Telecom Research Uses Message Fallbacks to Spot Network Instability

RCR Wireless News reports that independent telecom researcher Srinivasa Chary Maringanti is developing AI models that treat message and call fallbacks as signals of network instability. The models examine events such as an RCS message switching to SMS or a VoLTE call moving to an older system.

These fallbacks often go unnoticed because the message or call still works. Maringanti is working to map RCS events directly to SMS and MMS fallbacks, then combine them with signaling and transport data. Patterns across several services could indicate congestion or other localized problems that existing monitoring systems miss.

The research also applies geospatial indexing to divide areas into hexagonal cells. AI models can analyze these smaller areas instead of relying on city or market averages, which may hide issues affecting a building, street, or neighborhood.

Maringanti said operators must first connect data from messaging, voice, network, and application systems. His proposed approach uses common identifiers to link services within each geographic cell, with machine learning processing large data sets and large language models interpreting patterns across systems.