AlwaysOn Availability Groups - Getting Started with SQL Server 2014 Administration - page 61

With AlwaysOn Availability Groups, you can configure your secondary replicas
to be readable. This means that you can then offload your database's reporting
requirement to the secondary database. With read-only routing, you can ensure that
the client connections that specify a read intent connection get automatically directed
to a readable secondary. You will need to have configured an Availability Group
listener to achieve this automatic redirection.
AlwaysOn Availability Groups also allow you to run certain administrative tasks
on the secondary databases. You can run the following tasks:
• Copy only full database backups
• Transaction log backups
You cannot currently make use of differential backups on secondary replicas. If you
want to use differential backups on your AlwaysOn Availability Group databases,
you will need to ensure that the differential backups are set up and configured to
run on the primary database.
There is also a downside to creating a backup of a secondary. It adds additional
complexity to your backup plan—you must ensure you know where your database
backups are stored at all times, just in case you need them in an emergency.
Building AlwaysOn Availability Groups
To create an AlwaysOn Availability Group, you need the following:
• Windows Server Failover Cluster ( WSFC )
• Enable AlwaysOn Availability Groups on each instance
• An IP address for the listener
Windows Server Failover Cluster
In order to make use of SQL Server Availability Groups, you need Windows
Server Failover Cluster. The SQL Server instances will be standalone instances,
and it is important to understand that AlwaysOn Availability Groups are not
SQL Server clustering. In order to make use of the automatic failover and client
redirection, the servers that will act as replicas in your configuration will need to be
part of Windows Server Failover Cluster. This is not as traumatic as it sounds. There
is no need for shared storage, we are not clustering the SQL Server instance, and
clustering since Windows Server 2008 R2 is pretty painless to configure. If you are a
DBA in a relatively large organization, there is a high possibility that you won't have
the permissions to build up Windows Server Failover Cluster without the help of a
Windows Administrator.