1
Introduction : Everyone is dealing with large amount of data now, but how that big data impacts your business depends on what you are trying to do with it. Healthcare and pharmaceutical companies are faced with tight deadlines, complex analyses with long run-time. To solve that issue and reduce overall run-time and increase system availability, one option is the SAS Grid. The SAS Grid Manager uses the power of distributed computing across a cluster of server nodes that provide redundancy, scalability, and parallel (asynchronous) processing resulting in increased reliability, easy expansion, and reduction in overall run-time. Migrating towards this power of distributed computing does pose some challenges, but with careful planning and use of simple tools, the hurdles will quickly become a distant memory. We will address some of the key challenges we faced and its solutions while working on multiple SAS Grid implementations. Learning : Migrating to SAS grid into a new operating environment is a complex and challenging task. Especially, when it is across platform (different OS level), different OS Bit sizes and across SAS versions. But with careful planning and use of simple toolset it can be managed with little pain. Some of the key challenges and solution listed in this paper can help towards better understanding of things to consider when working on SAS Grid migration. Finally, do not underestimate the training requirement of your team of programmers on expected changes towards transition on to the SAS Grid. Across Platfor ms Across platforms FILE/DIRECTORY STRUCTURE SYSTEM/ ENVIRONMENT VARIABLES OS PLATFORM UTILITIES ADHOCH BUILD OS SCRIPT 32Bit/64Bit changes VALIDVARNAM E VALIDFMTNAM E SAS CATALOGS Code Handling DDE (DYNAMIC DATA EXCHANGE) FUNCTIONALITY SAS STATEMENTS THAT SPECIFY A FILE OR DIRECTORY PATH BACK SLASH AND FORWARD SLASH CALLING OS OR EXTERNAL COMMANDS SYSTEM MACRO NAMES AND ENVIRONMENT VARIABLES DBMS = EXCEL OPTION FOR EXCEL POP-UP WINDOW OR DISPLAY STATEMENTS SAS version changes EXTENDOBSCOUNTE R VALIDMEMNAME VALIDFMTNAME V6CREATEUPDA TE Data Handling INCOMPATIBLE DATA SET ENCODING SAS CATALGS Code Modification DM STATEMENTS ENDSAS AND ABORT STATEMENTS VALIDVARNAME MACRO CATALOG NAMING FORMAT CATALOG MACRO NAME CASE CHANGE OLE DB API Business process changes SECURITY AUTHENTICATION PERMISSIONS & PRIVACY Operation process changes JOB SCHEDULING JOB MONITORING PERFORMANCE MONITORING REGIONAL OVERLAP DURING JOB PROCESSING INTERFACE WITH REMOTE FILE SYSTEM AND SERVERS OUTSIDE GRID PhUSE 2016 Barcelona SAS Grid Migration - Challenges, Solutions and Learnings by Amol Waykar and Eric Brinsfield 32Bit/ 64Bit Change Acros s SAS Versio n Data Handlin g Code Handlin g Busines s Process Code Modificati on Operatio n Process Won’t work on SAS Grid and across platform Require changes to work on SAS Grid and across platform May not be required, since SAS Grid comes with this feature to manage it better

Sas Grid Migration and Roadmap

Embed Size (px)

Citation preview

Page 1: Sas Grid Migration and Roadmap

Introduction :Everyone is dealing with large amount of data now, but how that big data impacts your business depends on what you are trying to do with it. Healthcare and pharmaceutical companies are faced with tight deadlines, complex analyses with long run-time. To solve that issue and reduce overall run-time and increase system availability, one option is the SAS Grid. The SAS Grid Manager uses the power of distributed computing across a cluster of server nodes that provide redundancy, scalability, and parallel (asynchronous) processing resulting in increased reliability, easy expansion, and reduction in overall run-time. Migrating towards this power of distributed computing does pose some challenges, but with careful planning and use of simple tools, the hurdles will quickly become a distant memory. We will address some of the key challenges we faced and its solutions while working on multiple SAS Grid implementations.

Learning :Migrating to SAS grid into a new operating environment is a complex and challenging task. Especially, when it is across platform (different OS level), different OS Bit sizes and across SAS versions. But with careful planning and use of simple toolset it can be managed with little pain. Some of the key challenges and solution listed in this paper can help towards better understanding of things to consider when working on SAS Grid migration. Finally, do not underestimate the training requirement of your team of programmers on expected changes towards transition on to the SAS Grid.

Across Platforms

Across platformsFILE/DIRECTORY STRUCTURE

SYSTEM/ ENVIRONMENT VARIABLES

OS PLATFORM UTILITIES

ADHOCH BUILD OS SCRIPT

32Bit/64Bit changesVALIDVARNAME

VALIDFMTNAME

SAS CATALOGS

Code HandlingDDE (DYNAMIC DATA EXCHANGE) FUNCTIONALITY

SAS STATEMENTS THAT SPECIFY A FILE OR DIRECTORY PATH

BACK SLASH AND FORWARD SLASH

CALLING OS OR EXTERNAL COMMANDS

SYSTEM MACRO NAMES AND ENVIRONMENT VARIABLES

DBMS = EXCEL OPTION FOR EXCEL

POP-UP WINDOW OR DISPLAY STATEMENTS

SAS version changesEXTENDOBSCOUNTER

VALIDMEMNAME

VALIDFMTNAMEV6CREATEUPDATE

Data HandlingINCOMPATIBLE DATA SET ENCODING

SAS CATALGS

Code ModificationDM STATEMENTS

ENDSAS AND ABORT STATEMENTS

VALIDVARNAME

MACRO CATALOG NAMINGFORMAT CATALOG

MACRO NAME CASE CHANGE

OLE DB API

Business process changesSECURITY

AUTHENTICATION

PERMISSIONS & PRIVACY

Operation process changesJOB SCHEDULING

JOB

MONITORING

PERFORMANCE MONITORING

REGIONAL OVERLAP DURING JOB PROCESSING

INTERFACE WITH REMOTE FILE SYSTEM AND SERVERS OUTSIDE GRID

PhUSE 2016 Barcelona

SAS Grid Migration - Challenges, Solutions and Learnings

by Amol Waykar and Eric Brinsfield

32Bit/ 64Bit

Change

Across SAS

Version

Data Handling

Code Handling

Business Process

Code Modification

Operation Process

Won’t work on SAS Grid and across platform

Require changes to work on SAS Grid and across platform

May not be required, since SAS Grid comes with this feature to manageit better