Skip to main content

PartnerWorld > Products > Systems, servers, and storage > Technical >

AIX 6 best practice for SAS Enterprise Business Intelligence (SAS eBI) users on IBM POWER6

by Joseph Pu, Alfredo Mendoza, Frank Bartucca, Harry Seifert, Frank Battaglia

Last updated: 2009-01-06

PDF iconDownload the white paper (886 KB)
Links to external URLGet Adobe® Reader®

 
Rate this article
  Feedback

© Copyright IBM Corporation, 2008. All Rights Reserved.
All trademarks or registered trademarks mentioned herein are the property of their respective holders.

Abstract
Monitoring each SAS component and understanding how users use the environment helps understand how to best optimize the infrastructure to deliver the best performance. This paper focuses on best practice suggestions and performance settings of the IBM AIX 6.1 OS for SAS Enterprise Business Intelligence (SAS eBI).
Introduction

The SAS paper available at http://support.sas.com/resources/papers/happyIT.pdf provides system management guidelines for system (hardware and operating system) resources from a SAS eBI application’s perspective.

In this benchmark, the team targets a 500-user environment with varying types of workloads. The workload distribution is as follows:

  • 10 SAS Enterprise Guide sessions developing reports.
  • 20 advanced analytics sessions performing SAS statistical functions.
  • 10 SAS Enterprise Guide sessions developing reports and SAS Data Integration Server code.
  • 80 SAS Add-In for Microsoft® Office sessions.
  • 80 SAS Web Report Studio ad hoc, single-user sessions.
  • 70 SAS Web Report Studio static report sessions.
  • 80 SAS Stored Process user sessions.
  • 80 SAS OLAP cube viewer single-user sessions.
  • 70 SAS Enterprise BI Server portal and dashboard viewer sessions.
A typical SAS BI solution is made up of the following components:
  • SAS Metadata Server
  • SAS OLAP Server
  • SAS Server
  • The mid-tier server

Each component has its own characteristics on how it uses the underlying resources. For example, the SAS Server is both processor and I/O intensive, whereas the OLAP Server is mostly processor intensive. So while some servers need to have I/O tuning done, some require a more dynamic approach of rebalancing processor resources.

In real-world environments, when a performance concern arises, the methodology the team put forth in this document will demonstrate the effectiveness of a heuristic approach that examines each SAS component, the resources that it runs on and the different tuning parameters that need to be modified to get the best overall performance.