SAS® Scalable Performance Data Server
Partitioning and parallel processing
- Partitions across multiple CPUs, I/O channels and disks.
- Supports the dynamic clustering of tables with identical structures.
- Uses parallel CPUs and I/O for summarizations.
- Supports the use of table aliases.
- Integrates into the SAS SQL engine.
- Provides multithreaded WHERE clause processing.
- Evaluates subsetting criteria using WHERE clause planning.
- Supports full table scans where no indexes are defined.
- Launches multiple threads to process different portions of a table simultaneously.
- Evaluates strategies for query processing to determine the least resource-intensive option via WHERE planner costing feature.
- Uses memory lookups in phase two of a star join instead of index points to a table enhances performance.
- Enables parallel write to Hadoop, SAS PROC DS2 and SAS FedSQL capabilities for SAS Scalable Performance Data Server.
Unique indexing technology
- Combines the best features of B-tree indexes and bitmap indexes by using a hybrid bitmap index.
- Analyzes each table and determines the best indexing option.
- Builds and updates multiple indexes with a single read of a data table.
- Maintains a table of statistics about indexes.
- Offers a new SQL reset option that increases the maximum number of hash joins that are available.
Query optimization
- Provides star schema optimization and parallel and optimized join methods for data models other than star schemas.
- Uses transparent restructuring and recoding of nested SQL code to create temporary tables on the fly for accelerated processing.
- Performs fast sorting on BY statements.
- Sorts on subsets of data identified with a WHERE clause or KEEP list.
- Creates single sort bins larger than 2GB to reduce the number of bins needed to complete the sort.
- Provides star join optimizations for use in more unique query instances.
SQL pass-through support
- Sends server-specific SQL statements directly to the server for execution.
- Triggers multiple, independent SQL statements for parallel execution.
- Offers the option of using a SAS library (LIBNAME statement).
Management and security
- Handles large numbers of users and increasing amounts of data.
- Offers dynamic cluster tables for enhanced manageability and scalability of large tables.
- Allocates and addresses more than 2GB of memory at a time by using 64-bit, wide-mode support.
- Provides a sophisticated security implementation that gives administrators read, write and modification access to tables on a per-user or per-group basis.
- Ensures data integrity and consistency throughout the business intelligence and analytic application life cycle via the SAS common metadata framework.
- Offers a central point of control and administration via the SAS Management Console.
- Allows user names to begin with either a character or a number.
- Provides row-level security using WHERE constraints.