SAS Visual Analytics
SAS Visual Analytics runs on SAS High-Performance Analytics Server and on a
SAS Intelligence Platform if it is configured as an MPP environment. SAS Visual
Analytics provides the ability for business users and business analysts to perform a
wide variety of tasks using visualizations specifically designed for exploring big data
and deriving value. This enables you to go beyond descriptive statistics and use more
specialized analytics in a very scalable way. This provides more accurate insights
into the future and aids decision making. For example, users can visually explore
data, execute analytic correlations on billions of rows of data in just seconds, and
visually present results. This helps quickly identify patterns, trends, and relationships
in data that were not evident before.
SAS Visual Analytics can be combined with the SAS High-Performance Analytics
Server and SAS In-Database to provide a high-performance model lifecycle. For
example, use SAS Visual Analytics to explore the data and decide which information
to use. Use SAS High-Performance Analytics Server to build your models, and then
use SAS In-Database to push the models into an appropriate database for scoring.
SAS Visual Analytics requires a dedicated and specialized configuration of blade
hardware such as Teradata or EMC Greenplum appliances or Hadoop HDFS
configured as an MPP cluster. This environment is always threaded. SAS options
CPUCOUNT and NOTHREADS have no effect. Instead, the NTHREADS option in
the PERFORMANCE statement provides a way to throttle thread usage. See the the
SAS Visual Analytics: User’s Guide for product information.
SAS Visual Analytics relies on the SAS LASR Analytic Server to provide a highly
scalable analytics infrastructure that is optimized for large volumes of data and
complex computations.
SAS LASR Analytic Server
The SAS LASR Analytic Server is an analytic platform that provides a secure
environment for concurrent access to data. It loads the data into memory across the
computing nodes of a SAS High-Performance Analytics Server. The SAS LASR
Analytic Server executes on the SAS High-Performance Analytics Server root node
with worker nodes across the appliance that read data into memory in parallel very
fast. If the data is not from a co-located data provider, then the data is read from the
DBMS appliance or Hadoop cluster and transferred to the root node of the SAS
High-Performance Analytics Server. Then, it is loaded into the memory of the
worker nodes. The SAS LASR Analytic Server is not influenced by CPUCOUNT or
NOTHREADS. Instead, the NTHREADSoption in the PERFORMANCE statement
throttles thread usage. Refer to the SAS LASR Analytic Server: Administration Guide
for details.
For more SAS In-Memory Analytics products, see Products & Solutions/In-Memory
Analytics.
SAS High-Performance Analytics Product
Integration
The SAS High-Performance Analytics portfolio supports many SAS solutions and
products. Some SAS products are integrated with SAS Grid Manager and (to some
extent) other high-performance analytics products in order to take advantage of parallel
processing.
SAS High-Performance Analytics Product Integration 217
SAS Enterprise Guide:
SAS Enterprise Guide executes as a SAS client and provides a front end to SAS
servers to execute SAS programs. Users create programs that take advantage of the
parallel processing capabilities in Base SAS (for example, thread-enabled procedures
and stored processes). SAS Enterprise Guide can detect whether SAS Grid Manager
is managing the environment to provide workload balancing, resource assignment,
and job prioritization. SAS Enterprise Guide can run Process Flow branches in
parallel on different grid nodes. It enables parallel execution of tasks on the same
server, and you can run tasks at the project or individual level in a SAS Grid
Manager environment. See the SAS Enterprise Guide for details.
SAS Data Integration Studio
SAS Data Integration Studio runs on the SAS Data Integration Server and
automatically takes advantage of grid computing if SAS Grid Manager is installed.
SAS Data Integration Studio 3.4 was enhanced to automatically generate SAS
applications that are enabled to execute on a SAS Grid Manager managed grid. Users
can produce grid-enabled SAS applications without any programming knowledge or
knowledge of the underlying grid infrastructure.
These SAS applications detect the existence of a SAS Grid Manager environment at
run time and distribute the execution accordingly. These grid-enabled applications
can be saved as SAS stored processes and subsequently executed by the SAS
Intelligence Platform components including SAS Web Report Studio, SAS
Information Map Studio, and the SAS Add-In for Microsoft Office. (These
applications need SAS BI or SAS Enterprise BI servers, which depend on SAS
Foundation).
SAS Data Integration Studio can execute with in-memory products SAS High-
Performance Analytics Server and SAS Visual Analytics Server configured with the
SAS Intelligence Platform as an MPP environment, which is always threaded. SAS
Data Integration Studio provides High-Performance Analytics transformations for
SAS LASR Analytic Servers or HDFS. NOTHREADS and CPUCOUNT options
have no effect. See the SAS Data Integration Studio: User’s Guide for details. The
SAS Data Integration Server is administered as part of the SAS Intelligence
Platform.
SAS Risk Dimensions
The iterative workflow in SAS Risk Dimensions is similar to that in SAS Data
Integration Studio; they both execute the same analysis over different subsets of the
data. This workflow makes them ideal for taking advantage of SAS Grid Manager to
distribute the processing across the grid. For more information, see SAS Risk
Dimensions User’s Guide.
SAS Enterprise Miner
SAS Enterprise Miner can automatically generate SAS applications that are enabled
to execute on a SAS Grid Manager grid. These SAS applications detect the presence
of a SAS Grid Manager environment at run time and distribute the execution
accordingly. The applications can also be saved as SAS stored processes and
subsequently executed by the SAS Business Intelligence components such as SAS
Web Report Studio. The DMINE, DMREG, and DMDB procedures are thread-
enabled with THREADS as the default. NOTHREADS disables multithreaded
computation.
SAS Enterprise Miner includes a key set of high-performance statistical and data
mining procedures for tasks such as data binning, imputation, scoring, and
transformations, and others. These procedures execute in the highly threaded SAS
High-Performance Analytics Server. In MPP mode, SAS Enterprise Miner distributes
data, memory, and computation across the server nodes to build predictive models. If
the SAS High-Performance Analytics Server is installed and specific Hadoop HPDM
218 Chapter 13 • Support for Parallel Processing
code is enabled, an HPDM tab is available in SAS Enterprise Miner that permits
execution on the server nodes. Scoring code from Enterprise Miner can be run inside
the database using SAS In-Database technology. For more information, see SAS
Enterprise Miner: Administration and Configuration.
SAS High-Performance Analytics Product Integration 219
220 Chapter 13 • Support for Parallel Processing
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.