Introduction to Dynamic Cluster Tables :: SAS® Scalable Performance Data Server 5.3: User's Guide

Overview

SPD Server is designed to meet the storage and performance demands that are associated with processing large amounts of data using SAS. As the size of the data grows, the demand to process that data increases, and storage architecture must change to keep up with business needs. One way that the server enables access to large amounts of data is by offering dynamic cluster tables. A dynamic cluster table is a collection of SPD Server tables that are presented to the end-user application as a single table through a metadata layer acting similar to a view. The cluster table is Read-only. However, features are available that enable creators to add, remove, and replace member tables while keeping the cluster table available to end users.

This ability to update the cluster while it remains online provides for rolling updates. When SPD Server is used with SAS/CONNECT Multi-processing CONNECT (MP CONNECT) software, cluster tables can quickly and easily deliver consolidated results from analytic processes that are run on multiple grid nodes. The MP CONNECT software uses multiple CPUs to process tasks concurrently, in parallel. Each CPU delivers an output table, which becomes a member of the cluster table.

Benefits of Dynamic Cluster Tables

Dynamic cluster tables are virtual table structures. Because cluster metadata is used to manage the data that is contained in the members, the clusters do not require significant disk space beyond that already taken by the member tables.
Because a dynamic cluster table consists of numerous smaller server tables, the architecture enables parallel loading and processing. Cluster table loads and refreshes can be broken down into multiple tasks that are performed concurrently.
SPD Server provides a simple, straightforward interface for creating and managing dynamic cluster tables. These statements are available in PROC SPDO and as explicit pass-through statements. For reference information, see the SPDO Procedure in SAS Scalable Performance Data Server: Administrator’s Guide.

Server Authorizations and Cluster Tables

You must have Control access to any member tables that you use in a dynamic cluster table.
Access control lists (ACLs) can be defined on a dynamic cluster table after it is created. The permissions that are specified in the dynamic cluster table ACL are applied when the server accesses the dynamic cluster table. Any individual ACL that is defined on a member table does not apply during the time that the member table is part of a dynamic cluster table.
You must have Control access to the dynamic cluster table itself to destroy the table or undo the cluster.

Anonymous User

There is an ANONYMOUS user ID that any SPD Server user can specify with no password. Any resource (table, view, catalog) that is created with the ANONYMOUS user ID can be viewed and controlled by all users who have access to the domain in which the resource exists. As such, any table created by ANONYMOUS can be used in a dynamic cluster table created by any server user. In addition, all server users can read a dynamic cluster table that was created by ANONYMOUS.

The ANONYMOUS user can place ACLs on a resource to limit access to the resource. ANONYMOUS cannot access any other server user’s resources. Server administrators can remove ANONYMOUS access to the server. Consult with your server administrator to find out whether ANONYMOUS access is available on your server.