Our prime use is for warehousing and analytics.
Our prime use is for warehousing and analytics.
Being an appliance, the best features are its analytics and the performance level.
The underlying hardware that IBM provides with this appliance is made for a specific purpose, to serve performance on a large amount of data, and to do analytics as well. It is faster, when you compare it to any other product, like Hadoop or DXC Hadoop or Presto on AWS. It's made for a specific purpose and it serves that very appropriately.
It really serves its purpose. It meets the need for performance, it meets the need for robustness, and it also serves as a perfect data warehousing appliance. The only issue is that it's not expandable.
The new versions of Netezza, they are expandable. They can be on-prem and on the cloud like dashDB from IBM. It replaces the previous Netezza versions. We are currently on Mako because TwinFin support is going in 2019. Maybe in the next couple of years, we will be replacing Netezza with Redshift.
As a database administrator, and in warehousing part for the last 12 years, we always have stability issues in every technology. So the issue is how well they support their product. IBM does support their product very well. We have had issues where there were queries that were generating billions of rows. It's all about how you understand the underlying architecture. The coder who understands the architecture can write a better query. We sometimes need to educate them. So there are always stability issues but not in terms of hardware or support. It's always how well you write your queries.
If you don't know Netezza, if you don't SQL or you don't know the underlying architecture but you are a good SQL developer, you know all the business logic and you write a query - but it is not actually performing - it's because you have not understood the distribution part of it. You need to consider the distribution keys or the organization keys. That makes the difference.
The stability doesn't depend on the hardware, it depends on your coding in SQL.
If I was using TwinFin appliances, if my data grew to the limit that TwinFin supports, obviously I would have to buy a new appliance. I cannot just ask them to increase the memory or increase the storage or the CPUs. That would cost me another appliance. If our business side agreed to that, I would keep using Netezza and buy a bigger appliance, I'm okay with that. But the expandability is not there.
I would rate the technical support at seven out of 10. We have our own SLAs and they have their own SLAs and things go around that. If we want to a fix in two days, it might be that they are not able to provide it, until it becomes business critical.
Netezza was the first one which we incorporated and started using heavily. Before that, we did test the other data warehousing technology but it was just a PoC and PoV test.
In some of our business areas, they still use SQL Server data warehouses and Oracle data warehouses. But once we moved to Netezza, different business units bought in, and now we are up to Hadoop solutions and AWS solutions.
Once it has been shipped and installed in your data center, it's just the point of plugging it in and the initial configuration. IBM people come and when they install it, they initialize it and then they give us the password and from there we move on in. It's very simple from there.
Netezza is a costly solution. It does serve a specific purpose but it's costlier than what's available in the market, if you go to the cloud.
We are also doing a proof of value and proof of concept for AWS and Hadoop DXC.
My client is looking towards replacing Neteeza with one of the up and coming warehousing solutions like Presto. They don't want the in-house or on-prem cost of managing that particular appliance. When everything is available on cloud, we pay less. My client develops medical products. They wanted to concentrate on the medical part, not on how to manage their IT. So they're moving towards more and more towards the cloud to replace the on-prem solution.
My advice would be to fully categorize your needs. Why you need Netezza should be a specific question, because there are so many different analytic solutions and which provide performance and which are cheaper than Netezza. Until you figure out completely that you only need a PDA (pure data analytics) system, you should really look at other products and compare them.
I wouldn't choose Netezza in today's world when we have Redshift, Presto, EMR, when we have Teradata, and when we have Oracle Autonomous. In today's world, you should look at these solutions first. If they don't serve your purpose, then look to Netezza.
In the current world, data is the big question. Nowadays, we are receiving a lot of data. It's like the data generation has come. We have terabytes of data and it might be, in a year or so, you cross the petabyte scale. So go with a petabyte-scale solution instead of a non-expandable Netezza appliance.
We are currently working on the latest Mako version. After that - Mako retires in 2024 - I don't think they have anything on Netezza. What they have is dashDB and Sailfish, which is a completely different product for IBM, but similar to Netezza. And those are expandable.
Netezza is a good product in and of itself, aside from the fact it is not expandable. Overall, it's a good product but definitely has room for improvement.