Why UE policy Command is needed?

Hi All.
Could someone provide any idea on why UE policy Command is needed?
I am thinking we can control how data should flow using the PDU session, then what is the need for UE policy?
I see option such as match all types in UE policy traffic descriptor, which are also present in PDU session Flow description.

image

1 Like

I suppose by UE policy you mean, policy communicated and enforced by PCRF?
UE policy is used to tell UE and NW functions how to route various types of traffic.
Also UE policies are also used by Charging function to calculate bills and usage.

Thanks @ran_core_consultant.
Now I see traffic descriptor.option in UE policy.
But even PDU session we have Packet content filter… So.which one is used for what is confusing
Wouldn’t the UE just know abt where to.route traffic based on Packet content filter of PDU session?

Not really.
Each PDU gives UE list of IP ranges (TFT) and let UE know which bearer or PDU to use if destination IP is any of those ranges.
UE application and TCP stack is aware of these IP ranges.
It will send packet to destination based on APN configured in application and IP address associated with that APP.

But can I also define MAtchAlltype in PDU session and then every traffic goes throught that PDU session?

That is purpose of default bearer.
If nothing matches or misconfigured, UE will use default bearer.
I have seen issue where application couldn’t come online because application had wrong APN configured and UE was trying to send traffic via default.

Go through this blog. It explains how e2e signaling and flows work:

http://hongjoo71-e.blogspot.com/2015/07/e2e-volte-call-setup14-initial-attach.html?m=0

Remember UE will always send traffic on default, and PGW can simply dump the data if destination is unreachable.

1 Like

Got it.
So lets says I have two PDU sessions, one over IPV4 and one over IPV6.
I can define UE policy such that all default traffic goes through IPV4.
Is that correct understanding?

This is one way I can think of use case for UE policy.
Because if UE policy is not defined, traffic can go over IPV4 or IPV6, but with UE policy, we can control it.

Correct.
This is actually how some networks are implemented here.
My IMS is IPv6 and normal network is IPv4.
IMS app is configured to use IPv6 stack and do all signaling using that APN.
All other apps use default APN and use IPv4 stack.

1 Like

Thanks…I think it is becoming a bit clearer! :wink:

Continuing on this discussion:

I have a UE on which IPV4 PDU session is established.
I send it a UE policy to send all data (TrafficDescriptor set to MatchAllType) through IPV6 PDU session.
I would expect UE to now send a PDU session req IPV6 immediately.
But that doesn’t happen.
Please note no data is ongoing.
Is this behaviour correct?