type: saves space and query processing time
• Backup and restore to AWS S3
• Writing ROS data to Parquet in HDFS
• Advanced analytics: faster, scalable machine learning algorithms
• Management Console in AWS, and other new MC features
• Eon Mode Beta: separating storage and computing



Sumeet then dove right into how to improve query performance, covering the following topics:

 Cluster health: How to gather information about the health of your Vertica cluster
 Query plans: How to read and understand the plan that the Vertica Optimizer creates for executing each query
 How to query system tables: Use Vertica system tables to identify the cause of performance issues
 Projection design: How segmentation, sort order, and encoding of projection can optimize query performance

Sumeet used several real-world examples to demonstrate the impact of these improvements.



After the lunch break, Shrirang took over the presentation duties. His job was to provide tips for maintaining the health of your Vertica database. For many of the attendees, these topics were brand new.

The DBADMIN topics that Shrirang discussed were:

• Vertica catalog size
• Tuple Mover operations
• Node recovery
• Rebalancing
• Resource pool configuration
• Database performance over time

All the users I spoke to felt that they gained a lot of practical knowledge from the Deep Dive. They are looking forward to returning to their jobs and implementing what they learned.