SQL Database service tiers also differ based on capabilities related to business continuity and in-memory features.
Business continuity services
All service tiers provide mechanisms to enable a business to continue operating in the face of disruption to its computing infrastructure. The types of potential disruption scenarios from which these mechanisms protect you are:
Local hardware and software failures, such as disk failure
Data corruption and deletion, such as application bugs and user errors
Data center outage, such as a natural disaster
Upgrade or maintenance errors, such as unanticipated issues during an application upgrade requiring a rollback to a previous state
The next two sections discuss these business continuity mechanisms and how they differ across service tiers.
Automatic database backup and self-service geo-restore services
To protect your data and enable point-in-time restore services, SQL Database automatically takes full database backups every week, multiple differential backups every day, and log backups every five minutes. Backup files are stored in a geo-redundant storage account with read access (RA-GRS) to ensure backups’ availability for disaster recovery purposes in case of catastrophic disaster at any single data center. The first full backup is scheduled immediately after a database is created. After the first full backup, all further backups are scheduled automatically and managed silently in the background. The exact timing of full and differential backups is determined by the system to balance overall load.
SQL Database provides this service for all databases regardless of server tier, but they have different retention periods for the database backups as follows:
Basic: Any restore point within the past 7 days
Standard: Any restore point within the past 14 days
Premium: Any restore point within the past 35 days
SQL Database provides both database copy and point-in-time restore services on top of this automated backup system, which enable you to restore an existing or deleted database to a new database as of a specified point in time, including the current point in time, within the retention period supported by the service tier. This capability is offered within the current data center.
Storage geo-replication provides the ability to restore a database to another data center from the last geo-redundant backup to create a new database in any geographic region. This capability is called geo-restore. The geo-replication of the backup blobs from the automated backups guarantees that daily backups are available even after an outage or massive failure in the primary region.
In addition to automated backups and storage geo-replication, SQL Database offers active geo-replication for all databases regardless of service tier. Active geo-replication provides a mechanism to create and maintain asynchronous secondary replicas in up to four regions. These secondary replicas can be either offline or online and readable, and they can be in the same or different data centers.
readable secondary is priced at 1X price of the selected performance level. A non-readable secondary is charged at a discounted rate of 3/4X price of the selected performance level. To ensure that transactions being applied to the secondary do not bottleneck the primary, the secondary must be at the same or higher performance level than the primary. Non-readable secondaries are available for failover if needed due to any of the disaster scenarios discussed above. Readable online secondary replicas can be used for load balancing and/or low-latency read access to different geographic regions. When performing an application upgrade or maintenance operation, the continuous replication can be frozen immediately prior to the upgrade or maintenance operation so that you can easily fall back. In case of a data center outage, you can manually fail over to one of the secondary replicas. You also can design and deploy a small worker role application that monitors your primary database and triggers a failover if necessary. For a discussion of active replication topologies using a combination of readable and non-readable secondaries for both disaster recovery scenarios and application load-balancing scenarios.
In-memory services are available only in the Premium service tier, with different size limits for memory storage based on the performance level within the Premium tier. In-memory services encompass two features: in-memory OLTP and Operational Analytics.
In-memory OLTP is a feature that can greatly improve OLTP performance by storing selected OLTP tables in memory (5X–20X improvement) and by using compiled stored procedures (100X improvement). In-memory OLTP has been available in SQL Server beginning with SQL Server 2014 and has been enhanced significantly in SQL Server 2016. These enhancements also are available in SQL Database.
Operational Analytics provides the ability to run both analytics and OLTP workloads in the same database at the same time without a significant degradation in performance. Besides running analytics in real time, you can eliminate the need for loading data into a data warehouse. Operational Analytics is a feature introduced in SQL Server 2016, with the same feature available in SQL Database.
A two-minute video about the benefits of using the in-memory OLTP feature is available at Azure SQL Database - In-Memory Technologies.
Source of Information : Migrating SQL Server Databases to Azure