RoCE, or Remote Direct Memory Access over Converged Ethernet, is a high-speed networking technology that allows data to be transferred between servers, storage, and other devices over Ethernet. Persistent Memory is a form of non-volatile memory that can be used to store data without losing it when power is lost.
Oracle Linux 8.0 supports both RoCE and Persistent Memory, allowing Exadata systems to leverage the benefits of both technologies. For more information on Oracle Linux 8.0 and its features, refer to the Oracle Exadata Database Machine X9M Implementation Essentials official text book and resources.

According to Oracle documentation12, patchmgr is the utility used to update software for Exadata infrastructure components. It can apply patches in a rolling or non-rolling manner depending on the argument and input file provided.

Oracle Platinum Services provides patching features for the full software stack, which includes operating systems, virtualization, storage software and firmware, and databases. This service is typically performed up to twice per year, allowing customers to schedule the patching at their convenience. The patching is usually done to keep the software and firmware up-to-date and to fix any known security vulnerabilities. It's important to note that patching services are not done automatically during business hours, customer's involvement and schedule is needed, and the rolling and complete downtime approaches are options to minimize the impact on the system during patching.
The correct statement about the patching features provided in Platinum Services is that patching services are available for the full software stack up to twice per year. Oracle Platinum Service covers Exadata storage software and firmware patching, and customers must also perform the database patching. The rolling and complete down time approaches are two options for patching, and patching should be scheduled for times when the system is not being heavily used. Patching is not done automatically during business hours. This is covered in section 4.13 of the Oracle Exadata Database Machine X9M Implementation Essentials Official Text Book and Resources [1].
[1] https://docs.oracle.com/en/engineered-systems/exadata-database-machine/x9m/exad-implementing-database-machine-x9m.pdf Search results: [1] Oracle Exadata Storage Server Patching [1][2]. Oracle Exadata Storage Server patching is performed by a team of Oracle engineers [3][2], and includes the latest Oracle... [2] Patching and Upgrades. Oracle Database Machine provides automated patching and upgrades for the full Exadata software stack, up to twice a year, with... [3] Virtualized Exadata. Oracle Database Machine X9M-2 comes with a choice of two patching approaches to accommodate different customer needs: Rolling... [4] Jan 10, 2017 ... Exadata Patching Process. Oracle Exadata Storage Server patching is performed by a team of Oracle engineers [3][2], and includes the latest...

According to the Oracle Exadata Database Machine Technical Architecture1, the initial storage configuration after the standard (non-virtualized) deployment of a new Exadata Database Machine with High Capacity storage servers includes two ASM disk groups: DATA_<DBM_Name> and RECO_<DBM_NAME>. These disk groups are built on hard disks and flash disks inside the Exadata Storage Servers1.
The correct statements about this configuration are:
Option D: There is free space available on flashdisks inside the Exadata storage servers for possible use for storage indexes. Storage indexes are a feature of Exadata Storage Software that can improve query performance by avoiding unnecessary I/O operations. Storage indexes use a small amount of flash memory to store metadata about data blocks stored on disk1.
Option E: There is free space available on flashdisks inside the Exadata storage servers to configure Exadata Smart Flash Logs. Exadata Smart Flash Logs are another feature of Exadata Storage Software that can improve database performance by using flash memory as an extension of the database redo log buffer. This can reduce latency and increase throughput for redo log writes1.