Move to SAS Viya -service

As SAS 9 technology is entering the "Limited support"- status the current SAS customers face the need to decide what to do with these old SAS 9 systems they’ve developed over the years. One of the possible options for these old SAS 9 systems is to migrate them to SAS Viya -platform. SAS Viya is SAS Institute’s cloud-enabled platform for AI, analytics, and data management. 

Varjus Consulting provides SAS expert services for these SAS customers to help them to migrate these old SAS legacy systems to SAS Viya platform. Varjus Consulting services can be utilised for planning, designing and performing this migration work. 

There are several SAS tools to help companies with this migration. First of all the SAS 9 code can be executed “as-is” in SAS Viya Compute context and can be used with (almost) no alterations at all. There are also some migration tools to migrate the existing SAS DI Studio jobs and SAS Enterprise Guide projects to convert them to SAS Studio Flows in SAS Viya. Also, the SAS 9 Content Assessment -tool is useful support tool for planning the SAS Viya migration. 

Even though the migration of SAS codes, SAS DI jobs and SAS EG projects are quite straightforward there are some things to consider. First of all, SAS Viya utilises the UTF8 -character set, as the default character set for SAS 9 is Latin. This means that SAS 9 datasets need to be converted to UTF8, and it causes some issues especially for comparing and verifying the correctness of results of this migration work. Also, due to this change some of the character functions do not work in a similar manner in SAS Viya as they do in SAS 9. For example, character functions like length(), substring(), etc. needs to be changed to corresponding klength(), ksubstring(), etc. functions during this migration project. 

Second, just moving the current SAS 9 code to SAS Viya and running it “as-is” in Compute Server does not provide the full benefits of cloud computing. In order to make the code run parallel in memory the existing SAS 9 code needs to be altered to run in CAS- server (i.e. SAS Cloud Analytics Services). These alterations should be done carefully, as it requires loading the data into CAS and also the execution logic for the data processing is somewhat different in parallel processing. 

Note, that all SAS processing should not be moved to CAS as moving the data to memory also causes some overhead for the process. Only the steps that run long in the Compute server should be moved to CAS.