Optimising SAS on Mainframes: Challenges and Modernisation Strategies

SAS and the zSeries Mainframe History

The IBM zSeries Mainframe has been a vital processing platform for many major companies, known for its security, resilience, and performance, especially in handling large data volumes. SAS has been integrated with the mainframe since 1971, maintaining compatibility and functionality over the decades. Companies have substantial investments in bespoke SAS code developed over the years.

Many mainframe operations depend on MXG from Merrill Consultants for processing system logs and metrics, leveraging complex SAS macros. SAS’s ITRM is based on MXG and relies on it for processing. Strategies involving MXG require special attention due to its importance.

SAS 9.4 remains supported, but the focus is shifting to the newer Viya platform, which integrates with modern technologies like Python and R. Viya will not be available on zSeries mainframes, pushing customers to consider other platforms. SAS 9.4 may see limited enhancements but will continue to receive fixes and security updates. Upgrading to the latest SAS 9.4M8 is crucial for regulatory compliance, especially with newer z/OS versions.

Talent and Skills: While SAS skills are widespread, mainframe-specific SAS expertise is rarer. Universities produce graduates skilled in newer tools like Python and R, prompting businesses to integrate these into their data analysis functions. The mainframe workforce is aging, raising concerns about future talent. While SAS skills are transferable across platforms, there is still a potential skills gap to address.

For committed mainframe users:

1. Upgrade to SAS 9.4M8: Ensures continued support and compliance.
2. Switch to Altair’s SLC: Offers SAS language compatibility with potential cost savings.
3. Adopt Open-Source Tools: Python, supported by IBM’s Open Enterprise SDK on z/OS, offers powerful analytics capabilities.

For those planning to exit the mainframe:

1. Assess SAS Code Compatibility: Use tools to evaluate necessary changes, whether it be to execute on a SAS 9.4 or Altair SLC platform.
2. Integrate New Technologies: Consider Python, R, and SAS Viya features.
3. Control Costs: Balance software replacement savings with migration expenses.
4. Retrain Users: Prepare the workforce for new tools.
5. Evaluate MXG Transition: MXG can run on ASCII platforms under SAS or SLC.

Moving SAS workloads from the mainframe involves:

• Converting EBCDIC to ASCII codepages.
• Handling mainframe-specific data types.
• Adapting data source access methods.
• Addressing unsupported SAS procedures and functions.
• Rewriting embedded TSO or CALL statements for the new platform.

MXG can execute on ASCII platforms with minimal changes. Transition methods are well-documented, and despite the added data transfer time, new platforms may offer better processing times.

Mainframe SAS customers need a clear strategy for future data analytics, considering regulatory needs, modern technologies, and IT budget constraints. Understanding current usage and future goals is key to successful Modernisation.