What is our primary use case?
My solution has a big database with terabytes of data and we use Database In-Memory for a lot of our data. Normally, we partition it and create big tables, but we can use In-Memory for data that we use every day or every hour. We put some partitions in In-Memory from some tables and we use that. It normally has good performance.
What is most valuable?
Normally, every database server uses hard disks. In-Memory has a feature, apart from its database, which is very good. When we start our server, all your data needs loading memory. We can use that. It's a very good feature. I think they added this feature in 2019. We can mount memory in the partition, create partitions in there, and create tablespace from that spot to share. It's a really good feature. We use it a lot.
What needs improvement?
We use some partitions in In-Memory. We have a very large table and a low dose. It is very expensive in data to load all of them into In-Memory. It takes up more memory slots in the server, as well as a lot of RAM. We use the last partitions on the table. We always need to create a script and make a schedule that can load the last partition In-Memory. Oracle doesn't have features to do this automatically. I would like them to allow us to load last partitions, as well as other table partitions, in In-Memory. I think a good feature would do that automatically, letting you see a table, load a large partition, and monitor loading memory. It's quite a good feature.
For how long have I used the solution?
I have been using this version for about four months.
What do I think about the stability of the solution?
We have used this feature for four months and it's stable. I used to load one terabyte of data in my memory servers each time. I think it worked okay. I don't have any problem for now.
How are customer service and technical support?
I live in Iran. Iran is sanctioned by the US government so Oracle can't provide any services to us. I study on my own. I use Oracle documentation and watch YouTube videos. I don't have any company to support us.
How was the initial setup?
It is very easy to deploy. I always work in Linux, Oracle Linux. If you want to work from the GNOME GUI, it's really simple. You must state a parameter in Linux and then use their installer. Click next, next, next, then finish. It's simple. In Oracle 80 and 90, we have an installation system in a common UI. It's very simple. You must have the prerequisites and after that, install RPM in just one comment. Check your configuration and set up the database. It's simple.
Deployment in the GUI version, if you want to use that, is I think 14 to 20 minutes. In the command line, it would take 10 to 15 minutes. I can't remember exactly, but something close to that.
What other advice do I have?
Oracle is the best database, but I love open-source software. Oracle always has the first original features for three or four years and we use them because they are stable and we can buy in a large scale and use it for our office. It has no problems. I think Oracle is ten out of ten.
About Oracle Database In-Memory, in particular, I would rate it as eight out of ten. It's a new feature. I think it's improved from the last version three years ago.
Oracle's new features and data are very useful for us for storing data, loading it, etc. Oracle features based on processes are good. In Oracle, we just have four functions based on data types, but in post-production, we have more than ten functions. That is very useful for us. We'll add more functions and features like index and categorization based on data type, output, and large data. That would be very useful.
Which deployment model are you using for this solution?