Vertica Reviews

The speed of Vertica out of the box with the ability it has to perform complex analytics queries. In other databases, information will return in hours or even days while in Vertica it will be finished in minutes or even seconds. This is the best feature it has.

Vertica is in our core technology stack. We are serving reports and dashboard to clients from it. It's very important to us that it fulfills its function correctly and provides us with an advantage over our competitors.

The internal documentation. As a DBA, I really want to understand how its internals works. It needs to handle high concurrency short queries better as Vertica is not handling these well, and we have had to develop our own tool to help us with our dashboard.

I've been using it for five years.

The only thing is that it is very sensitive to network glitches and every time it happens, a node will leave the cluster and we need to re-connect it. Apart from this, the stability is very good.

It is very easy to scale Vertica as you need.

It used to be 10/10 but now it's 8/10. Maybe it's because they were bought by HP which is a big company and the transition is hard.

The last company I worked for shifted from Oracle to Vertica. For our BI queries it's a huge win. Vertica is much better than any other raw store as it is built exactly for complex analytic queries with a huge amount of data.

The earlier version wasn't that good to deploy, but now it's pretty easy to install. If you follow the documentation, you will be OK.

We implemented it in-house. Since it's different than a raw store database, you need to understand the architecture in order to get the most out of Vertica. This means that you will need to design you data model to suit the Vertica architecture otherwise you will get the same performance as the solution you're replacing, or worse. If your implementation is not complex you can just put it in and you will get out of the box improvements, but for complex ones you need to know how Vertica works and build the right design.

It's a great tool but to get the most out of it you will have to design your models to fit it.

The primary use of Vertica is as a data warehouse to perform aggregate and summary reports.

This solution has allowed us to reduce the creation of summarized tables, as the user can perform queries on the fly. It has provided for the faster retrieval of records.

The performance is very good and the aggregate records are fast.

Vertica Auto-Projections reduce the knowledge required by the DBA for tuning performance.

When it is about to reach the maximum storage capacity, it becomes slow.

Before we used Vertica we used another columnar database which turned out to be very unstable and its performance was inconsistent. Vertica turned that around, to the point that it is now our go-to database. We became Vertica partners.

Vertica is a columnar database, this support our developments in analytics, advanced analytics, and our ETL process with large sets of data.

I believe the installation process could be streamlined.

No stability issues.

No scalability issues.

They are very professional and responsive.

See "Improvements to organization," above.

There are some considerations to be evaluated before you start the installation, but the installer does the respective checks so things will function properly. And there are a lot of options.

The first TB is free and you can use all the Vertica features. After 1TB you have to pay for licensing. The product is worth it, but be aware of this condition, and plan. The compression ratio is explained in the documentation.

Infobright and MonetDB.

The technical requirements for the product are really important. The design tool for vertica is the core of the database for performance. Never forget to use it to create projections to optimize the storage compression and response times. Better compression means the 1TB mark takes longer to be reached.

We use the product for compressed data store, fast reporting, and self healing analytical data workloads. It also helps with big data ingestions, processing, and reporting.

Bulk loads, batch loads, and micro-batch loads have made it possible for our organization to process near real-time ingestions and faster analytics.

• The tool for performance tuning and recommendations
• Any novice user can tune vertical queries with minimal training (or no training at all).

• It should provide a GUI interface for data management and tuning.
• Monitoring tools need to be lightweight. They should not take up heavy resources of the main server.

Partition and join back to node are easy and simple for DBAs.

DBAs don’t need to add a partition every month/quarter like with other DBs.

There are a lot of limitations within this product and it makes things extremely hard for developers. It lacks Stored Procedure, packages, and triggers like other RDBMs.

Yes, we have encountered issues with Projections and performance.

Very bad support, I would rate it two out of 10.

We use DB2, Oracle , MySQL, MSSQL. We switched to Vertica to explore it for  future projects.

Easy setup. Much easier than setting up Oracle RAC.

Licensing is based on size of the database.

They did a good PoC and we were impressed with Vertica. However, when we implemented, it was nightmare with bad support.

My advice regarding this product is a definite "no", due to bad support.

Query Performance.

More analytical functions. Optimization around DML operations such that we be able to use it more. I understand that Vertica is not meant for that, but I would really love it if it has capability to support procedural processing.

Optimization around DML refers to fast DELETE and UPDATE statement such that we can leverage Vertica around those operations. I do understand that Vertica is not meant as an OLTP system, neither I'm asking to have it similar but if DML operations can be optimized, that would be admirable.

Regarding, my comment on the capability to support procedural processing - Vertica as of now is mainly used via SQL only. If we have to use any procedure based operation, we do it via User Defined Functions. If within Vertica itself, we have the capability to create & execute procedure similar to that for functions, it would be a plus. Again, I do understand that this may be against the architecture of Vertica but if there is anything that can be revised to get these supported, that would be preferred.

3+ years

Initially yes, but now we're used to it. Challenges do come up but it's all well understood.

We had confronted couple of issues during Vertica upgrades for which we do align with your support group.

Speed of columnar data.

Performance tuning; user community is needed.

1 year

Node recovery is very inconsistent and impacts performance.

Concurrency

Terrible

Terrible

SQL Server and Oracle.

Simple

User Defined Extensions

Analytic Functions

We were able to implement new algorithms without having to move data out of Vertica into a compute cluster. This allowed us to offer Analytics for Cybersecurity to our customers.

More Machine Learning algorithms--Random Forest for sure!

Very responsive

Excellent

In-house