CraftWorld · Transport Domain

Goal-driven optical
& IP backbone.

Five-nines is a goal, not a hope. Declare backbone availability, latency and protection targets across DWDM, OTN and IP/MPLS — and let Ritam (our AIOps engine) hold them, pre-empting faults before the span drops.

Optical and IP as one layer — a physical-layer event and its routing impact are one story, not two tickets.

DWDM / OTNIP / MPLSBGPProtection
Availability · LiveHOLDING
6 pre-emptive re-routes / qtr1 span at risk · MUM-PUN-04
The problem

The backbone fails silently, then all at once.

Optical and IP are managed in separate NMS by separate teams. Degradation is invisible until a span drops or an LSP congests — and by then the SLA is already breached across every service riding that path.

Reactive
protection
You learn a span is failing when protection switches — or doesn't.
Optical ≠ IP
no correlation
An optical fade and a BGP flap look like two unrelated tickets.
Manual
capacity planning
Exhaustion discovered when a trunk congests, not 90 days ahead.
Hours
to localise
Which of hundreds of spans? Which LSP? Investigated by hand.
Availability is the goal; reactive protection can’t guarantee it. A five-nines backbone must see degradation coming and re-route before the fault — across optical and IP as one layer.
The thesis

The backbone, managed by goals — not protection alarms.

You declare the availability, latency and protection goals the backbone must hold. Ritam correlates optical and IP as one layer, predicts degradation, and re-routes before the span drops — so the SLA never dips.

Alarm-driven — reactive
Goal-driven — pre-emptive
Protection switches after a span fails
OSNR slope detected 45–90 min before the fault
Optical and IP faults chased separately
One cross-layer causal graph — optical → LSP → service
Capacity discovered when a trunk congests
Exhaustion forecast 90 days ahead, CapEx pre-planned
A BGP hijack noticed after traffic is impacted
Anomalous origin auto-contained in <3 min
“With you in command.” Pre-planned FRR and re-routes execute autonomously to hold the SLA; capacity and topology changes are approved.
The scope

Optical and IP — one backbone, one control plane.

CraftWorld manages the full transport stack across every major vendor, correlating the optical layer with the IP/MPLS layer that rides on it — so a physical-layer event and its routing impact are one story.

LayerCraftWorld monitorsRitam controlsSLA / goal
DWDM channelsPer-channel OSNR · optical powerROADM express-port healthchannel OSNR
OTNOTN trail · ODU continuityProtection switchingtrail protection
IP / MPLSLSP continuity · RSVP-TE · FRRLSP re-route, TE re-balancelatency SLA
BGPPeer & prefix · route-leak detectRoute-filter, origin containment0 route-leaks
CapacityPer-LSP util · TE headroomDemand forecast, CapEx trigger> 20% headroom

An OSNR fade on a Nokia span and a BGP flap on a Cisco edge on the same physical route resolve to one root cause — in under 60 seconds.

The backbone, as Ritam sees it

Optical and IP — on one correlated map.

A cross-layer view of the backbone — every span's optical health and the LSPs that ride it. When a span's OSNR slopes, Ritam pre-positions traffic on the protected path before the fault.

Span MUM-PUN-04 · Live

OSNR sloping

OSNRtrending down — outside baseline
Predicted faultOTN LOF/LOS in ~60 min
Protected pathvalidated — OSNR healthy
Ritamre-route queued for approval
The capstone

Declare the backbone goals. Hold five-nines.

The transport goals you own — availability, latency, protection and capacity — tracked continuously, with Ritam flagging and pre-empting anything drifting toward breach.

GoalTargetNowStatus
Backbone link availability99.999%99.999%On track
Latency SLA — core wavelengths< 8 ms6.2 msOn track
Unprotected spans01At risk
TE-trunk capacity headroom> 20%23%On track
BGP route-leak events00On track
99.999%
Backbone availability
target met · 5.2 min downtime / year
1 goal at risk — an OSNR slope on span MUM-PUN-04; Ritam has traffic pre-positioned and a re-route queued for approval. 6 pre-emptive re-routes this quarter.
Worked example · optical

A span fails in 60 minutes — Ritam acts in 10.

TSLAM4b learns each span's optical fingerprint. When OSNR slopes, it knows a fault is coming — and pre-positions traffic before the customer ever sees it. The availability goal never dips.

T–60 min

TSLAM4b detects an OSNR slope-change on span MUM-PUN-04 — outside its learned baseline.

T–58 min

ARGUS confirms the fault trajectory; predicts OTN LOF/LOS within ~60 min.

T–55 min

Ritam pre-positions traffic on the protected ROADM path; validates the new-path OSNR.

T–52 min

Re-route executed on approval, ticket auto-raised — the span drops later to zero customer impact.

45–90 min
Prediction lead before the optical fault impacts service.
< 10 min
Ticket closed — traffic moved before the drop.
99.999%
Availability — never dipped below five-nines.
The span still failed — but traffic had already moved, so backbone availability never dropped below five-nines.
How it works

Autonomous across optical and IP.

Ritam operates the backbone end-to-end — DWDM/OTN optical, IP/MPLS routing and BGP — with autonomous, standards-based actions on approval.

Optical — DWDM / OTN

Per-channel OSNR & optical power
Amplifier / EDFA gain, ROADM health
OTN trail, ODU continuity, protection
Ritam: protection switch, path re-route

IP / MPLS / BGP

LSP continuity, RSVP-TE, FRR validation
BGP peer & prefix, route-leak detection
Per-LSP utilisation, TE headroom
Ritam: LSP re-route, route-filter

Assurance & OAM

BFD / LSP-Ping, protection-switch validation
G.709 OTN overhead, TCM
Cross-layer correlation, optical → IP
Ritam: pre-position, capacity forecast
Multi-vendor

Native across every transport vendor.

One control plane over a multi-vendor backbone — no middleware. Every integration is production-validated and correlates optical and IP regardless of which vendor sits at each layer.

Optical / DWDM / OTN

Ciena · Infinera · Nokia · Fujitsu · Huawei OptiX · Lumentum · ADTRAN · Ribbon

IP / MPLS / Routing

Cisco IOS-XR · Juniper JUNOS · Nokia SR OS · Huawei VRP · Arista EOS

Assurance & OAM

BFD / LSP-Ping · RSVP-TE · G.709 OTN overhead · TCM · protection-switch validation

Interfaces & standards
NETCONF / YANG · RESTCONF · SNMP v2c / v3 · gRPC / gNMI streaming · OpenConfig · syslog · TL1 (optical) · BGP-LS · OTN G.709
Autonomous execution — LSPs re-routed, ROADM paths switched, route filters applied — all fully audited, on approval.
Product tour

See CraftWorld Transport in action.

The optical and IP backbone as one layer — span health, OSNR trend and the LSPs riding it.

CraftWorld Transport — console screenshot
The measurable delta

What goal-driven transport delivers.

From reactive protection to a pre-emptive, five-nines backbone — measurable from the first quarter.

99.999%
Backbone availability
held as a goal, not hoped for
45–90 min
Optical fault lead
OSNR slope detected before service impact
< 3 min
BGP hijack containment
anomalous origin auto-filtered
90 days
Capacity lead time
exhaustion forecast, CapEx pre-planned
< 60 s
Cross-layer RCA
optical → IP → service, one root cause
< 15 min
Target MTTR
autonomous re-route vs 5–8 hrs legacy
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 backbone console.

Optical + IP correlated topology, OSNR sloping on a span, re-route queued.

The backbone console
See it on your estate

See Ritam hold five-nines.

A performance-guaranteed PoC on your optical & IP backbone — baselined, KPI-gated, reversible.