SAS Scalable Performance
Data Server 5.2 on SAS 9.4 offers the following new features and enhancements:
-
SPD Server 5.2 can read, write
and update tables in the Hadoop environment. SPD Server support for
Hadoop is enabled by the SPD Server administrator. Changes to the
SPD Server configuration must be made in the SPD Server LIBNAME parameter
file and the SPD Server server parameter file. New SPD Server Hadoop
options have also been introduced to specify Hadoop operational parameters.
Hadoop access is supported only on SPD Server 5.2 running on Linux.
For more information
about SPD Server and Hadoop, see Using SPD Server with Hadoop in SAS Scalable Performance Data Server: Administrator's Guide.
-
SPD Server 5.2 supports WHERE-processing
optimization in the Hadoop cluster using MapReduce. SPD Server WHERE
processing enables you to conditionally select a subset of observations,
so the software processes only the observations that meet specified
conditions. To optimize the performance of WHERE processing, you can
request that data subsetting be performed in the Hadoop cluster. SPD
Server instantiates the WHERE expression as a Java class and submits
the Java class to the Hadoop cluster as a component in a MapReduce
program. By subsetting the data in the Hadoop cluster, performance
might be improved by the filtering and ordering capabilities of the
MapReduce framework.
For more information,
see Configure SPD Server for Hadoop WHERE Processing Optimization with MapReduce in SAS Scalable Performance Data Server: Administrator's Guide.
-
When accessing a Hadoop domain,
parallel reads are supported without requiring a WHERE clause.
For more information,
see
SPDSHPRD=.
-
In order to support enterprise
computing environments that have existing authentication processes
and password management systems, SPD Server 5.2 provides support for
performing non-native user authentication via the SAS Metadata Server.
When user authentication is requested via the SAS Metadata Server,
the back-end server that is specified in the SAS Metadata Server configuration
will perform the user authentication.
Performing authentication
through the SAS Metadata Server can provide enhanced LDAP support
over the LDAP support provided natively by the SPD Server. Using SAS
Metadata Server to integrate SPD Server user IDs and passwords with
the framework of the platform’s authentication provider allows
the SPD Server administrator to maintain only one set of user IDs
and passwords.
For more information, see Configuring SPD Server for SAS Metadata Server Authentication in SAS Scalable Performance Data Server: Administrator's Guide.
-
The SPD Server PROC SPDO CLUSTER
LIST command supports an OUT= option with the SPD Server 5.2 release.
The new CLUSTER LIST OUT= option enables you to specify the name of
a destination file to be created with the contents of the CLUSTER
LIST output.
-
The SPD Server PROC SPDO CLUSTER
LIST command now enables the SAS ODSDEST= option. The SAS ODSDEST=
option configures the default output display system content formatting.
This is favorable because the default ODSDEST= setting is HTML, which
produces easy-to-read HTML output with embedded style sheets. This
makes it easier to read and share CLUSTER LIST results.
For more information
about the CLUSTER LIST command, see
Querying and Reading Member Tables in a Dynamic Cluster.
-
Additional protection is provided
against intrusion from unknown clients. SPD Server processes communicate
with one another and clients via TCP/IP sockets. Improvements in
TCP/IP socket monitoring allow SPD Server to identify and handle these
unknown client requests in a secure and consistent manner.
-
SPD Server 5.2 supports the sharing
of SPD Server user proxy processes with different credentials in the
same SAS session. Sharing proxy processes mitigates some of the resource
and performance issues that can occur when redundant proxy processes
are created.