Choosing where a database lives is not only a question of cost or convenience — it is a question of how much downtime an organization can tolerate, how quickly it can recover when something fails, and who is accountable when data is exposed. This topic brings those concerns together under three headings: availability (will the database respond when users need it), reliability (can it survive hardware failure, human error, or a regional outage), and security (who can reach the data, and how is it protected at rest and in transit).
In a cloud deployment, much of this is expressed contractually through a Service Level Agreement, and technically through provider features such as redundant replicas, automated backups, and geo-replication. In a local deployment, the same guarantees have to be engineered and funded by the organization itself: backup schedules, standby servers, off-site copies, failover testing, and physical and network security controls. Neither model removes the work — the cloud shifts part of it to the provider under a shared responsibility model, while on-premises keeps all of it in house.
As you watch the videos, look for three things. First, the vocabulary: RTO (recovery time objective), RPO (recovery point objective), high availability versus disaster recovery, failover, and uptime percentages. Second, the mechanisms: what specifically provides redundancy in each model. Third, the division of labor: note each time a task is performed automatically by a provider versus performed manually by a DBA, because that distinction is the heart of the shared responsibility model and of the recommendations you will be asked to make later in this module.
A short lecture introducing the Service Level Agreement as the formal contract between a cloud provider and a customer. Based on its title and description, it explains what an SLA specifies regarding expected performance and availability, why such an agreement is needed, and how it is used in practice when evaluating or holding a provider accountable.
It gives you the contractual vocabulary behind uptime guarantees in a compact, lecture-style format, which is the natural starting point for the Uptime Guarantees and SLAs subtopic before moving to the technical mechanisms.
An overview episode from Microsoft's Data Exposed series covering high availability and disaster recovery concepts as they apply to Azure SQL Database. It is positioned to explain the difference between HA and DR, what availability means in a managed cloud database, and the options the platform offers for implementing each.
It comes directly from the platform vendor and addresses Disaster Recovery in Cloud Deployments using a concrete managed service, letting you see which availability tasks the provider handles for you and which remain yours to configure.
A Data Exposed MVP Edition session on automating disaster recovery for SQL Server running on-premises. Framed around the question of whether you are truly prepared if your primary data center is destroyed, it addresses the gap between simply having database backups and having a tested, automated recovery process.
It is the on-premises counterpart to the Azure video and supports Disaster Recovery in Local Deployments and Business Continuity Considerations by showing the engineering effort an organization must take on itself when no provider is doing it.
Together these videos supply the availability and security half of the deployment comparison this module builds toward. The SLA lecture frames the guarantees a cloud provider contractually offers, while the paired Azure SQL and on-premises SQL Server sessions let you contrast how the same recovery goals are met in each infrastructure model (MO1) and where the operational burden and cost of achieving them falls (MO2). Watching the cloud and local disaster recovery episodes side by side is the core preparation for weighing maintenance, availability, and security trade-offs when recommending a hosting strategy (MO4), and each time a video shows a task handled automatically by the platform rather than manually by a DBA, you are collecting evidence for classifying duties under the shared responsibility model (MO5).
These videos were chosen from YouTube on October 2, 2026. Each has captions; the player turns them on. Videos on YouTube can be changed or removed — if one no longer plays, tell your instructor. If a player does not appear, use its Watch on YouTube link.