Skip to main content

Hello,

I would like to know if this architecture is available with FMEServer :

  • 2 Servers are independent, each have their own FME Interface / Configuration
  • Nevertheless, these servers share the ENgine.

In this configuration , we can imagine for example that these servers share 2 static licences and 7000 hours of dynamic licences.

I think it is possible using 3-tier architecture , but I'm not sure 100%.

FME is getting more and more popular in my job and maybe, if technically possible, this architecture would be a solution to ease the administration.

Thank you

 

 

I don't think you can easily do that. And to be honest, it sounds like a major hassle.

 

What would the business rationale behind this be?

 

 


I agree, I don't think this is possible.

The bottleneck is almost always the number of available engines, not the number of cores, so I'm also wondering what the business case is.


I don't think you can easily do that. And to be honest, it sounds like a major hassle.

 

What would the business rationale behind this be?

 

 

thank you for your answer. I understand of course if it is not possible, but I try to answer your question :

  • On Safe side, I think it would be the same principle than sharing floating Desktop licence in an organisation. It means users need less licences for the same services, but it is already the case for FME Desktop.
  • On our side, we face a bigger amount of administrators and users. For example, the administration of the server is getting more and more complex because it uses multiple applications, API... We face also security issue because I work for the army (multiple server depending on security access). What's more,several FME Server applications deal with billions of data (bathymetry) using sometimes an engine for hours. Other jobs are very fast : That's why sharing licences would optimise the use of these engines.

We think about multiple ways to deal with the bigger amount of FMEServer users in our organisation. Of course, If this one is not technicaly possible or not desired by Safe, we won't choose this one . 🙂


thank you for your answer. I understand of course if it is not possible, but I try to answer your question :

  • On Safe side, I think it would be the same principle than sharing floating Desktop licence in an organisation. It means users need less licences for the same services, but it is already the case for FME Desktop.
  • On our side, we face a bigger amount of administrators and users. For example, the administration of the server is getting more and more complex because it uses multiple applications, API... We face also security issue because I work for the army (multiple server depending on security access). What's more,several FME Server applications deal with billions of data (bathymetry) using sometimes an engine for hours. Other jobs are very fast : That's why sharing licences would optimise the use of these engines.

We think about multiple ways to deal with the bigger amount of FMEServer users in our organisation. Of course, If this one is not technicaly possible or not desired by Safe, we won't choose this one . 🙂

Based on your description of having jobs with very varying execution times, it sounds like this webinar from yesterday could be of interest to you:

https://www.safe.com/webinars/a-safers-guide-best-practices-for-optimizing-jobs-on-fme-server/

Also: https://docs.safe.com/fme/html/FME_Server_Documentation/WebUI/Queue-Control.htm


Thank a lot you both for the answer and for the links 🙂 : "The bottleneck is almost always the number of available engines, not the number of cores" --> I totaly agree with you but there are other issues in our organisation that explain the questions about our current architecture (security; servor administration...). We try to think about multiple way. Indeed , the architecture described above is maybe not bussiness rational.


Reply