Running High-Performance Analytical Procedures in Asymmetric Mode

This section provides examples of how you can run high-performance analytical procedures in asymmetric mode. It also includes examples that run in symmetric mode to highlight differences between the modes. For a description of asymmetric mode, see the section Symmetric and Asymmetric Distributed Modes.

Asymmetric mode is commonly used when the data appliance and the computing appliance are distinct appliances. In order to be able to use an appliance as a data provider for high-performance analytical procedures that run in asymmetric mode on another appliance, it is not necessary that SAS High-Performance Statistics be installed on the data appliance. However, it is essential that a SAS Embedded Process be installed on the data appliance and that SAS High-Performance Statistics be installed on the computing appliance.

The following examples use a 24-node data appliance named data_appliance.sas.com, which houses a Teradata DBMS and has a SAS Embedded Process installed. Because SAS High-Performance Statistics is also installed on this appliance, it can be used to run high-performance analytical procedures in both symmetric and asymmetric modes.

The following statements load the simData data set of the preceding sections onto the data appliance:


libname dataLib teradata
        server  ="tera2650"
        user    =XXXXXX
        password=YYYYY
        database=mydb;

data dataLib.simData;
   set simData;
run;

Note: You can provision the appliance with data even if SAS High-Performance Statistics software is not installed on the appliance.

The following subsections show how you can run the HPLOGISTIC procedure symmetrically and asymmetrically on a single data appliance and asymmetrically on distinct data and computing appliances.