Skip to main content
Announcements
Qlik Introduces a New Era of Visualization! READ ALL ABOUT IT
cancel
Showing results for 
Search instead for 
Did you mean: 
mwallman
Creator III
Creator III

ENVIRONMENT PERFORMANCE Question: How to manage network latency in a Qlik Sense environment?

Hi,

How can we manage good network latency for a Qlik Sense environment?

In Qlik help: https://help.qlik.com/en-US/sense/November2018/Subsystems/PlanningQlikSenseDeployments/Content/Sense...

It says:

The current persistence model does not support geographical deployments. For best performance we recommend that you locate all your Qlik Sense servers in the same geographic location or data center as the repository database and the file share with a network latency below 4 milliseconds.

How can we achieve a network latency like this?

I am admiteddly don't know much about network latency and how to achieve it so am asking for advice. Sorry for my ignorance.

2 Replies
ganapati_h
Partner Ambassador Alumni
Partner Ambassador Alumni

@mwallmanIf you have your Qlik Sense components(proxy, engine, repository scheduler) in same network you ideally will have low network latency and all read/writes to repository database and shared file will be fast.

If they are not then you need to have high speed connectivity over WAN.

Regards,
GH

swallace104
Contributor III
Contributor III

What's your actual set up like? This may not be an issue for you to even be concerned about. Are you setting up multiple nodes or placing the different services on different servers? Are they virtual or physical servers? 

If you are setting everything up on a single server then it's not really an issue. There isn't network latency involved since everything is on the same box. If you're using multiple vm's hosted on the same blade deployment then you probably won't have a serious issue either. Even if they are separate blades using the same switch I doubt you would have an issue. Where you would really run into an issue is having the main node in one datacenter and then another node in a datacenter in another state - then you're going to have to really work out your latency.