Cloud Databases
Cloud databases are managed database services provided by cloud platforms that handle provisioning, patching, backups, scaling, and high availability. They let teams focus on data modeling and queries while the provider operates the underlying infrastructure.
itCloud computing | OpenSkills.info
Intro
Cloud Databases
A cloud database stores and serves application data through cloud infrastructure. You provision it through an API or console. You reach it over a network.
The main benefit is not a new data model. It is a different operating model. A managed service can handle hardware replacement, software maintenance, backups, monitoring hooks, and failover mechanisms. You still own the data model, queries, access rules, recovery objectives, and application behavior.
That division matters. A managed database reduces infrastructure work. It does not make database design, security, performance, or recovery automatic.
The three layers of the decision
Start with three separate questions.
- What data model fits the workload? Choose relational, document, key-value, graph, time-series, or another model from the required access patterns.
- How much control do you need? Choose a self-managed engine on cloud compute, a managed engine, or a more abstract cloud-native service.
- What operating outcome must you meet? Define latency, throughput, consistency, availability, recovery, security, residency, and cost requirements.
Do not begin with a provider logo. Begin with the reads, writes, transactions, and failure outcomes your application needs.
Common data models
Relational
A relational database organizes data into tables with a defined schema. It supports constraints, joins, and transactions across related records. It fits order processing, billing, inventory, and other work that depends on relational integrity.
Document
Continue the course
This section is part of the paid course.
See pricing to subscribe, or log in if you already have access.
Where this skill leads
Relevant careers
See how this topic contributes to broader role-level skill maps.
Sources
- https://csrc.nist.gov/pubs/sp/800/145/final
Supports
- Cloud computing provides on-demand network access to pooled configurable resources
- Cloud services can be provisioned and released with limited provider interaction
- Cloud computing includes defined service models and measured-service characteristics
- https://learn.microsoft.com/en-us/azure/architecture/data-guide/technology-choices/understand-data-store-models
Supports
- Access patterns should guide selection of a data model before selection of a service
- Relational databases support defined schemas, integrity, joins, and multi-row transactions
- Document, key-value, graph, and time-series stores fit distinct data shapes and access patterns
- A production system can use several storage models when access patterns or life cycles diverge
- Each additional model adds operational complexity that must be justified
- https://docs.cloud.google.com/sql/docs/introduction
Supports
- Cloud SQL is a managed relational service for MySQL, PostgreSQL, and SQL Server
- A managed relational service can operate backups, high availability, failover, connectivity, maintenance, monitoring, and logging
- Customers still configure databases, users, resources, flags, high availability, and connectivity
- Managed-service updates can include hardware, online, and maintenance updates with different interruption behavior
- Cloud SQL cost depends on provisioned storage, compute, memory, location, network egress, and address use
- https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/Welcome.html
Supports
- RDS manages backups, patching, failure detection, and recovery for managed relational instances
- Instance classes and storage types have different performance and price characteristics
- VPC placement and security groups control database network access
- Multi-AZ deployments provide data redundancy and failover with deployment-specific read behavior
- Cross-Region replication can be asynchronous
- Database monitoring includes health, metrics, load, logs, and events
- https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/UsingWithRDS.html
Supports
- Security for a managed database follows a shared-responsibility model
- Customer responsibilities depend on the service, data sensitivity, organizational requirements, and law
- Database protection includes VPC placement, identity policies, security groups, transport encryption, storage encryption, and engine permissions
- https://learn.microsoft.com/en-us/azure/cosmos-db/partitioning
Supports
- A partition key determines the logical partition for an item
- Partition-key choice affects distribution and application performance
- A logical partition defines transaction scope for supported Cosmos DB transactions
- Physical partitions distribute storage and throughput and are managed by the service
- Poorly distributed partition keys can create scaling and query problems
- https://learn.microsoft.com/en-us/azure/cosmos-db/how-to-manage-consistency
Supports
- A distributed database can expose configurable consistency at account, client, or request scope
- Session tokens support read-your-write behavior and are associated with partitions
- Strong account-level consistency uses synchronous replication across configured regions
- Applications must account for consistency configuration in client behavior
- https://docs.cloud.google.com/spanner/docs/true-time-external-consistency
Supports
- Spanner assigns transaction timestamps using TrueTime
- Multi-version concurrency control supports consistent reads without blocking writes
- External consistency preserves a serial transaction order that matches observable commit order
- Cross-region database design requires verification of documented transaction consistency semantics
- https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_WorkingWithAutomatedBackups.html
Supports
- Automated backups have a configured retention period
- Point-in-time recovery restores to an eligible point inside the retention period
- Manual snapshots have a lifecycle independent from automated backups
- Backup retention and deletion behavior differ from the behavior of live replicas
- Recovery claims require a usable retained backup and a tested restore path
- https://docs.cloud.google.com/architecture/dr-scenarios-planning-guide
Supports
- Recovery time objective limits acceptable application outage duration
- Recovery point objective limits the acceptable period of data loss after an incident
- Smaller recovery objectives generally increase cost and administrative complexity
- High availability and disaster recovery overlap but are not identical
- Recovery planning should include well-tested procedures
- https://docs.cloud.google.com/architecture/database-migration-concepts-principles-part-1
Supports
- Database migration moves data from source databases to target databases and switches clients to the target
- Migration can restructure or transform data
- Migration approaches include migration systems, engine replication, and custom functionality
- Migration systems extract, transport, and optionally transform data
- Migration planning must account for sources, targets, ongoing changes, and client cutover
- https://docs.aws.amazon.com/prescriptive-guidance/latest/best-practices-migration-cutover/cutover-stage.html
Supports
- Cutover preparation includes functional and non-functional testing
- Operational readiness includes monitoring, alerting, support, and test completeness
- Rollback planning should define triggers, data handling, and decision ownership
- Writes accepted after cutover complicate a return to a stale source
- Restore timing should be tested rather than inferred from configuration