What is most valuable?
Vertica’s analytic capabilities are its key strength. It can aggregate and analyze data at massive scale and neatly bring the calculation logic to the data with external procedures in C, Java and R.
The Vertica architecture means it can process/ingest data in parallel to reporting and analyzing because of its in-memory Write-Optimized Storage sitting alongside the analytics optimized Read-Optimized Storage.
Which brings us to projections and the DB designer which intelligently structures how data is actually stored on disk to improve the queries you actually run against it. So tables are a logical construct which are operated on as per other DBMS systems, but there’s a whole next level of intelligence in optimization for querying that puts Vertica in another league.
How has it helped my organization?
Our consultancy has introduced Vertica to a number of clients, from small scale ones who benefit from the free tier and per TB pricing model to have a powerful analytics cluster fairly cheaply to large investment banks who have been able to handle data at a scale that wouldn’t otherwise be possible.
What needs improvement?
We’ve built a data ingestion tool to sit alongside Vertica for easy data loading, and I would personally like to see extended developer tooling suited to Vertica – think published PowerDesigner SQL dialect support, IDE with IntelliSense, and stored procedures which we’ve also had to build a work-around module for.
For how long have I used the solution?
Personally, I've used it for three to four years (since v6), but a few others in Thorium Data Science have used it for longer.
What was my experience with deployment of the solution?
We've had no issues. You do need to invest a little time to understand how to set things up and optimize for your workload, but it’s all well documented and there are consultancy firms who will happily help with that.
What do I think about the stability of the solution?
We've had no issues with the stability.
What do I think about the scalability of the solution?
We've had no issues scaling it.
How are customer service and technical support?
It's very good. HP have some technically smart guys and are willing to give access to them when you start using Vertica. We’ve had some great support from their engineering team with things like telling us about upcoming features (snapshotting, in this case), which were spot on for a need a client of ours had. We were looking into engineering a solution ourselves and HP happened to have just what we needed coming down the pipeline in the next version.
Which solution did I use previously and why did I switch?
We previously used Exadata, which is typically very expensive by comparison. This is because Oracle throw top end hardware at the problem as opposed to
HP Vertica’s commodity hardware and smart software approach.
How was the initial setup?
It takes some time to come to grips with the various considerations. I’d suggest bringing in a consultant if you don’t have the time or inclination to do it yourself as it takes going through and install and configuration one or two times to really understand the implications of the different options.
What other advice do I have?
The implementation itself is excellent with fantastic features, speed and scalability. They lose a point only for the development experience which relies on third party tooling like squirrel, and not having SQL based stored procedures.
Go for it! Try the pre-installed VM which HP offers to have a play with it and get a feel for it. It can certainly scale better than any other RDBMS and pushes the envelope of SQL analysis so you can query/analyze/report “BIG-DATA” without having to resort to the complications associated with Hadoop & unstructured data analysis. If your data is structured and large Vertica is what you need.
Which version of this solution are you currently using?