Planning Your Hardware and Software |
Overview |
Data sources and warehouse data stores can reside on any host that can run any currently supported release of SAS. Your SAS installation coordinator will help you choose the best data host configuration for your project. Some typical configurations are described in this section.
Note: For simplicity, the figures in this section show SAS/Warehouse Administrator and its metadata repositories installed on the same host. This is not a requirement. Any of the metadata host configurations described in Metadata Host Configuration could also be used with the data host configurations described in this section.
SAS and SAS/Warehouse Administrator can be used to read or write data in a proprietary data base management system (DBMS), such as DB2 or Oracle. Under Version 7 or later of SAS, you can use a SAS/ACCESS LIBNAME statement to read and write DBMS data as if it were SAS data. Under Version 6 or later of SAS, you can use other SAS/ACCESS methods, such as an SQL Pass-through view to read DBMS data, and PROC DBLOAD to load data into DBMS tables.
If your project plan requires you to read or write DBMS data, your data host configuration must include the software to support at least one of the methods in the following two tables.
Note: By default, for new DBMS data stores, SAS/Warehouse Administrator generates code that uses SAS/ACCESS LIBNAME statements to read and write data in DBMS format.
Product | Purpose | Where Installed | |
---|---|---|---|
Base SAS, Version 7 or later | Supports SAS/ACCESS LIBNAME statement. | On the host where the SAS/ACCESS LIBNAME statement is executed. | |
SAS/ACCESS engine for target DBMS | Read or write DBMS data as if it were SAS data. | same | |
Client software for target DBMS | Handles local or remote connection to DBMS. | same |
Note: For DBMS tables created with 1.x releases of SAS/Warehouse Administrator, or for new DBMS tables with code generation level 1.1 Load Steps, SAS/Warehouse Administrator creates an SQL Pass-through view to access DBMS data and uses PROC DBLOAD to load data into DBMS tables.
Product | Purpose | Where Installed | |
---|---|---|---|
Base SAS, Version 6 or later | Supports SAS/ACCESS | On the host where the SAS/ACCESS statement is executed. | |
SAS/ACCESS | Read or write DBMS data. | same | |
SAS/CONNECT | Supports connections to files, hardware resources, and SAS software on various remote hosts to use with a SAS session on a local host. | On both the local and remote hosts. |
A third alternative is to use SAS/Warehouse Administrator add-ins that generate the code required to load a warehouse table into a DBMS, such as Oracle, using native loading software. For instructions on obtaining add-ins, see Customizing SAS/Warehouse Administrator.
Local Job, Local Data |
In this configuration, SAS/Warehouse Administrator, its data sources, and its warehouse data stores are installed on the same machine. The Job and all of its processes execute on the SAS/Warehouse Administrator host.
The configuration shown in Figure 3.5 supports the following scenario:
SAS/Warehouse Administrator submits the Job to its SAS session on Host 1.
The Job calls one or more extraction processes that execute on Host 1 and send their output to Host 1.
The Job calls one or more Load processes that execute on Host 1 and send their output to Host 1.
To implement a data host configuration similar to the one shown in Local Job, Local Data, you will need the following software:
Software | Host | |
---|---|---|
Base SAS, Version 8 or later | Host 1 | |
SAS/Warehouse Administrator | Host 1 |
If your project plan requires you to read or write DBMS data, your configuration must include the software to support at least one of the methods described in Access to Data In DBMS Format.
Local Job, Local Data with Support for Job Scheduling and Tracking shows the same configuration with additional software to support the scheduling and tracking of Jobs through SAS/Warehouse Administrator. Note that a SAS/SHARE server controls access to the Jobs Information libraries.
Local Job, Local Data with Support for Job Scheduling and Tracking
The configuration shown in Figure 3.6 supports the following scenario:
SAS/Warehouse Administrator submits the Job to a scheduling server application on Host 1, such as CRON on a UNIX host, or the AT command (Schedule service) on a Windows NT host.
The Job calls one or more extraction processes that execute on Host 1 and send their output to Host 1.
The Job calls one or more Load processes that execute on Host 1 and send their output to Host 1.
The Job updates the Jobs Information library with the status of the Job.
Local Job, Remote Data |
In this configuration, SAS/Warehouse Administrator is installed on one host, and its data sources and data stores reside on one or more remote hosts. The Job executes on the SAS/Warehouse Administrator host. The processes within the Job can execute on the same host, or they can be submitted remotely from the Job to run on the remote hosts.
The configuration shown in Figure 3.7 supports the following scenario:
SAS/Warehouse Administrator submits the Job to its SAS session on Host 1.
The Job calls one or more extraction processes that execute on Host 1, read data from Host 2, and send their output to an appropriate host.
The Job calls one or more Load processes that execute on Host 1 and send their output to Host 3.
To implement a data host configuration similar to the one shown in Local Job, Remote Data, you will need the following software:
Software | Host | |
---|---|---|
Base SAS, Version 8 or later | Host 1 | |
Base SAS, any supported version | Host 2, Host 3 | |
SAS/Warehouse Administrator | Host 1 | |
Remote communication software | See Note below |
Note: In the data host configuration shown in Local Job, Remote Data, if the processes on Host 1 must read or write SAS data on a remote host, SAS/CONNECT is required on Host 1 and the remote host. If the processes on Host 1 must read from or write to a proprietary DBMS, your configuration must include the software to support at least one of the methods described in Access to Data In DBMS Format.
Local Job, Remote Data with Support for Job Scheduling and Tracking shows the same configuration with additional software to support the scheduling and tracking of Jobs through SAS/Warehouse Administrator. Note that a SAS/SHARE server controls access to the Jobs Information libraries.
Local Job, Remote Data with Support for Job Scheduling and Tracking
The configuration shown in Figure 3.8 supports the following scenario:
SAS/Warehouse Administrator submits the Job to a scheduling server application on Host 1.
The Job calls one or more extraction processes that execute on Host 1, read data from Host 2, and send their output to an appropriate host.
The Job calls one or more Load processes that execute on Host 1 and send their output to Host 3.
The Job updates the Jobs Information library with the status of the Job.
Remote Job, Local Data |
In this configuration, SAS/Warehouse Administrator is installed on one host; its data sources reside on one or more remote hosts, and its data stores reside on the SAS/Warehouse Administrator host. The Job is submitted to a remote host, and all of its processes execute on remote host(s), where the source data is located.
The configuration shown in Figure 3.9 supports the following scenario:
SAS/Warehouse Administrator submits the Job to a SAS session on Host 2.
The Job calls one or more extraction processes that execute on Host 2, read data from Host 2, and send their output to an appropriate host.
The Job calls one or more Load processes that execute on Host 2 and use PROC UPLOAD to send their output to Host 1.
To implement a data host configuration similar to the one shown in Remote Job, Local Data, you will need the following software:
Software | Host | |
---|---|---|
Base SAS, Version 8 or later | Host 1 | |
Base SAS, version appropriate for the processes being executed | Host 2 | |
SAS/Warehouse Administrator | Host 1 | |
SAS/CONNECT | Host 1, Host 2 |
SAS/CONNECT is required on the SAS/Warehouse Administrator host and the Job host in order to support Remote Compute Services.
If the processes on Host 2 need to read from or write to a proprietary DBMS, your configuration must include the software to support at least one of the methods described in Access to Data In DBMS Format.
Remote Job, Local Data with Support for Job Scheduling and Tracking shows the same configuration with additional software to support the scheduling and tracking of Jobs through SAS/Warehouse Administrator. Note that a SAS/SHARE server controls access to the Jobs Information libraries on Host 1. The Job on Host 2 uses SAS/SHARE to communicate with the SAS/SHARE server. Accordingly, Host 2 must have a license for SAS/SHARE.
Remote Job, Local Data with Support for Job Scheduling and Tracking
The configuration shown in Figure 3.10 supports the following scenario:
SAS/Warehouse Administrator submits the Job to a scheduling server application on Host 2.
The Job calls one or more extraction processes that execute on Host 2, read data from Host 2, and send their output to an appropriate host.
The Job calls one or more Load processes that execute on Host 2 and send their output to Host 1.
The Job updates the Jobs Information library with the status of the Job.
Remote Job, Remote Data (Example 1) |
In this configuration, SAS/Warehouse Administrator is installed on one host; its data sources reside on one or more remote hosts, and its data stores reside on one or more remote hosts. The Job is submitted to a remote host, and all of its processes execute on remote host(s), where the source data and the data stores are located. You could use this configuration to execute the SAS/Warehouse Administrator Job on a remote host where data stores are located.
Remote Job, Remote Data (Example 1)
The configuration shown in Figure 3.11 supports the following scenario:
SAS/Warehouse Administrator submits the Job to a SAS session on Host 2.
The Job calls one or more extraction processes that execute on Host 2, read data from Host 3, and send their output to an appropriate host (probably Host 3).
The Job calls one or more Load processes that execute on Host 2 and send their output to Host 2.
To implement a data host configuration similar to the one shown in Remote Job, Remote Data (Example 1), you will need the following software:
Software | Host | |
---|---|---|
Base SAS, Version 8 or later | Host 1 | |
Base SAS, version appropriate for the processes being executed | Host 2 | |
Base SAS, any supported version | Host 3 | |
SAS/Warehouse Administrator | Host 1 | |
SAS/CONNECT | Host 1, Host 2, perhaps Host 3 |
SAS/CONNECT is required on the SAS/Warehouse Administrator host (Host 1) and the Job host (Host 2) in order to support Remote Compute Services.
If the extraction processes on Host 2 need to read SAS data on a remote host (Host 3), SAS/CONNECT or SAS/SHARE is required on Host 2 and the remote host. If the processes on Host 2 needs to read from or write to a proprietary DBMS on a remote host, your configuration must include the software to support at least one of the methods described in Access to Data In DBMS Format.
Example 1 with Support for Job Scheduling and Tracking shows the same configuration with additional software to support the scheduling and tracking of Jobs through SAS/Warehouse Administrator. Note that a SAS/SHARE server controls access to the Jobs Information libraries on Host 1. The Job on Host 2 uses SAS/SHARE to communicate with the SAS/SHARE server. Accordingly, Host 2 must have a license for SAS/SHARE.
Example 1 with Support for Job Scheduling and Tracking
The configuration shown in Figure 3.12 supports the following scenario:
SAS/Warehouse Administrator submits the Job to a scheduling server application on Host 2.
The Job calls one or more extraction processes that execute on Host 2, read data from Host 3, and send their output to an appropriate host (probably Host 3).
The Job calls one or more Load processes that execute on Host 2 and send their output to Host 2.
The Job updates the Job Information library with the status of the Job.
Remote Job, Remote Data (Example 2) |
This configuration is similar to the one described in Remote Job, Remote Data (Example 1). You could use this configuration to execute the SAS/Warehouse Administrator Job on a remote host where data sources are located.
Remote Job, Remote Data, Example 2
The configuration shown in Figure 3.13 supports the following scenario:
SAS/Warehouse Administrator submits the Job to a SAS session on Host 2.
The Job calls one or more extraction processes that execute on Host 2, read data from Host 2, and send their output to an appropriate host.
The Job calls one or more Load processes that execute on Host 2 and send their output to Host 3.
To implement a data host configuration similar to the one shown in Remote Job, Remote Data, Example 2, you will need the following software:
Software | Host | |
---|---|---|
Base SAS, Version 8 or later | Host 1 | |
Base SAS, version appropriate for the processes being executed | Host 2 | |
Base SAS, any supported version | Host 3 | |
SAS/Warehouse Administrator | Host 1 | |
SAS/CONNECT | Host 1, Host 2, perhaps Host 3 |
SAS/CONNECT is required on the SAS/Warehouse Administrator host (Host 1) and the Job host (Host 2) in order to support Remote Compute Services.
If the Load processes on Host 2 need to write SAS data on a remote host, SAS/CONNECT or SAS/SHARE is required on Host 2 and the remote host (Host 3). If the processes on Host 2 need to read from or write to a proprietary DBMS on a remote host, your configuration must include the software to support at least one of the methods described in Access to Data In DBMS Format.
Example 2 with Support for Job Scheduling and Tracking shows the same configuration with additional software to support the scheduling and tracking of Jobs through SAS/Warehouse Administrator. Note that a SAS/SHARE server controls access to the Jobs Information libraries on Host 1. The Job on Host 2 uses SAS/SHARE to communicate with the SAS/SHARE server. Accordingly, Host 2 must have a license for SAS/SHARE.
Example 2 with Support for Job Scheduling and Tracking
The configuration shown in Figure 3.14 supports the following scenario:
SAS/Warehouse Administrator submits the Job to a scheduling server application on Host 2.
The Job calls one or more extraction processes that execute on Host 2, read data from Host 2, and send their output to an appropriate host.
The Job calls one or more Load processes that execute on Host 2 and send their output to Host 3.
The Job updates the Jobs Information library with the status of the Job.
Copyright © 2012 by SAS Institute Inc., Cary, NC, USA. All rights reserved.