It is a scalable system when it comes to performance, but unfortunately Oracle is no longer supporting the edition that we are running as it is now end-of-life, and figuring out the way forward isn't easy.

Our total user count is around 3,000 end-users, including everyone who uses our organization's applications such as normal end-users and the many administrators who work on it.

It is scalable, but as mentioned, if you want to update your SuperCluster to the next version, you are out of luck as there will be no future development on it from Oracle.

As it is now, both the banking application and a few other applications are running fine for our users on SuperCluster. 

While the general consensus was to maintain about 10 percent free space on ZFS file systems, it became clear that 30 percent free was actually required in order to maintain acceptable performance. Having to keep that headroom available impacted scalability to some extent.

Within the same infrastructure it's scalable but, as I said, its integration with other systems in our datacenter is where there's more of a challenge.

The problem with the scalability at this point is that, while it is field upgradable, in order to do a field upgrade at this point in time, the entire SuperCluster has to be rebuilt. They're working on a solution to have field upgrades available so that it doesn't require being rebuilt. However, at this point in time, that's the only option. Previously, that wasn't even an option. Once you bought the SuperCluster, it was stuck in the configuration that you bought it at for its entirety of its lifetime.

They have improved it, but they still have some work to do.

We faced issues with ZFS failover, due to misconfiguration. After correcting the settings, ZFS failover worked like expected.

This solution is scalable and performs well.

We can expand to our requirements because the storage can be expanded and also the computer with us can be expanded. It depends upon the free space available on the rack.

We can expand the rack to superclass a rack and populate what the requirements are regarding the compute, the storage, and we can connect between the two racks using the InfiniBand switches.

It's partially scalable. We have less than 20 users using this solution.