Training / Course 04
Aims — what you can do after
- Read initial Filter Criteria and predict which Application Server a request hits — and what happens when it is down.
- Trace the full Cx command set, including the two the base course skipped.
- Distinguish re-INVITE from UPDATE, and explain the QoS-preconditions mechanism.
- Execute the MMTEL services: hold/resume, call waiting, diversion, barring and identity.
- Follow transfer (REFER) and conference call-flows.
- Route an emergency call: emergency registration, the service URN, and PSAP selection.
- Explain terminating access-domain selection, EPS fallback and single-radio continuity.
- Resolve a called number the way a network does, and understand interconnect topology hiding.
- Compute voice quality from captures, and correlate charging across nodes.
The idea that unlocks the course
Application Server (MMTEL, forwarding, ...) ^ | | ISC (SIP) | Sh (to the HSS) | v INVITE --> S-CSCF --[ evaluate filter criteria in priority order ]--> onward # matched a rule? hand the request to that AS, then continue
The service layer: CSCFs route, Application Servers implement — filter criteria connect them.
Flow — services on both sides of a call
UE-A S-CSCF(orig) AS(orig) S-CSCF(term) AS(term) UE-B | | | | | | |--INVITE-->|--filter------>| | | | originating services | |<--(changed)---| | | | | |------------- INVITE -------->|--filter------>| | terminating services | | | |<--(changed)---| | | | | |------------ INVITE ------>| delivered # a service that 'works from my phone but not for customers' tested only one side
One call crosses the service layer twice — originating and terminating.
Special path — emergency calls
NORMAL: UE -> P-CSCF -> S-CSCF -> (filters -> AS) -> callee EMERGENCY: UE -> P-CSCF -> E-CSCF -> (location) -> PSAP # forwarding, barring and identity-restriction must NOT apply to an emergency call
An emergency call takes a different path — it skips the whole service layer on purpose.
The MMTEL suite you exercise
Roughly half the course is offline reference call-flows — 20 annotated ladders that render with no lab, so it covers elements a single-VM lab cannot host.
Module breakdown
0
Bootstrap & the call-flow renderer20 min
The tooling that draws every ladder — pure text, no lab required.
1
The service layer: filter criteria & Application Servers35 min
How an IMS network gets services at all.
2
Registration deep dive35 min
The reg-event package, network-initiated de-registration, restoration.
3
Session modification35 min
Reliable provisional responses, UPDATE, QoS preconditions, early media.
4
MMTEL I — hold & call waiting30 min
Hold as an SDP direction change, driven live.
5
MMTEL II — diversion & barring35 min
The forwarding family, loop prevention, and barring codes.
6
MMTEL III — identity25 min
Caller-ID presentation and restriction, asserted identity, privacy.
7
Transfer & conference30 min
REFER, dialog replacement, the conference focus.
8
Emergency calls30 min
The service URN, the emergency CSCF, location and callback.
9
Continuity30 min
Access-domain selection, EPS fallback, single-radio voice-call continuity.
10
Interconnect30 min
Number-to-URI resolution, the border gateway, number portability.
11
Advanced media35 min
Codec negotiation, DTMF, receiver reports, a quality score.
12
Charging25 min
Offline vs online, and the identifier that ties a call together.
13
Resilience & the clinic40 min
Overload control and multi-layer fault localization.
What mastery of the service layer looks like
- Predict an outage from a single configuration field before a maintenance window.
- Read a forwarded / transferred call from its history headers.
- Route an emergency call on paper, and know which services must be switched off.