OpenText Analytics Database (Vertica) Reviews

The fact that it is a columnar database is valuable. Columnar storage has its own benefit with a large amount of data. It's superior to most traditional relational DB when dealing with a large amount of data. We believe that Vertica is one of the best players in this realm.

Large-volume queries are executed in a relatively short amount of time, so that we could develop reports that consume data in Vertica.

Speed: It's already doing what it is supposed to do in terms of speed but still, as a user, I hope it gets even faster.

Specific to our company, we do store the data both in AWS S3 and Vertica. For some batch jobs, we decided to create a Spark job rather than Vertica operations for speed and/or scalability concerns. Maybe this is just due to the computation efficiency between SQL operations vs. a programmatic approach. Even with some optimization (adding projections for merge joins and grouped by pipelined), it's still taking a longer time than a Spark job in some cases.

I have personally used it for about 2.5 years.

I have not recently encountered any stability issues; we have good health checks/monitoring around Vertica now.

I have not encountered any scalability issues; I think it's scalable.

N/A; don't have much experience on this.

We do have some pipelines accessing raw data directly and process it as a batch Spark job. Why? I guess it's because the type of operations we do can be done easily in code vs. SQL.

I would recommend using Vertica for those people/teams having large denormalized fact tables that need to be processed efficiently. I worked around optimizing the query performance dealing with projections, merge joins and groupby pipelines. It paid off at the end.

Speed and performance: Vertica stands top among its peers in the MPP world, delivering unparalleled speed and performance in query response time. Its distributed architecture and use of projection (materialized version of data) beats most of its competitors.

This product is used for in-database analytics for reports and queries that require very fast response times. Complicated multi-table queries perform very well, and the company has improved on business operations looking at hot data from various dimensions.

Projections take up a lot of space and hence, compression can be improved. Installation can be more intuitive via a simple, lightweight web client instead of the command line.

I have used it for two years.

While Vertica is otherwise stable, sometimes there are issues with restores to the last checkpoint.

I have not encountered any scalability issues.

Technical support is very good and knowledgeable.

I previously used Postgres; switched as performance suffered due to data growth.

Initial setup was straightforward through the command line.

Negotiate; with HDFS, storage is cheap. Vertica charges per terabyte of compressed data. But the underlying architecture materializes data in a different order and hence space utilization is always heavy, even for a single table; add to that the replication factor.

Before choosing this product, we evaluated Netezza and ParAccel.

Make sure the data and table structures are compact. Vertica will create many versions of the same data as a projection and isolated tables will increase size, increasing licensing cost.

• Compression / speed with highly complex queries

We use it for analytics (marketing).

• Performance tuning
• Not much by way of any documentation: The explain plans are very difficult to read / understand. I tried to diagnose some specific queries using the DBD Vertica utility, etc. For one example of using the explain plan, the query was complex with lots of joins and so on (the query took up about three A4 pages), but the explain plan I printed out took up in excess of 32 A4 pages. How on earth would you read that? No visual tools were available that I could find.
• Very little if any training available in the UK: Our company wasn't able to find any on the topic. We found very little if any documentation (from the vendor) that was of much use.
• Cloning / export was not well documented; poor examples.

I have used it for three years. I worked with versions 4-7.x.

I occasionally encountered stability issues (more so in earlier versions).

I have not encountered any scalability issues.

Technical support is excellent.

Initially when I first started, the documentation, etc. available was scarce. However, this has improved substantially.

I was used to OLTP and DWH solutions based on technology such as Oracle, so some of the concepts are quite different.

Before choosing this product, other options were considered, e.g., Kognitio.

It’s still not mainstream (especially in the UK) and I would say to some extent still ‘improving’ at each release, but it is enterprise ready and a hugely cheaper option than some.

We did some like-for-like comparisons between HP Vertica and Oracle Exadata (work load / timings) and the two compared favourably, with Vertica being faster than Oracle in all but the biggest and most complex of queries.

The biggest, most valuable feature for us is the clustering aspect with a share-nothing mentality. Most clusters usually require their own shared storage, shared subnet, etc. and this becomes a pain and a nightmare to maintain.

The second most valuable feature is that it's very easy to maintain. It's a breeze once you know how to handle it with your scenario in mind.

Loading raw data and leveraging column store to quickly aggregate the values as well as run a general analysis were the biggest improvements we found. Before, we had to scrub the data or reformat, load it, possibly scrub it some more, and then run the first set of analysis, and so on.

With Vertica, we were able to combine some of these steps, such as loading gzip data directly into the table and leveraging R in Vertica to run all of the analysis.

Developer Tools - Vertica really needs some kind of IDE plugin for a system such as Eclipse or IntelliJ. Developing external functions in Vertica can kind of be like shooting in the dark sometimes. Also, an improved monitor or monitoring with alerting built-in that actually works would be a welcome addition.

They truly need a Python or some script that can handle all of the low-level system changes for you and find out how the customer has heavily modified their nodes before the install. Some automation here would help a lot.

The product overall is a great product, however management tools as well as monitoring tools are lacking. The product does, however, offer a lot of information in the form of system views and tables, but most of the data is hard to translate with out the help of their support team.

I have used HP Vertica in multiple companies over the last four years. We currently have it running on a three-node Centos cluster and a six-node Centos cluster.

There have been no issues with the deployment.

There have been no issues with the stability.

We have had no issues scaling it for our needs.

Like everything else HP has support for, the support is very poor. You normally have to threaten to leave, not buy support renewals, or call your sales rep to talk
to anyone who knows anything about the product. The community normally knows more than support and most of my questions or issues were resolved by searching the old community boards while I wait for overseas support to ask me to send them the logs again for the 50th time.

I have previously tried SQL PDW, Mongo, Cassandra for alternatives. Even though all of those products are in different landscapes, the Vertica column store ended up being the best thing that was needed.

It is straightforward if you read the documents and have mid to senior-level knowledge of Linux. Non-Linux admins will find the setup complex and cumbersome since most are Windows admin and they want point-and-click.

We implemented through our in-house team. You need to read the docs, then read them again, and then make yourself a cheat sheet. Once you have done the setup for a two-node cluster, do some Research and Development before taking the time to do a large production cluster or buy the license.

ROI is great compared to the previous solution, SQL Server.

TCO is much lower given the Linux OS and the fact that Vertica is licensed by data size and not node count. The best advice for licensing is to make sure you have a proper data retention policy in place and well-documented as well as some growth expectations before buying. Following this, it will make sure you don't over or under buy.

If you are not Linux savvy, find a person that is. Make a cheat sheet with the commands and/or steps for your environment. If you are in the cloud, make sure to understand the networking aspect is completely different in AWS from it will be in your local data center. Failure to plan is planning to fail with Vertica implementation, and try not to mess up the spread as it's a pain to fix. If you read the documents, you will see what I am talking about.

Analytical features are amazing, the integration is wonderful.

It has helped us escalate, we need information almost real-time.

Documentation, there are functions that are not documented. UDF SDK, I'd like to see a step by step simulator example in a manual. The read-me code is good, however, an example would be great for starters.

3 years

Yes, cluster migration takes time.

Yes, in data streaming ROS containers is a pain to work with.

Not yet.

It is good, they answer in good time. There are times that they really don't come with a proper answer.

Decent.

Yes, regular relational databases. We switched for scalability reasons.

In-house.

Yes, Netezza.

I like the new things they are introducing. I want to see more with Python.

Fast inserts, queries way faster than in SQL Server.

Certain research which was unattainable beforehand, now is in reach.

Some GUI Tools out of the box, better python integration. I would love to see some nice query engine, tooled specifically to Vertica extensions of SQL (with IntelliSense). We currently use TOAD, but it has a lot of bugs and does not provide full support of all Vertica features. For example with the "copy to" command it is buggy - extremely hard to debug.

Other issues would be better query plan display and better management tools like command line admin tools. (MC would probably solve a bit here, I saw a demo on the conference). We have several avid Python users in the company, but this how maybe lower priority after I reviewed my conference materials and found that Vertica now has native Python driver.

2 years.

I was acting as my own DBA for a while, so a lot of hurdles before. But things are getting easier as more people in the company bought into the solution. I also got HP Training in house. (Thanks Herb!)

We can scale a lot, wish it was a bit more affordable.

It complements our SQL Server solution.

Own efforts.

Switch to per node from per TB.

Scale-out, analytical functions, ML.

We are an HP partner. A SQL-based compute platform like Vertica enables far less human overhead in operations and analytics.

More ML, both data prep, models, evaluation and workflow.  Improved support for deep analytics/ predictive modelling with machine learning algorithms. This area of analytics need a stack of functionality in order to support the scenario. The needed functionality includes:

• Data preparation. Scaling, centering, removing skewness, gap filling, pivoting, feature selection and feature generation
• Algorithms/models. Non-linear models in general. More specifically, penalized models, tree/rule-based models (incl. ensambles), SVM, MARS, Neural networks, K-nearest neighbours, Naïve bayes, etc.
• Support the concept of a “data processing pipeline” with data prep. + model. One would typically use “a pipeline” as the overall logical unit used to produce predictions/scoring.
• Automatic model evaluation/tuning. With algorithms requiring tuning, support for automated testing of different settings/tuning parameters is very useful. Should include (k fold) cross validation and bootstrap for model evaluation
• Some sort of hooks to use external models in a pipeline i.e. data prep in Vertica + model from Spark/R. 
• Parity functionality for the Java SDK compared to C++. Today the C++ SDK is the most feature rich. The request is to bring (and keep) the Java SDK up to feature parity with C++.
• Streaming data and notifications/alerts. Streaming data is starting to get well supported with the Kafka integration. Now we just need a hook to issue notifications on streaming data. That is, running some sort of evaluation on incoming records (as they arrive to the Vertica tables) and possibly raising a notification.

Two years.

No, not really.

No.

No.

Postgresql, MySQL, SQL Server. Switched because of scalability and reliability, analytics functionality. V being a better engineered product.

Straightforward. Good docs helped a lot.

Its reasonably priced for non-trivial data problems.

Yes, Hadoop / Spark, SQL Server.

See additional functionality above.

Super-fast aggregated results from massive data.

We can process vast amounts of data, fast and with a high degree of reliability.

Better feedback from installation. I would like to see more meaningful errors returned and more graceful handling of those. Thankfully, we don't often hit error conditions.

4 years.

No

No

Depends on the environment. Generally pretty good. If you have a large catalog, you can get timeouts adding nodes. Large catalog issues have been dealt with it recent releases so this should make scaling up even more robust.

Excellent. It can take some time to get to the right people but generally our issues are all addressed in an acceptable timeframe.

Greenplum. It was less stable.

Vertica is very robust and recovers predictably from unexpected infrastructure failures.

Great overall solution.