Fast query processing for historical data analytics. Write Optimized Store (WOS) continuous data loading without drastically impacting performance of OLAP queries. It's one of the few columnar databases that has the capability to provide near real time data delivery for analytics with minimal delay sourcing data from traditional databases or NoSQL data stores or any unstructured data sources.
Improvements to My Organization:
With traditional RDBMS historical data analysis or any complex queries took minutes to complete. With the addition of Vertica to handle big data queries, these reports are now returned in under 15 seconds. The biggest performance improvements obviously are for queries that have to analyze a large amount of historical data.
Room for Improvement:
Stability, scalability (3 node Community Edition) and backup/restore all need to be worked on. Without proper work load management and resource pool allocation, any batch/ETL or streaming jobs which refreshes data frequently will impair OLAP query performance.
Use of Solution:
We've been using the three node cluster for about one and a half years.
We had several incidents where SQL queries with UDF predicates would shutdown the cluster or sometimes a single node. We worked with HP support to get these things fixed with subsequent versions of Vertica.
With the Community Edition we are restricted to three nodes. We have a lot of enterprise clients who stress our cluster to its limits. The only advice I would give to new adopters is that if you want superior performance and reliability you are better off going all-in with the enterprise edition and a large number of nodes; assuming you have a lot of clients who run queries concurrently.
Setup and administration are very easy. Vertica was designed to be operational with minimal Database Administrator effort.
Other Solutions Considered:
We evaluated various other solutions but we chose Vertica because its SQL implementation is very similar to PostgreSQL, and therefore it saved us lot of development time re-writing SQL queries. Vertica seems to be one of the few columnar database which can handle both ETL/Batch jobs and OLAP queries simultaneously. We stream data into Vertica from RDBMS frequently than what is typically recommended for Columnar databases.
Which version of this solution are you currently using?