Managing Core SQL Server 2008 R2 Technologies - Microsoft SQL Server 2008 R2 Administration - page 189

Using the trace-connections feature from SQL Azure, you may trace the connectivity problem
when necessary. The CONTEXT_INFO value can be used to obtain this value which is set
with a unique session GUID value. To obtain more information on the trace-connectivity code
example, refer to http://msdn.microsoft.com/en-us/library/ee730903.aspx .
Use the TRY…CATCH block statements to catch transactions on database operations. SQL
Azure rolls back the transaction automatically to free up the resources held by that transaction.
Further, best practices in coding include catching the transient connection termination errors,
keeping the connection open and short, and executing the transaction in a continuous loop.
For data transfer methods, always use the Bulk copy utility (bcp.exe) or the SSIS package. For
best results on data synchronization, take advantage of the SQL Azure DataSync application
from : http://www.microsoft.com/en-us/SQLAzure/datasync/default.as px .
The recommended pattern is to build your database on-premise in SQL
Express and then generate a script and execute the script on the cloud
database either using SQL Server Management Studio or SQLCMD.
When creating a logical SQL Azure server, choose the location that
is closest to you, to avoid extra bandwidth costs and achieve better
performance.
How it works...
The three services of SQL Azure are responsible for all the activities of data access. The
services, as mentioned in the previous sections, are Infrastructure, Platform, and Services.
The infrastructure layer is responsible for providing the administration of hardware and
operating systems required by the services layer. This works similar to a data-center layer,
which is shared across multiple services in a data center.
The platform layer consists of the SQL Server instances and the SQL Azure fabric
components, and required Management services. The instances are represented as the
deployed databases and their replicas. The SQL Azure fabric is the internal framework, which
automates the deployment, replication, failover, and load balancing of the database servers.
This is the most important layer that is responsible for creating three replicas of the database
instance to provide automatic failover capabilities to these instances.
The services layer, by its name, comprises external (customer) facing machines and performs
as a gateway to the platform layer. The TDS protocol (which uses port 1433 over SSL) is
responsible for providing the metering and account provisioning services to the customers.
In addition to these functions, it maintains the connectivity-routing to the primary database
instance and runtime information about your database replicas and routes the TDS coming
from client applications to the appropriate primary instance.