Oracle Multicloud Architect Associate flashcards
Oracle Multicloud Architect Associate · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- Oracle Database@Azure
- Oracle-managed Exadata and Oracle Database services running in Azure regions, administered through OCI tooling while billed and surfaced through Azure.
- Multicloud split responsibility
- Oracle patches Exadata and database software, the hyperscaler runs the datacenter and host network, and the customer owns data, schemas, and IAM policy design.
- Oracle Support Rewards
- Program that offsets some Oracle cloud consumption against on-premises license support costs; eligibility varies per multicloud service, so verify terms.
- Delegated subnet
- An Azure subnet delegated to Oracle so the Exadata VM cluster can place client and backup network interfaces inside the customer VNet.
- Exadata VM cluster
- OCI-administered compute resource hosting Oracle Database instances on multicloud Exadata; cluster members must stay in one region.
- Autonomous Database multicloud
- Fully managed Oracle Database offered inside partner clouds with automated patching, scaling, and backup, trading administrative control for operational simplicity.
- Exadata smart scan
- Offload of row filtering and column projection to Exadata storage cells; behavior is preserved in multicloud placements because the same stack is used.
- Data Guard cross-region standby
- A standby database in a second region receiving redo over the private interconnect, providing region-level disaster recovery for a multicloud primary.
- Private interconnect
- Provider-managed private link carrying traffic between the partner cloud application network and the Oracle database subnet, avoiding public internet egress.
- CIDR non-overlap rule
- Application VNet or VPC ranges, database client and backup subnets, and OCI VCN ranges must not overlap or interconnect routing fails.
- Redo transport latency budget
- Cross-cloud round-trip latency determines whether SYNC or ASYNC Data Guard transport is viable; measure before choosing a protection mode.
- Private DNS zone
- Internal name resolution that maps the database endpoint to its private interconnect address, preventing application traffic from resolving to a public listener.
- Egress cost avoidance
- Database traffic kept on the private interconnect incurs no hyperscaler internet egress charge; public-path database traffic is both costly and a design defect.
- Entra ID to OCI IAM federation
- Mapping Microsoft Entra ID users and groups into OCI IAM so Azure identities receive least-privilege OCI policies for database administration.
- TDE with customer-managed key
- Transparent Data Encryption using a key held in OCI Vault, giving the customer control over rotation and residency of database encryption keys.
- Cloud IAM versus database roles
- Cloud identity governs resource operations such as starting a VM cluster; Oracle Database users and roles govern SQL access, and neither substitutes for the other.
- Segregation of duties
- Separating OCI IAM policy administration, database schema administration, and production change approval so no single principal holds all three.
- GoldenGate versus Data Guard
- Choose GoldenGate for non-Oracle sources, filtered or bidirectional replication, and reversible cutover; choose Data Guard for Oracle-to-Oracle full-database HA and DR.
- Data Pump network_link
- Import mode that pulls data directly from the source database over the private interconnect, avoiding staged dump files in the partner cloud.
- Planned switchover cutover
- Replication target catches up, a switchover runs in a maintenance window, and the application is repointed by DNS with a defined rollback window.
- Endianness precheck
- Verifying source and target platform byte order before migration; a mismatch forces logical migration instead of RMAN physical backup and restore.
- DBMS_CLOUD ingestion
- Package used to load files from object storage into Autonomous Database; the bucket must be reachable over the private path.
- Split observability planes
- Database metrics and Data Guard status live in OCI monitoring, while infrastructure metrics live in the partner cloud, so alerts must be forwarded to one channel.
- Data Guard apply lag
- Difference between primary and standby redo application; alert on it because growing lag erodes the recovery point objective.
- Exadata provisioned-capacity billing
- Exadata infrastructure is billed for the shape provisioned rather than utilization, so right-size at design time because downsizing requires a rebuild.
- Azure MACC eligibility
- Oracle Database@Azure spend can count toward the Microsoft Azure consumption commitment; confirm eligible SKUs and reconcile monthly.
- Patch window ownership
- Oracle applies Exadata and database patches, but the customer must schedule application-compatible windows and validate afterward; the hyperscaler does not patch Oracle Database.
- RAC within one region
- Real Application Clusters provide instance-level high availability but all cluster nodes must reside in the same VM cluster and region.
- Fast-Start Failover observer
- Third-site process that monitors primary and standby and triggers automatic failover; it must be reachable from both to prevent split-brain.
- MaxAvailability protection mode
- Data Guard mode using synchronous redo transport for zero data loss, sensitive to interconnect latency and requiring adequate standby bandwidth.
- FAN and TAF
- Oracle notification and failover features that let connection pools discard stale endpoints and reconnect to the new primary after role transition.
- Standby capacity headroom
- Sizing the standby region to absorb full production load after failover; an undersized standby turns a disaster recovery event into a capacity incident.