SAS Administration from the Ground Up. Anja Fischer
own tasks. One function is providing your users with the data they are requesting, or, running a Stored Process, or creating a report. And that’s really all there is to it: the tier these SAS processes (SAS servers) are running on is referred to as Server Tier or Compute Tier.
Next, we have the web tier, called the middle tier.
Middle Tier aka Web Tier
The middle tier – also called the web tier – coordinates the web traffic. It enables access to data and functionality using web clients, a browser. Think SAS Studio (users using a web browser), SAS Environment Manager (web client) or any other web-based SAS APIs. We will revisit the middle tier in a little bit.
Client Tier
The client tier runs your desktop clients and web browsers. Examples for SAS clients are SAS Enterprise Guide, SAS Add-In for Microsoft Office, SAS Enterprise Miner, and so on.
Data Tier (Data Sources) – The “Plus One” Tier
The data tier is where the data sources are stored. It is different from the SAS tiers we just described because even though it is important to SAS, it does not come from SAS. It is not used for SAS to run but for users to consume using SAS.
A data tier can be any machine that stores data that you want to access from within SAS.
Data sources can be SAS data sets, OLAP cubes, web content, DBMS data (SAS can access databases such as Oracle as one example out of many), and more. We will talk more about data in Chapter 5, Metadata Library Administration.
Table 2.1 depicts the three plus one tiers. Even though the tiers are pictured in four different boxes, the boxes do not represent physical machines. It simply pictures the different layers a SAS metadata deployment consists of, whether it is installed on one machine or multiple machines.
Table 2.1: Architecture of the SAS Intelligence Platform
| Data Tier | SAS Server Tier | Middle Tier | SAS Client Tier |
| SAS Data SetsOLAP CubesSAS Scalable Performance Data Server (SPDS)SAS Web Infrastructure Platform Data ServerThird-Party Data Sources (such as Oracle, Teradata, etc.)HadoopEnterprise Resource Planning (ERP) Systems | SAS Metadata ServerSAS Workspace ServerSAS Pooled Workspace ServerSAS OLAP ServerSAS Stored Process ServerThese servers are running SAS processes for distributed clients | SAS Web ServerSAS Web Application ServerSAS Web Infrastructure PlatformTask: providing services and applications for SAS web applications.It includes:SAS Content Server to store digital content (such as reports)andother infrastructure applications and service (such as SAS Deployment Backup and Recovery tool, and more)Web Clients:(run in an instance of the SAS Web Application Server). The SAS web clients are:SAS Web Report StudioSAS Information Delivery PortalSAS BI PortletsSAS BI DashboardSAS Help Viewer for the WebOther SAS Web Applications and SolutionsSAS Environment Manager(server process that includes a web application server, providing a web-based administrative interface) | Desktop Clients:SAS Add-In for Microsoft OfficeSAS Data Integration StudioSAS Enterprise MinerSAS Forecast StudioSAS Enterprise GuideSAS Information Map StudioSAS Management ConsoleSAS Model ManagerSAS OLAP Cube StudioSAS Workflow StudioJMPOther SAS analytics products and solutionWeb Browser to surface web ApplicationsMobile Devices, if applicable, to view certain type of reports |
Figure 2.3 shows this from a very simplified layer perspective.
Figure 2.3: A layered view of SAS platform architecture.
To take a closer look at these tiers, take a look at the Architecture Overview section in SAS® 9.4 Intelligence Platform: Administration / SAS Intelligence Platform: Overview, available at: .https://go.documentation.sas.com/?cdcId=bicdc&cdcVersion=9.4&docsetId=biov&docsetTarget=titlepage.htm&locale=en
SAS Server Tier
Let’s come back and talk a bit more about the SAS Server tier. I would like to spend some time discussing its components because as a SAS administrator, this is super important!
To refresh your memory, SAS servers are SAS Server processes running on one or more physical or virtual server machines. The SAS server tier is nothing but a bucket of SAS processes that are based on “sas.exe”. Every “sas.exe” has a different role or functionality, such as:
metadata server (in-memory)
workspace server (interactive sessions)
stored process and pooled workspace server (trusted sessions)
if you have SAS Visual Analytics for example, the SAS LASR Analytic Servers (in-memory for SAS Visual Analytics). We will not address SAS Visual Analytics; this is simply meant as an example.
Let’s discuss each of these in turn.
Metadata Server
The metadata server is the heart of your environment, the foundation. If it’s not working, SAS is not working. Going back to the house-building analogy: if the foundation is weak or breaks, the house will crumble.
The metadata server is an in-memory server. It keeps your environment running and stores all your assets: metadata about where to find your assets, where assets can be data, users and groups, etc. It is a centralized resource for storing, managing, and also providing metadata for all your SAS applications, for all your users.
In-memory: when your users request data, a copy of the physical data is stored into memory. From there on, your users are going “against” memory. This speeds up the process so that the access time is quicker. Memory is flushed when you pause and resume, or, stop and restart the metadata server.
When speaking about metadata assets, we are referring to libraries, users, groups, SAS folders – everything you create using SAS Management Console or SAS® 9.4 Open Metadata
Interface (metadata server programming language). SAS applications also create and manage metadata.
To learn more about the SAS 9.4 Open Metadata Interface, take a look at https://go.documentation.sas.com/api/docsets/omaref/9.4/content/omaref.pdf
These assets are stored in a metadata repository. A repository is like a box in which you save all your belongings. Your main repository is Foundation. You can also create custom and project repositories. When SAS is installed, only one repository is created, which is the Foundation one.
A custom repository can be used – as one example – in cases where you might have very sensitive data that should only be available to a certain group of users. The custom repository and its contents could be created in a directory where only these users have access to.
Project repositories are used for Change Management and are available for SAS Data Integration Studio only. Users can check out and lock metadata so that metadata can be modified and tested, to be checked back in afterward, which then unlocks the metadata.
In this book, we will focus on the Foundation repository. To learn more about custom and project repositories, check out About Repositories at https://go.documentation.sas.com/?cdcId=bicdc&cdcVersion=9.4&docsetId=bisag&docsetTarget=p1b7gxgkbu04zbn1dhnh3pb8c6zs.htm&locale=en
When you log on to SAS Management Console, Foundation is chosen per default as shown in Figure 2.4.
Figure 2.4: SAS Management Console
| Note: The path defaults to SAS-conf-dir/Lev1/SASMeta/MetadataServer/MetadataRepositories/. Do not rename or delete the metadata server |