Architecting, installing and maintaining your SAS environment

SAS Visual Analytics and Visual Statistics Deployment Options

New Contributor
Posts: 2

SAS Visual Analytics and Visual Statistics Deployment Options


We have number of SAS 9.4 modules deployed on our SAS implementation e.g. enterprise guide, DI Studio, Metadata repository etc. We are looking to deploy SAS visual analytics and visual statistics. Just want to check the best deployment method. The options we are considering are.


  1. Deploy 9.4 VA/VS co-locating with existing SAS 9.4 installation. i.e. sharing the common modules like metadata repository, mid/web servers etc keeping the SAS VA/VS compute nodes separate to SAS compute nodes
  2. Deploy 9.4 VA/VS as a separate SAS instance i.e do not share the common modules. SAS VA/VS will have its own version of metadata repository server, mid/web servers and compute nodes.
  3. Deploy Viya VA/VS which is a new SAS ground up technology and is not compatible with existing SAS 9.4 product suite.

Option 3 may not fly and may take long time to decide as we have fairly large user base. So it is the choice between 1 & 2. Trying to identify the pros and cons of these options.


Option 1 seems to have following pros. But do they also become bottleneck and having a seperate instance will have flexibility? Just want to understand the best practices in SAS deployment. Do you go down seperate instances route for new modules or try to build on top with exisiting SAS estate.

  1. Single place for user administration & access control.
  2. Software patching will be simpler.
  3. Single integration point to LDAP
  4. Less cost for hardware required for common modules e.g. metadata repository & mid/web server.
  5. Can developers benefit from having a shared metadata repository?


Many thanks for your help.




Trusted Advisor
Posts: 1,757

Re: SAS Visual Analytics and Visual Statistics Deployment Options

[ Edited ]

Hello @kiks,


I would recommend to ground this decission more on the business requirements than on the technical ones, meaning:


- Do I want to have available always the latest updates on VA?

- Do I want to have a standard environment, to reduce maintenance and follow standard procedures?


At this moment, VA releases new version or maintenances quicker than other solutions. Therefore if you want to have the latest version on each of your DTAP environments, you can only chose to have a separated version (2).


In many corporations, they try to have every SAS solution on the same maintenance as much as possible, and generally latest menus one, to deliver a service as mature as possible. On this situation (1), there are many PROs, as having fully integrated environment. And if still you want to have a taste of the latest versions, to start preparing your developments to the next iteration, you can still have a single standalone VA server.


Of course, to roll in Viya might be very interesting as well! (3) Actually, it could be even another single standalone server, for now. The general recommendation is to not use the Viya version of VA/VS on production systems for now, but I would encourage you to start learning and playing with it. You would get a lot of support from SAS as well.


All in all, probably the best you can do is to contact your SAS Representative, to help you address your questions as best as possible.

New Contributor
Posts: 2

Re: SAS Visual Analytics and Visual Statistics Deployment Options

Posted in reply to JuanS_OCS

Thanks a lot for the reply.

Super User
Posts: 3,866

Re: SAS Visual Analytics and Visual Statistics Deployment Options

[ Edited ]

As well as SAS VA being on a quicker upgrade cycle than SAS BI (as mentioned by @JuanS_OCS) there is another good reason to have separate installations. That is applying maintenance on SAS BI takes much longer than with SAS VA. We haven't bothered applying maintenance on our SAS BI platform since moving to 9.4 simply because it takes so long and is too disruptive. We can't afford the downtime.


It is a lot quicker to upgrade a separate VA installation and leave traditional SAS BI just to major upgrades only.



Ask a Question
Discussion stats
  • 3 replies
  • 3 in conversation