Explanation
On an Exadata DB system, all databases share dedicated storage servers which include flash storage. By default, the databases are given equal priority with respect to these resources. The Exadata storage management software uses a first come, first served approach for query processing. If a database executes a major query that overloads I/O resources, overall system performance can be slowed down.
The I/O Resource Management (IORM) allows you to assign priorities to your databases to ensure critical queries are processed first when workloads exceed their resource allocations. You assign priorities by creating directives that specify the number of shares for each database. The number of shares corresponds to a percentage of resources given to that database when I/O resources are stressed.
Directives work together with an overall optimization objective you set for managing the resources. The following objectives are available:
1) Auto - Recommended. IORM determines the optimization objective and continuously and dynamically determines the optimal settings, based on the workloads observed, and resource plans enabled.
2) Balanced - For critical OLTP and DSS workloads. This setting balances low disk latency and high throughput. This setting limits disk utilization of large I/Os to a lesser extent than low latency to achieve a balance between good latency and good throughput.
3) High throughput - For critical DSS workloads that require high throughput.
4) Low latency - For critical OLTP workloads. This setting provides the lowest possible latency by significantly limiting disk utilization.

A fault domain is a grouping of hardware and infrastructure within an availability domain. Each availability domain contains three fault domains. Fault domains provide anti-affinity: they let you distribute your instances so that the instances are not on the same physical hardware within a single availability domain. A hardware failure or Compute hardware maintenance event that affects one fault domain does not affect instances in other fault domains. In addition, the physical hardware in a fault domain has independent and redundant power supplies, which prevents a failure in the power supply hardware within one fault domain from affecting other fault domains.
To control the placement of your compute instances, bare metal DB system instances, or virtual machine DB system instances, you can optionally specify the fault domain for a new instance or instance pool at launch time. If you don't specify the fault domain, the system selects one for you. Oracle Cloud Infrastructure makes a best-effort anti-affinity placement across different fault domains, while optimizing for available capacity in the availability domain. To change the fault domain for an instance, terminate it and launch a new instance in the preferred fault domain.
Use fault domains to do the following things:
Protect against unexpected hardware failures or power supply failures.
Protect against planned outages because of Compute hardware maintenance.
Reference:
https://blogs.oracle.com/cloud-infrastructure/using-availibility-domains-and-fault-domains-to-improve-application-resiliency