Best Practices When Moving SAS to Microsoft Azure

Moving your SAS (Statistical Analysis System) applications to the cloud can unlock a world of benefits, but to ensure a seamless transition and optimal performance, it's crucial to follow best practices. Microsoft Azure, with its powerful infrastructure and services, is a popular choice for hosting SAS applications. In this article, we'll explore key considerations and best practices to make your SAS migration to Azure successful.

1. Understand Your SAS Application

Before making the move, it's imperative to have a comprehensive understanding of your existing SAS application. Know all the layers, components, and success criteria of your business users. Identify the source data, third-party tools, authentication methods, and relational databases associated with your SAS application. Understanding these elements will help in planning a successful migration.

2. Define Success Criteria for Cloud Migration

Clearly define the success criteria for your SAS application in the cloud. Recognize the expectations of your business users, especially concerning speed and performance. Notably, moving to the cloud doesn't guarantee faster operations by default, so set realistic expectations and understand what success looks like in the cloud environment.

3. Azure Instance Matching

Understand the criteria of your SAS application, including core memory, I/O bandwidth, and network bandwidth. Match these requirements with the appropriate types of SAS 9 or Viya servers on Azure. Choosing the right Azure instances ensures optimal performance for your SAS applications.

4. Processor Type Considerations

Pay close attention to the processor types associated with Azure instances. For example, some instances may offer a choice of Intel processors like Broadwell, Skylake, or Cascade Lake. Be aware that the Azure portal might not allow you to select a specific processor type, so post-instance creation, run checks to ensure uniformity, especially if multiple instances are in use.

5. Premium Support and Unified Support

Consider opting for Azure Unified (premium) support when running SAS on Azure. This ensures that in case of issues, you have direct access to Microsoft Azure for assistance. Unified support can be a valuable investment to address challenges promptly.

6. Understanding VCPUs and Physical Cores

When assessing Azure instances, note that the VCPUs (virtual CPUs) mentioned on the Azure website are hyper-threaded, not physical cores. Since SAS 9 specifically runs on physical cores, divide the VCPU number by 2 to determine the physical core count.

7. Storage Considerations

Distinguish between permanent and temporary SAS data storage. Ensure that the storage configurations align with SAS application needs, especially concerning I/O throughput. Understand SAS's recommendations for megabytes per second per physical core and validate the storage setup accordingly.

8. Shared File Systems

If using SAS Grid, implement a shared file system. Evaluate options like DDN's EXAScaler, IBM's Spectrum Scale, or Azure NetApp Files based on your specific requirements. Ensure shared file systems align with the performance needs of your SAS application.

9. Azure Proximity Placement Group

Keep SAS components, security, data, and shared file systems in the same Azure Proximity Placement Group to minimize latency. This ensures that all components are in close proximity within the Azure data center.

10. Optimizing Performance

Configure instances for optimal performance by utilizing Azure Proximity Placement Groups, avoiding NMI lockups, enabling Azure Accelerated Networking, and maintaining consistent MTU sizes. Regularly refer to best practices documentation for updates and optimizations.

11. Flexibility and Future Updates

Stay informed about updates and new instance types on Azure. Cloud environments are dynamic, and newer options may provide better performance or cost savings. Plan for flexibility and periodic assessments of your SAS setup on Azure.

Conclusion

Moving SAS to Microsoft Azure can offer scalability, reliability, and enhanced performance, but success lies in careful planning and adherence to best practices. By understanding your SAS application, aligning with Azure instance types, optimizing storage, and staying informed about updates, you can ensure a smooth transition to the cloud. Embrace the power of Azure while maintaining the efficiency and reliability of your SAS applications.

For ongoing insights and support, consider becoming part of the SAS communities, where administrators and deployment experts share valuable experiences and knowledge for optimizing SAS on Microsoft Azure.