Vertica is a column-oriented analytical database that uses massively parallel processing to distribute storage and query execution across multiple nodes. The platform supports SQL-based analytics and is designed primarily for reporting, business intelligence, data warehousing, and large-scale analytical workloads.
Column-oriented storage organization, which increases performance of sequential record access at the expense of common transactional operations such as single record retrieval, updates, and deletes.
Massively parallel processing (MPP) architecture to distribute queries on independent nodes and scale performance linearly.
Read the full text
Standard SQL interface with many analytics capabilities built-in, such as time series gap filling/interpolation, event-based windowing and sessionization, pattern matching, event series joins, statistical computation (e.g., regression analysis), and geospatial analysis.
In-database machine learning including categorization, fitting and prediction without down-sampling and data movement. Vertica offers a variety of in-database algorithms, including linear regression, logistic regression, k-means clustering, Naive Bayes classification, random forest decision trees, XGBoost, and support vector machine regression and classification. It also allows deployment of ML models to multiple clusters.
High compression, possible because columns of homogeneous datatype are stored together and because updates to the main store are batched.
Automated workload management, data replication, server recovery, query optimization, and storage optimization.
Native integration with open source big data technologies like Apache Kafka and Apache Spark.
Support for standard programming interfaces, including ODBC, JDBC, ADO.NET, and OLEDB.
High-performance and parallel data transfer to statistical tools and built-in machine learning algorithms.
Vertica's specialized approach aims to significantly increase query performance in data warehouses, while reducing hardware costs.
In 2018, Vertica introduced Eon Mode, an architecture that separates compute and storage resources. Under this model, data is stored in shared object storage while compute resources can be added or removed independently. The design enables multiple workloads to access a common data repository while operating in isolated compute environments.
Vertica includes built-in analytical and machine learning capabilities that can be executed directly within the database. Supported functions include statistical analysis, time-series operations, geospatial analysis, clustering, regression, classification, and model deployment.
Vertica integrates with a range of data engineering, business intelligence, and analytics tools. The platform supports Apache Kafka for streaming ingestion, Apache Spark for distributed data processing, containerized deployments using Docker and Kubernetes, and visualization platforms such as Grafana.
