What needs improvement?
My main suggestion for Oracle is the configuration and key values that come for JSON files. When we create a table, especially if you see in our RedShift or some other stuff, if I create a table on top of a JSON file with multiple array columns or superset columns, those column values create some difficulty in Oracle.
In terms of architecture and pricing structure, I feel it is a little bit costly compared to Azure. It's fine compared to RedShift, but compared to Azure, it's a bit pricey when you calculate for one TB storage plus around five hours of reporting with the frequency of 1TB data. The cost adds up, making Oracle a bit expensive.
There are two clusters I'm expecting to improve. The super set datasets need a bit of enhancement, especially the column values that come with the super in these. There's a separate column called super. Those kinds of elements have to be improved in Oracle.
Another aspect is the pricing strategy, which should come with different options. Cloud users prefer keeping their applications in OCI cloud. Many are moving to Postgres for savings. That's a great deal for us.
We recently moved our Oracle database to Postgres, saving a significant amount. Oracle Autonomous Data Warehouse should consider more flexible pricing strategies based on usage to accommodate the diverse needs of users.
In future releases, I would like to see better support for various file systems when creating tables. Plus, a bit more focus from the Oracle community on improving things. And lastly, the pricing strategy could be tweaked to leverage the power of Oracle Autonomous Data Warehouse further.
For how long have I used the solution?
I have been using it for two years.
How are customer service and support?
The community support is top-notch. We can almost always find answers, even for new file formats. That's what initially drew us to Oracle. Knowing a strong community was there to help us overcome any challenges was a major deciding factor. So, I'd rate the support a solid nine out of ten, maybe even nine and a half.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
I've also used AWS Redshift and Azure Synapse Analytics. I worked with Redshift and now Snowflake Query Service. On top of Snowflake, I used Resideo and others. I even played with some open-source options like PNOC.
In the market, the main players are Azure, AWS Redshift, and Oracle Autonomous. But compared to the others, configuration is much easier in Autonomous.
Now, in terms of creating and reporting for our current needs, it's pretty easy.
We're planning to keep working on Oracle Autonomous. We'll apply it to historical accounts only for now. Once this project wraps up, we plan to move all our accounts to Oracle.
How was the initial setup?
The initial setup is fine if compared to RedShift; this is easy.
On AWS or Azure, spinning up a data warehouse involves things like Glue jobs, schedules, file transfers from S3, crawler configurations, and processing shifts. It's a whole orchestration!
With Oracle Autonomous Data Warehouse, things are much simpler. Creating a structure, initializing the servers, extending the servers, those are all things that are very, very easy. That's the main reason we use it.
Postgres is the first reason, and the second one is it's easy. That's one of the main reasons we haven't migrated away yet. It frees us from the complexities of server and configuration management, allowing us to focus on what truly matters: reporting and queries.
What's my experience with pricing, setup cost, and licensing?
Cost-wise, it's a solid seven out of ten. A bit costly, but it is a good tool.
What other advice do I have?
I recommend it, mainly because we don't need to work on configuration compared to others. There is no need for those manual setup steps.
However, users should be aware of potential file strategy differences and some technical nuances versus other data warehouses. Overall, though, it's a smooth transition.
Configuration and administrative side? I would rate it as nine and a half out of ten.
But for development, data fetching, fixing, and applying algorithms, I would rate it a seven out of ten due to compatibility issues, extra coding, and other needs.
Performance tuning, however, shines at eight to eight and a half! It automatically indexes and tunes based on my tables, saving me loads of work.
*Disclosure: My company has a business relationship with this vendor other than being a customer: Partner