SS7 signaling link dimensioning for a national network
SS7 signaling link dimensioning determines how many signaling paths a national telecommunications network needs during its busiest operating period. The Erlang B model offers a practical starting point for estimating capacity, blocking probability and spare resources before a network is deployed.
The calculation is especially useful when planning links between service switching points, signal transfer points and international gateways. Australia adds particular planning considerations: traffic is concentrated around Sydney, Melbourne and Brisbane, while long distances to Perth, Darwin and regional communities make resilience and maintenance access just as important as raw capacity.
Why Erlang B remains useful
Erlang B estimates the probability that an arriving transaction finds every available channel occupied and is cleared without waiting. The classic formula is:
[ B(N,A)=\frac{A^N/N!}{\sum_{k=0}^{N} A^k/k!} ]
Here, (N) is the number of available circuits or equivalent service channels, (A) is offered traffic in Erlangs, and (B) is the blocking probability. A network designer normally selects a target grade of service, then solves for the required value of (N).
SS7 traffic is made up of short messages rather than long voice calls, so Erlang B is an approximation rather than a complete protocol model. It remains valuable when signaling attempts can be represented as concurrent service demand and when the design also applies a margin for bursts, retransmissions and link failures.
Establishing the offered load
The first step is measuring the busy-hour signaling rate. For a call-control service, a useful estimate is:
[ A=\lambda \times h ]
where (\lambda) is the arrival rate and (h) is the average channel occupation time in hours. For SS7, the occupation interval may represent message handling, transaction processing or the time a signaling resource is committed.
Traffic should be segmented by service. ISUP call setup, SMS delivery, mobile location updates, roaming queries and supplementary services have different message patterns. A national operator serving Melbourne’s dense business districts may see a very different busy-hour profile from a regional exchange supporting a smaller population but fewer alternative routes.
Converting SS7 traffic into link capacity
A 64 kbit/s signaling link has finite message capacity. Designers need the average message size, protocol overhead, transaction frequency and expected retransmission rate. MTP2 framing, MTP3 routing labels, SCCP, TCAP and application-layer content all contribute to the actual bit rate.
For a practical calculation, estimate messages per second, multiply by average octets per message and add a conservative overhead factor. The resulting bit rate can be compared with the usable capacity of each link. Training material from the SS7 training resource can help relate this calculation to MTP3, ISUP, timers and call-routing behaviour.
Selecting the blocking target
A target such as 0.1% blocking may be suitable for a critical signaling route, but the correct value depends on service obligations and network architecture. Emergency call handling, inter-operator traffic and mobile authentication can justify tighter limits than lower-priority application traffic.
Busy-hour figures should include abnormal but credible conditions. Australian networks must account for major sporting events, public holidays, bushfire evacuations and sudden demand around Sydney or Melbourne transport disruptions. A design based only on an annual average can appear efficient while failing during the exact periods when reliable signaling matters most.
Designing for national resilience
Capacity should be calculated for normal operation and failure operation. If a linkset contains four links, the Erlang B calculation should be repeated with three links available. The same test should be applied to a failed signaling point, a maintenance event and the loss of a transmission route.
Geographical diversity is important across Australia. Separate paths may be needed between metropolitan sites and remote facilities, particularly where fibre routes share long corridors through regional New South Wales, Queensland or Western Australia. STP pairs, alternate linksets and independent power arrangements reduce the risk that one cable cut or exchange outage isolates a large service area.
Checking the model against operations
The mathematical result should be compared with live counters from MTP2 and MTP3. Useful measures include link utilisation, congestion indications, retransmission counts, transfer prohibited messages, route-set changes and peak messages per second. Timer expiries can reveal delay or overload that a basic Erlang B result does not show.
Where networks are migrating from TDM signaling to IP transport, the mapping between legacy SS7 behaviour and SIGTRAN components must be validated carefully. The M2PA mapping guide provides relevant context for relating traditional signaling links to IP-based transport and network architecture.
| Planning item | Typical calculation input | Engineering decision |
|---|---|---|
| Busy-hour demand | Transactions per second and holding time | Offered traffic in Erlangs |
| Link capacity | Bit rate, message size and overhead | Number of active 64 kbit/s links |
| Service target | Blocking probability, such as 0.1% | Minimum linkset size |
| Failure case | One link, route or node unavailable | N+1 or larger resilience margin |
| Growth allowance | Forecast traffic and new services | Additional links or IP capacity |
| Operational validation | Peak utilisation and retransmissions | Recalibration of the model |
What the design should retain
Erlang B gives a disciplined baseline: measure the busy hour, convert signaling demand into offered load, select a blocking objective and calculate capacity for both normal and failure states. It should be paired with protocol-level analysis because SS7 congestion can result from routing errors, timer interactions or retransmissions as well as insufficient bandwidth.
For an Australian national network, the final design must balance metropolitan concentration, remote distances, shared infrastructure and service continuity. The key point to remember is that a link count is adequate only when it meets the Erlang B target during the busy hour and still carries essential signaling after the most credible failure.
