CraftWorld · Mobility Domain

Goal-driven RAN,
core & subscriber.

Experience is the SLA. Declare session success, call quality and slice targets across 5G core, RAN and subscriber edge — and let Ritam (our AIOps engine) hold them, pre-empting congestion before the customer feels it.

Core and RAN as one causal graph — UPF → gNB → session — not three separate tickets.

5G / 4G CoreRANMassive MIMOSlicing
Session Success · LiveHOLDING
14 congestion pre-empts / wk1 cell at risk · gNB-05
The problem

The subscriber feels it before the NOC sees it.

Core, RAN and subscriber KPIs live in separate tools. Congestion, dropped sessions and slice-SLA breaches surface as customer complaints and churn — not as a signal the network acts on in time.

Complaints
as the alarm
Dropped calls and slow data reported by users, not detected.
Core ≠ RAN
no correlation
A UPF issue and a cell degradation look like two unrelated tickets.
Peak-hour
congestion
Cells saturate at busy hour; SON reacts only after QoE has dropped.
Slice SLA
unverified
Per-tenant isolation and SLA assumed, not continuously proven.
Experience is the goal; reactive SON can’t guarantee it. Subscriber-experience SLAs must be predicted and pre-empted — across core, RAN and edge as one, before the customer churns.
The thesis

Mobility, managed by experience goals — not KPIs.

You declare the subscriber-experience goals the network must hold. Ritam correlates core, RAN and edge, predicts congestion and quality drops, and acts — SON, PRB and slice adjustments — before the customer ever notices.

KPI-driven — reactive
Goal-driven — pre-emptive
SON reacts after QoE has already dropped
Congestion predicted; PRBs re-allocated before busy hour
Core and RAN faults chased separately
One causal graph — UPF → gNB → session, in <60s
Slice SLA assumed, checked at audit
Per-slice SLA verified continuously, breach pre-empted
Beam failures degrade cells silently
Beam failure detected, fallback beam auto-activated
“With you in command.” SON and PRB re-allocations execute autonomously to hold QoE; slice re-provisioning and capacity change are approved.
The scope

Five layers, one control plane — each with its SLA.

CraftWorld covers the full mobility stack from the 5G core to the subscriber's device — monitoring and controlling every layer, and holding a declared experience SLA at each.

LayerCraftWorld monitorsRitam controlsSLA / goal
SlicingPer-slice KPI · NSSF · isolationSlice reassignment, SLA enforceper-DNN SLA
5G / 4G CoreAMF/SMF/UPF · UDM · IMSNF auto-scale, pod restoresession > 99.9%
RANgNB/eNB · RRC · BLER · PRBSON, power, PRB re-allocationcell > 99.95%
Massive MIMOBeam mgmt · beam-failure · SINRFallback beam auto-activationbeam recovery
CPE / EdgeSignal · bearer · per-sub SLAAuto-reprovision, bearer resetper-sub SLA

A UPF session failure, the gNB it rode and the subscriber bearer all resolve to one root cause in <60 sec — core, RAN and edge correlated.

The RAN, as Ritam sees it

Coverage, load and experience — on one live map.

A correlated view of the RAN — each cell's load, quality and the core session it serves. When a cell approaches congestion at busy hour, Ritam sees it early and offloads to neighbours before QoE drops.

Cell gNB-05 · Live

Congestion building

PRB util.88% ▲ — busy hour in ~25 min
Session success99.6% — trending down
5G sliceeMBB SLA holding
RitamPRB offload to gNB-06 staged
The capstone

Declare the experience goals. Own the QoE.

The mobility goals you own — session success, call quality, cell availability and slice SLA — tracked continuously, with Ritam pre-empting anything drifting toward breach.

GoalTargetNowStatus
5G session-setup success> 99.9%99.94%On track
VoLTE / VoNR call quality (MOS)> 4.04.2On track
Peak-hour cell availability> 99.95%99.91%At risk
Data-session setup time< 100 ms78 msOn track
Slice SLA compliance (per tenant)100%97%At risk
99.94%
5G session-setup success
target > 99.9% · held through busy hour
1 goal at risk — cell gNB-05 approaching busy-hour peak; Ritam has a PRB offload to gNB-06 staged, 14 congestion pre-empts this week.
Worked example · RAN

Busy hour hits in 25 minutes — Ritam acts now.

TSLAM4b learns each cell's daily load pattern. It sees congestion building before the peak and re-allocates capacity autonomously — so session success and QoE hold through the busy hour.

T–25 min

TSLAM4b sees gNB-05 PRB utilisation trending to saturation ahead of busy hour.

T–23 min

ARGUS confirms: session-setup success will breach 99.9% within ~20 min if unaddressed.

T–22 min

Ritam triggers SON — adjusts tilt/power, re-allocates PRBs, offloads to gNB-06.

T–20 min

Executed on approval — busy hour passes with session success held at 99.94%.

20–30 min
Prediction lead before congestion breaches the QoE goal.
99.94%
Session success — the goal held through the peak.
0
Subscribers who saw a dropped or degraded session.
The cell still peaked — but capacity had already shifted, so no subscriber ever felt it.
How it works

Autonomous across the 5G stack.

Ritam operates the mobility stack end-to-end — 5G core network functions, RAN radio management, and the subscriber edge — with autonomous, 3GPP-compliant actions on approval.

5G / 4G Core

AMF/SMF/UPF session lifecycle & QoS
UDM/HSS consistency, IMS / VoNR quality
Per-slice NSSF state & isolation
Ritam: NF pod restore, UPF auto-scale

RAN — gNB / eNB

RRC setup, PDCP throughput, BLER, PRB
Massive MIMO beam mgmt & recovery
ORAN O-DU/O-RU health, fronthaul timing
Ritam: SON, power, PRB re-allocation

Subscriber Edge

CPE / device signal, RSRP/RSRQ, SINR
Bearer drop, QoS marking, APN health
Per-subscriber & per-account SLA
Ritam: auto-reprovision, bearer reset
Multi-vendor

Native across every mobility vendor.

One control plane over a multi-vendor mobile network — public and private 5G — correlating core, RAN and edge regardless of which vendor sits at each layer.

5G / 4G RAN

Nokia AirScale · Ericsson · Samsung · Huawei AAU · CommScope · Baicells

Core (public)

Nokia · Ericsson Cloud Core · Huawei 5GC · Samsung · Mavenir

Core (private 5G)

Open5GS · Athonet · Druid · OAI · Free5GC · Mavenir

Interfaces & standards
3GPP northbound APIs · O-RAN (O1 / A1 / E2) · NETCONF / YANG · gRPC / gNMI · SNMP · ONAP / ETSI MANO · NWDAF analytics · 3GPP Rel.17 alarms
Autonomous execution — SON actions, PRB re-allocation, NF scaling, slice re-provisioning — all 3GPP-compliant, on approval.
Product tour

See CraftWorld Mobility in action.

The RAN and core, correlated — cell load, session success and slice SLA on one live view.

CraftWorld Mobility — console screenshot
The measurable delta

What goal-driven mobility delivers.

From reactive SON to pre-emptive, experience-first operations — measurable from the first quarter.

> 99.9%
5G session success
held through busy hour, not just off-peak
20–30 min
Congestion lead time
cells re-balanced before QoE drops
Continuous
Slice SLA assurance
per-tenant SLA proven, not assumed
< 2 min
Industry-4.0 recovery
OT-critical slice restored on breach
< 60 s
Core→RAN RCA
UPF → gNB → session, one root cause
43%
MTTR reduction
Ritam production reference at Tier-1 (ARKNOC)
CraftWorld · the carrier suite

Three domains, one autonomous engine.

Mobility, transport and access run the same goal-driven model on one engine — Ritam, the carrier hub. Prove one, extend across the network.

See it live

The RAN console.

Live cell grid with gNB-05 congested, PRB offload staged, session-success goal tracking.

The RAN console
See it on your estate

See Ritam hold session success.

A performance-guaranteed PoC on your core & RAN — baselined, KPI-gated, reversible.