Vertica 11 delivers on the vision of unified analytics. This major release of the Vertica Unified Analytics Platform delivers GA support for Docker containers and Kubernetes, advanced machine learning and time series capabilities, and increased analytical performance.
Let’s dive into some details.
Broadest Deployment Support on the Market
Vertica in Eon Mode has expanded from AWS and Google Cloud, and is now in full production on Microsoft Azure cloud. Vertica also welcomes Dell EMC ECS as a certified object storage partner for Eon Mode deployments for on-premises private data centers. Vertica now supports the Eon Mode database in a Kubernetes StatefulSet fully tested and generally available on Dockerhub, further delivering on the commitment to deployment flexibility across multi-clouds and on-premises data centers.
Vertica 11 introduces stored procedures for the purpose of automating the information lifecycle from ELT, through data preparation, to the ML pipeline. Vertica has a rich set of functions at each step in the lifecycle and usage pipeline. Stored procedures will enable users to automate their execution and facilitate metadata collection for auditing and forensics.
Time series data analytics functionality expansion
Time series forecasting functionality expanded greatly in this version to include support for autoregression, moving average, stationarity tests, and automatic generation of correlation plots in-database using SQL and VerticaPy. A lot of Vertica customers support IoT use cases including Jaguar Racing.
“Running in-database analytics and machine learning will be a real game changer for Jaguar Racing,” says Phil Charles, Technical Manager, Jaguar Racing. “With access to all the data, 700+ in-database functions, and the blazingly fast query results, Vertica helps us to make time-critical decisions to ultimately deliver an edge that leads to more points, podiums, and wins for Jaguar Racing.”
Improved external model support
Vertica now supports the latest TensorFlow version 2.5, and generalized linear model import from PMML.
Introducing VerticaPy Delphi – Auto-ML in VerticaPy
VerticaPy is an open-source Python library that exposes Pandas and Scikit-like functionality to conduct data science projects on data stored in Vertica and using the full power of the parallel Vertica engine for execution on complete data sets. VerticaPy combines the scalability of Vertica with the flexibility of Python. The newest version includes VerticaPy Delphi, an auto-ML capability that can greatly shorten machine learning project time to value. Delphi auto-prepares data, trains, and evaluates multiple algorithms at once. In a few minutes, it can show you a graph of several algorithms by accuracy and efficiency.
New improved Apache Spark connector
This new Spark connector not only provides fast parallel read/write support, but since Vertica contributed it to open source, other people like the folks at Agoda are already diving in to make it even better.
- Supports Spark 3.X
- More efficient bi-directional data flow
- Projections, filters, and SQL push-down
- Enterprise SSO support including Kerberos
- Contributed to open source
Full support for complex data types in ORC and Parquet
- Query files with complex types in place without modification or import
Several improvements to the Management Console including
- Quick launch templates to get you working faster
- Simplified workflows to help you work more efficiently
- Customizable alerts and monitoring metrics to keep Vertica running smoothly
Extensive query engine optimizations
- Clear WITH clause speedups
- Multiple performance improvements
And those are just the highlights. Check out the Vertica 11 documentation for details.
And if you haven’t tried Vertica yet,
Try It Now!