Dist-Management

[maherhodali]

Hello,

Is it possible to limit the para shared from Slave project to Master project, I have done the same but in the opposite direction and it works fine.
In other words, master project to manage only selected tags from slave project in distributed architecture no need for full para.

Regards

[adaneau]

Hi,

Maybe by using Authorization Check Plug-in on a dist manager level. However I dont see the benefit of doing this...

What's your goal here? Is it about visualization? Then you should restrict on server directly depending of user permissions.

If you are worried about performances, the Identification caching makes it very fast and you shouldnt worry about it.

BR
Alexandre Daneau

[leoknipp]

Which datapoint types and datapoint types do you want to hide in the PARA module?
Is it only related to the PARA module or to any datapoint selector?

Best Regards
Leopold Knipp
Senior Support Specialist

[maherhodali]

Hello adaneau,

I'm connecting two big para's platforms together, so no need for me to see it all. in regarding the performance i have too many distributed projects connected and i'm afraid that it would be heavy for the event manager in the master project to handle.

Hello Leo,

I'm planning to visible selected data points which i choose to be shared with the master (No need for internal data points).
the data point selector shall be able to see the shared data points only.


Regards
Maher

[adaneau]

Hello Maher,

What would be the problem to see all para together in engineering? And if you use dp widget in runtime, then just filter what you want using object properties.

If it is for performances, no worries, dist has impact only at first connection and in case of configuration change to upgrade dpid, but not continuously.

If it is for security reason, ex keeping point protected, then use Auth check plugin.

BR
Alexandre

[leoknipp]

As Alexandre has already written the impact on performance is only related to data which needs to be transferred between distributed systems, e.g. the hotlink messages for dpConnect()/dpQueryConnect*() functions.
Data between distributed systems is only transferred on demand, there is no exchange for the complete data all the time.
Only the identification for all connected systems is loaded to the memory for every process running with "distributed = 1".

Best Regards
Leopold Knipp
Senior Support Specialist