top button
Flag Notify
    Connect to us
      Site Registration

Site Registration

How does MME and SGW maps the GTPv2 messages responses to each user or IMSI

+1 vote
448 views

How does MME and SGW maps the GTPv2 messages responses to each user or IMSI not all of these responses carry IMSI values or any other user identifier, for example Create Session Request has Mandatory IMSI field, but the answer don't have IMSI, so how does MME know this particular response belong to this particular request?

posted Feb 10, 2017 by anonymous

Share this question
Facebook Share Button Twitter Share Button LinkedIn Share Button

1 Answer

+1 vote

Response message come on the tunnel formed during Create Session Request handling.
So, MME, SGW, PGW keep a tunnel record with them, in which they store numerous parameters like Tunnel ID, UE Record, Peer Record, etc.
Inside UE Record, which gets allocated during Create Session Request handling, IMSI is also stored.

When Create Session Response is received, based on the TEID, the corresponding tunnel record is looked up and IMSI is fetched from UE Record inside tunnel record.

Hope this is clear.

Nodes also keep transaction record to store additional information like for which bearers request message was sent. But transaction record lives only till response is received, and deallocated afterwards. Whereas tunnel record lives till the entire call gets disconnected (through DBReq, DSReq, local purge, path failure, etc.)

answer Feb 11, 2017 by Hardik Tanna
Similar Questions
+1 vote

I recently come across advantages of SCTP over TCP and In telecommunication most of the real time interfaces require SCTP connection example eNB to MME.

But MME to SGW and SGW to PGW GTP-C is used as signaling protocol. Why SCTP not being used? What is the advantage of using GTP-C?

+2 votes

Hello,

I am working on GTPV2 control plane. I want to implement negative test scenarios for GTPV2-C messages in my project. Can anybody list out the possible negative test case scenarios for GTPV2-C messages.??

0 votes

Implementation of user plane processing at the user space of node is time consuming and cpu intensive. Is there any library/solution available through which GTP encapsulation/de-capsulation can be done at kernel level to reduce delay ?

+2 votes

GTPv2 header has reserved only 1 byte for message type field.
What could be the reason behind it ?

...