Select to view content in your preferred language

Implement license concurrent model for named users on ArcGIS Online/Portal for ArcGIS

12-13-2013 06:09 AM
Status: Open
Esri Contributor

Example: Consider organization with 300 field workers. They are divided into 6 different ArcGIS Online groups. Each group is tight to specific region, so members of each group should see just one specific region. Only 100 workers can work concurrently at the same time = need to have access to the AGOL (to particular group) and do the same work. Their work is very limited - they need to just display data.
In such case concurrent license model would be seriously helpful. To have named user for each worker is not very efficient.

Another scenarios is for organizations that utlize concurrent desktop licensing. Currently as administrator of both Desktop and AGOL I have no "fair" way to distrubute named user accounts to AGOL and not everyone needs it or uses it everyday. Some projcts are very short in nature and then these users might not use the portal for a months or even a year. I agree with the authors suggestion on some more flexibility in named users and implementing AGOL concurrent licensing. 
I can easily envision a license manager in the cloud. ESRI knows how many users accounts we have, so just keep track of how many are connected. If we have 30 AGOL connections then throttle us. If #31 tries to connect then deny access. Of course, this goes counter to the current LM technology/philosophy which does not time-out/disconnect users for inactivity (another unmet enhancement!!)

I also have no way to fairly distribute our AGOL named user accounts.
Our whole system is built on the concurrent use model (shared resource). Just because AGOL is a cloud SaaS does not change how we work.
We will not be implementing user based access to AGOL until something is done to bridge the differences in licensing. I have a small handful of "content publishers" but they do not have admin rights, they cannot make maps or apps public.  


FlexLM is circa 1988, and it's amazing that this performs so much better than the named user model. I have many users in my Org Account. 70% of them don't sign in in a 30 day window. But ALL users need a named user account-- the total bucket of those named users is limited. That means that I have to delete old user accounts to make room for new ones. This creates a ton of overhead for the administration of this tool. 

A true concurrent usage model would work so much better.