An oversimplified definition of a database cursor might be this: A cursor is database functionality that allows you to point to a certain location within a result set and allows you to move forward (and sometimes backward, depending upon the cursor type) through the result set one row at a time. [...] In the paging [scenario], I'll use a static cursor since that cursor type would seem to satisfy the requirements of many web-based applications.

He shows how to execute a simple query with a dynamic cursor by specifying it in the connection call. He then uses the sqlsrv_num_rows to find the number of records returned and a sqlsrv_fetch_array call to pull just the page you need. He also includes some handy code to paginate the results, complete with links.