Hosting QlikView on AWS Server (Which Instance types to recommend ? )

[herard_bertrand]

Hi,

has anyone already used an AWS EC2 Server to host an on-premise QlikView Server ?

I'm completely lost in the list of Instance type :

https://aws.amazon.com/fr/ec2/instance-types/

If someone have advices or best practices to help to choose, don't hesitate 🙂

Best regards

[Chip_Matejowsky]

Moving this thread from the Qlik Sense forum to the QlikView forum for greater visibility

[Levi_Turner · Accepted solution]

There isn't a single answer to this, but my thoughts:

• r5 / r5a* series
  • This is a memory heavy instance type ideal for consumption nodes (QVS) or reload nodes where larger apps are being reloaded.
• m5 / m5a* series
  • This is a balanced VM type. In deployments with more modest apps with high reload concurrency, often the m5 series can work better for back-end nodes (QDS).

Hope that helps

* In the abstract, AMD EPYC chips have better density of cores (aka more cores available) but lower clock speeds compared to Intel's Xeon Scalable series. This surfaces in single threaded heavy operations (e.g. app open events are the common reference here). So in deployments with larger apps, it becomes wiser to prize clock speed vs. # of cores. This # of cores delta between AMD and Intel is mostly neutered in modern cloud platforms where VMs are provisioned based on fix specifications rather bare metal instances.

[herard_bertrand]

Thanks for the reply 👍