Traditional thinking assumes telecommunications needs a central number database. One authoritative source of truth. One place to query. This assumption has blocked progress for decades because building such a database is prohibitively expensive.

The key insight: all the data already exists, distributed across every Communication Provider's database. Each CP knows which numbers they host, where calls should be routed, and customer details. We don't need to duplicate this into a central location.

Computer science has solved this problem. Distributed databases have powered the internet for decades. Google, Facebook, Amazon all use distributed systems at massive scale. The telecom industry can apply these same proven techniques.

PSTN2 uses eventual consistency. Changes don't propagate instantly to all systems. Instead, information spreads gradually as CPs query each other and cache results. Within minutes to hours, everyone has consistent information. This is fast enough for telephony.

When CP1 queries CP2 about a number, CP1 caches the result. Next time that number is called, CP1 uses the cached information without querying again. Cache entries have time-to-live values, typically a few hours. This provides speed while ensuring freshness.

When number porting occurs, the previous CP can proactively invalidate cached entries by sending update notifications. This accelerates convergence. Even without notifications, cached entries expire naturally and get refreshed on next use.

CPs can optionally maintain directory servers that aggregate cached information. These serve as secondary lookup sources. But they're not authoritative, just helpful caches. The authoritative source remains the CP that actually hosts the number.

If CPs disagree about number ownership, the system uses timestamps to resolve conflicts. The most recent authoritative update wins. Porting records include timestamps precisely for this reason. Mathematical consistency is maintained without central coordination.

Because there's no central database, there's no single point of failure. If one CP goes down, others continue operating. If a directory server fails, CPs query authoritative sources directly. The system is inherently resilient and self-healing.

Distributed systems scale horizontally. As more CPs join, capacity increases proportionally. There's no central bottleneck. No single database to overload. The system's capacity grows with the number of participants.

Despite being distributed, the system achieves global consistency. Every CP eventually learns about every number. Updates propagate through the network. Information converges to a consistent state. This happens automatically without central coordination.

Building a central database might cost 5 billion pounds. Operating it would cost hundreds of millions annually. PSTN2's distributed approach costs nothing centrally. CPs use their existing infrastructure. The savings are dramatic and enable rapid deployment.