BSS/OSS Academy
2.615 min read

Worked Example: Connectivity Service

Worked Example: Residential Broadband

This section walks through a complete, realistic example of how a residential broadband product decomposes through every catalog layer — from the Product Offering displayed in the online shop, all the way down to the physical resources on the network. This ties together every concept from Module 2.

Follow the Chain
Pay attention to how each layer references the layer below it. This decomposition chain is exactly what the fulfilment system traverses when activating a new service. If you can follow this example, you understand catalog-driven architecture.
SuperFibre 200 HomeProduct OfferingResidential Broadband AccessProduct SpecificationCustomer-Facing ServicesInternet AccessCFSWiFi MgmtCFSCPE MgmtCFSResource-Facing ServicesGPON-BearerRFSVLAN-ServiceRFSIP-ProfileRFSQoS-ProfileRFSCloud-ControllerRFSTR069-ConfigRFSCPE-ProvisionRFSResourcesOLT PortResourceONTResourceVLAN IDResourceIPv4 AddrResourceIPv6 PrefixResourceBW PolicyResourceACS EndpointResourcePOPS

Full decomposition of SuperFibre 200 Home -- from commercial offering down to network resources

Figure 2.6 — Complete decomposition of "SuperFibre 200 Home" from offering to resources

Layer 1: Product Offering

Product Offering: SuperFibre 200 Home
Name: SuperFibre 200 Home. Price: $89/month + $0 installation (promotional). Contract: 24 months. Channel: Online, Retail, Telesales. Segment: Residential. Includes: 200Mbps internet + WiFi management app + router. References: Product Specification "Residential Broadband Access".

Layer 2: Product Specification

Product Specification: Residential Broadband Access

ComponentTypeCardinalityLinked CFS
Internet AccessMandatory1:1CFS:Internet-Access
WiFi ManagementOptional (included in this offering)0:1CFS:WiFi-Management
Router EquipmentMandatory (choice of model)1:1CFS:CPE-Management
Static IPOptional (not included)0:1CFS:Static-IP

The Product Specification defines the structure: what components exist, which are mandatory, and which CFS each component links to. The Product Offering selects which optional components are included and sets characteristic values (speed = 200Mbps).

Layer 3: Customer-Facing Services

For the "SuperFibre 200 Home" offering, three CFS instances are created (Static IP was not selected):

AttributeValue
Service TypeInternet Access
Download Speed200 Mbps
Upload Speed50 Mbps
Data QuotaUnlimited
IP VersionDual Stack (IPv4 + IPv6)
SLAStandard residential (99.5% availability)

Decomposes into RFS based on access technology at the customer's address. In this example: GPON technology.

Layer 4: Resource-Facing Services

The three CFS instances decompose into a set of RFS instances. This is where technology-specific detail appears:

RFS Decomposition

Parent CFSRFSPurpose
Internet AccessRFS:GPON-BearerConfigures the GPON connection (bearer channel, bandwidth profile)
Internet AccessRFS:VLAN-ServiceAssigns and configures customer VLAN on the access network
Internet AccessRFS:IP-ProfileConfigures DHCP, IPv4/IPv6 addressing, DNS delegation
Internet AccessRFS:QoS-ProfileApplies bandwidth policy (200/50 Mbps shaping)
WiFi ManagementRFS:Cloud-Controller-LinkRegisters CPE with cloud WiFi management platform
CPE ManagementRFS:TR069-ConfigSets up TR-069 auto-configuration server (ACS) profile for the ONT
CPE ManagementRFS:CPE-ProvisioningPre-provisions ONT with customer config (SSID, password, LAN settings)

Layer 5: Resources

Finally, each RFS requires specific physical or logical resources to be allocated from inventory:

Resource Allocation

RFSResourceTypeAllocated Value
GPON-BearerOLT PortPhysicalOLT-EAST-07, Slot 3, Port 12
GPON-BearerONT DevicePhysicalHuawei HG8245H, S/N: HWT2340098
VLAN-ServiceVLAN IDLogicalVLAN 1042 (from range 1000-1999)
IP-ProfileIPv4 AddressLogical10.42.17.22 (from DHCP pool)
IP-ProfileIPv6 PrefixLogical2001:db8:42:17::/64
QoS-ProfileBandwidth PolicyLogicalPolicy ID: BW-200-50-RES
TR069-ConfigACS EndpointLogicalacs.example.com/device/HWT2340098

The Complete Picture

Stepping back, here is what happened when the customer clicked "Order Now" on "SuperFibre 200 Home":

End-to-End Fulfilment Flow

1
Commercial Order Captured
BSS

COM creates a Product Order for "SuperFibre 200 Home" with selected options (router model, no static IP). Product Inventory creates a pending Product Instance.

2
Product → CFS Decomposition
BSS → OSS

The catalog maps the Product Specification to 3 CFS types. SOM creates Service Orders for Internet Access, WiFi Management, and CPE Management.

3
CFS → RFS Decomposition
OSS — SOM

The service catalog decomposes each CFS into RFS. Access technology = GPON determines the specific RFS set. 7 RFS instances are created.

4
Resource Allocation
OSS — ROM

ROM allocates resources from inventory: OLT port, VLAN, IP addresses, bandwidth policy. Each resource is reserved for this service.

5
Network Activation
OSS — Network

ROM drives configuration to the network: OLT port configured, VLAN created, DHCP updated, QoS policy applied, ONT configured via TR-069.

6
Service Activation Complete
OSS → BSS

Service Inventory marks all CFS instances as Active. Product Inventory marks the Product Instance as Active. Customer receives confirmation.

API Chain in This Example
This flow uses the following TM Forum APIs: TMF622 (Product Ordering) for the commercial order, TMF641 (Service Ordering) for service orders to SOM, TMF652 (Resource Ordering) for resource orders to ROM, TMF638 (Service Inventory) for service instance tracking, TMF639 (Resource Inventory) for resource allocation, TMF637 (Product Inventory) for the customer's product instance.

When Things Go Wrong: Failure Scenarios

The happy path above shows how the chain works when everything succeeds. In production, failures occur at every layer. A catalog-driven architecture handles these failures systematically because the decomposition chain is modelled and traceable — the system knows exactly what was requested, what succeeded, and what needs to be compensated.

Failure Scenarios and Recovery

Failure PointWhat HappensSystem ResponseRecovery Action
OLT Port unavailableROM attempts to allocate a GPON port on OLT-EAST-07, but all ports on the nearest OLT are in use.ROM returns a resource allocation failure to SOM. The RFS:GPON-Bearer order item fails. SOM marks the CFS:Internet-Access as blocked.ROM queries Resource Inventory for alternative OLTs in the same serving area. If an alternative exists, ROM re-attempts allocation. If none available, SOM raises a capacity exception — the order is suspended pending network build.
VLAN pool exhaustedThe VLAN range (1000-1999) for this access node is fully allocated. RFS:VLAN-Service cannot be provisioned.ROM returns allocation failure. SOM holds dependent RFS items (IP-Profile, QoS-Profile depend on VLAN assignment).Network planning is notified to extend the VLAN range or re-architect the access layer. The order remains in a "waiting for resource" state with full traceability to the blocking constraint.
Partial activation — Steps 1-4 succeed, Step 5 failsOLT port configured, VLAN assigned, IP allocated, QoS applied — but TR-069 configuration of the ONT fails (device not reachable).ROM reports partial success. SOM identifies which RFS succeeded and which failed. Service Inventory records partial activation state.SOM triggers a compensating action: retry TR-069 configuration (the ONT may not be powered on yet). After N retries, SOM can hold the order for technician dispatch. Critically, the already-allocated resources (port, VLAN, IP) are reserved, not released — they belong to this service instance.
Commercial order cancelled after decompositionCustomer cancels the order after SOM has created service orders and ROM has allocated resources, but before network activation completes.COM sends a cancellation to SOM. SOM triggers compensating service orders (cancel each CFS). ROM releases allocated resources back to their pools.Each layer reverses its work in the correct dependency order: release resources → cancel RFS → cancel CFS → cancel Product Instance. The catalog-driven decomposition chain provides the exact reversal sequence.
Why Catalog-Driven Failure Handling Works
In a legacy environment without modelled decomposition, partial failures create orphaned resources, inconsistent states, and manual cleanup. In a catalog-driven system, the decomposition chain works in both directions: forward for activation, reverse for rollback. Every resource, service, and product instance is traceable, so the system knows exactly what to compensate. This is not a theoretical benefit — it is the difference between a 10-minute automated recovery and a 3-day manual investigation.

Modification Path: Speed Upgrade

What happens when this customer upgrades from SuperFibre 200 to SuperFibre 500? In a catalog-driven architecture, the system calculates the delta between the current state and the target state, and only modifies what needs to change.

Upgrade Impact Analysis: 200 Mbps → 500 Mbps

LayerEntityChange Required?Action
Product OfferingSuperFibre 200 Home → SuperFibre 500 HomeYesProduct Inventory updates the product instance to reference the new offering. Pricing changes from $89 to $119/month.
Product SpecificationResidential Broadband AccessNo — same specThe specification is unchanged. Only the downloadSpeed characteristic value changes from 200 to 500.
CFS: Internet AccessSpeed characteristicYes — characteristic updateSOM issues a modify service order for the existing CFS instance. Download speed updated from 200 to 500 Mbps.
CFS: WiFi ManagementNo changeNoWiFi management is unaffected by a speed change. No service order generated.
CFS: CPE ManagementNo changeNoThe same ONT supports 500 Mbps. No hardware swap needed. No service order generated.
RFS: QoS-ProfileBandwidth policyYes — policy updateROM issues a modify resource order. Bandwidth policy updated from BW-200-50-RES to BW-500-125-RES.
RFS: GPON-BearerBearer configurationPossiblyIf the bearer profile needs adjustment for higher speeds, ROM modifies the bearer. If the existing bearer already supports 500 Mbps, no change needed.
RFS: VLAN, IP, TR069, Cloud-ControllerNo changeNoThese resources are unaffected by a speed upgrade. The customer keeps the same VLAN, IP address, and device management.

The catalog-driven system identified that a speed upgrade only requires changes to 2 out of 7 RFS and 1 out of 3 CFS. The OLT port, VLAN, IP address, ONT device, and cloud controller link are all unchanged. This precision is impossible without a modelled decomposition chain — in a legacy system, a speed upgrade might trigger a full re-provision, disrupting the customer's service unnecessarily.

Key Takeaways

  • A single product offering ("SuperFibre 200 Home") decomposes into 3 CFS, 7 RFS, and 7+ resources
  • Each layer adds more technical specificity while the layer above stays technology-neutral
  • The catalog defines all decomposition rules — fulfilment systems execute them at order time
  • Resource allocation (finding free ports, VLANs, IPs) happens at the bottom of the chain
  • The full chain is traceable: any resource can be traced back to the product that requires it
  • Failure handling is systematic — the decomposition chain provides exact rollback sequencing and resource compensation
  • Modifications use delta analysis: only entities that actually change are updated, minimising disruption and cost
  • This traceability is essential for impact analysis, troubleshooting, and lifecycle management