When the network is unavailable
Offline communication and local coordination: how information can move when internet and mobile service are unavailable.
Research / David Bentley
Design for the conditions.
Keep the record.
Reliable systems need more than a successful demonstration. They need clear limits, people who can operate them, and a record of what actually happened.
An English introduction to my research alongside commercial engineering.
The programme and my role
El Hilo que Persiste
I am a collaborating technical researcher on El Hilo que Persiste, a programme working with the Laboratory of Algorithms and Data Analysis (LAAD) at UNSAAC in Cusco, Peru.
The work concerns communication and record-keeping where connectivity is unreliable. Academic and community roles are documented on the programme website.
Two research lines
The message. The record.
When information must remain trustworthy
Local custody, tamper-evident records and later synchronisation: preserving information without taking control away from the people who produce it.
Field example / 26 July 2026
A demonstration in Pacchanta
The programme’s public field record describes an offline communication prototype demonstrated with the community and five UNSAAC students at around 4,300 metres.
A message reached the other endpoint and a bell sounded. That established a point-to-point demonstration, not reliable coverage across a whole corridor.
The record also reports unreliable node timestamps and unanswered questions about overnight connectivity and whether the bell can be heard from homes. The equipment remains under community control.
Status: research prototype. Demonstrated, not operationally validated. It is not an early-warning system or an emergency service.
Read the programme’s field recordWhat I bring into commercial work
Make failure visible.
Make responsibility clear.
A business integration and an offline research prototype are different systems. The practical questions I bring to both are straightforward:
- What happens when a connection or dependency fails?
- What was observed, and what are we only assuming?
- Who controls the information and approves the next action?
- Can another person understand and maintain the work?
These questions inform how I design client-owned workflows. Research participation is not certification of a commercial service, and a prototype demonstration is not a client outcome.
Read further
This page is a short introduction, not a translation of the full papers. The original publications and their individual status remain the reference.
- Programme, research lines and publications Spanish and English programme text.
- Public field record The demonstration, its limits and the link to the original report.
- David Bentley’s ORCID research record Supporting author and publication reference.
Have a business workflow
that needs this care?
You do not need a research project or a technical brief. Describe what should happen, which tools are involved and where the work gets stuck.