Quick Check

1.	Name and explain the two basic kinds of partitioning.
2.	What is the purpose of a partition scheme?
3.	What two methods can be used to specify the name of a linked server in a query?
4.	What is the purpose of a Distributed Partition View (DPV)?
5.	Which scale-out technique uses middleware to route requests to the appropriate database server?
6.	Which SQL Server 2005 features can applications built with the Service-Oriented Database Architecture (SODA) take advantage of?
7.	Which scale-out technique is best for transparently distributing data across multiple servers?
8.	What does SQLCLR provide?

Answers

1.	Vertical partitioning involves removing columns from a table and creating separate tables to store the data. Horizontal partitioning involves removing rows from a table based on a key partitioning column.
2.	A partition scheme is used to specify to which filegroups a partition will be assigned. The partition function associated with this scheme will also be provided because this specifies a range of values that will fall within each partition.
3.	For a query involving a linked server, you can either use a four-part name or the OPENDATASOURCE function, which is used to execute an ad hoc query against a remote server.
4.	A Distributed Partition View (DPV) is used to bring together data that was horizontally partitioned into a single view for the client. The query for the view is built using UNION statements that join together one or more linked servers that contain the partitioned data.
5.	Data Dependent Routing (DDR) uses the application or middleware to route requests to the appropriate server where the data has been partitioned. Lookup tables are used to associate the data with the table based on a partition key value.
6.	Applications built using the SODA scale-out technique can take advantage of SQLCLR, Notification Services, Query Notifications, Service Broker, and Native XML Web Services.
7.	A Scalable Shared Database is a read-only database that was built on a SAN and can service up to eight different instances of SQL Server. No application changes are needed to implement this scale-out technique.
8.	SQLCLR allows developers to write managed code using the .NET Framework and execute the assembly on SQL Server. Stored procedures, user-defined functions, and triggers can be written using Visual Basic, C#, or J#. In cases where complex operations must be performed on the data, SQLCLR can enhance application performance.