Route Diversity

Route diversity is the use of independent communication paths so that one credible failure does not remove every path needed by the service. It applies to fiber, microwave, wireless, satellite, packet transport and industrial communication networks.

The important word is independent. Two circuits from different providers are not diverse if they share the same bridge crossing, site-entry duct, patch room, tower, power plant, timing source, software configuration or maintenance crew. Route diversity is an evidence claim, not a topology drawing.

Physical Versus Logical Diversity

Logical diversity means the network has more than one configured path. Physical diversity means those paths do not share the failure mechanism being protected against. The distinction is central because many outages occur in the physical layer below the logical network map.

Examples of shared risk include:

• two fibers in one duct bank;
• two radios on one overloaded tower;
• two paths through one building entrance;
• two packet links powered by one battery plant;
• a primary route and backup route using the same upstream aggregation switch.

If one event can disable every path, the design is not diverse for that event.

Availability With Independent Paths

For two independent parallel paths, service availability can be screened as:

If both paths have:

then:

or:

This attractive number is valid only if the failure assumptions are independent enough for the service decision.

Shared-Risk Correction

If both routes pass through a shared segment with availability:

then the path-level availability is limited by that common segment:

Using the values above:

or:

The design still has two logical paths, but the common segment dominates the service risk. This is why shared-risk link groups must be mapped before availability claims are accepted.

Capacity During Failover

Route diversity also has a capacity condition. A backup path is useful only if it can carry the protected traffic during the fault. If protected load is:

and backup capacity is:

then utilization during failover is:

The protected service has spare capacity. If unfiltered traffic is:

then:

The backup path is overloaded unless traffic policy protects the service classes that matter.

Shared-Risk Evidence

Route diversity should be documented as shared-risk evidence, not as a simple yes/no field. A useful review identifies which assets are common to both paths and which hazards they create: excavation, bridge impact, building fire, tower collapse, regional power loss, software misconfiguration, maintenance error or provider outage.

The evidence normally includes geographic route records, site photos, splice or patch records, provider demarcation points, power feeds, cabinet locations, radio-path drawings and failover test logs. If any record is missing, the availability claim should state the uncertainty instead of assuming independence.

Diversity should also be rechecked after repairs, provider reroutes, cabinet moves and capacity upgrades.

Boundary With Redundancy

Redundancy means extra components or paths exist. Route diversity means those paths avoid shared failure modes relevant to the service. A network can be redundant but not diverse, diverse but under-capacity, or diverse on a map but operationally unavailable because restoration authority, spares or configuration are missing.

Route diversity also differs from wireless antenna diversity. Antenna diversity reduces fading correlation in a radio channel. Route diversity reduces common failure exposure across service paths. Both depend on independence, but the failure mechanisms and validation evidence differ.

Validation Evidence

A defensible route-diversity claim includes route drawings, duct or pole records, site-entry records, building paths, patch-panel mapping, provider handoff points, power and cooling dependencies, timing dependencies, radio weather exposure, common software dependencies, protected traffic classes, failover capacity, failover time and restoration authority.

Common mistakes include accepting provider names as proof of diversity, checking long-haul routes while ignoring the last hundred meters, proving optical loss but not physical path separation, omitting power and timing from the shared-risk map, and testing failover without realistic traffic. A strong design shows which failures are covered and which are explicitly not covered.