Summary

  • Rogers' recovery problem extended beyond the failed routers. Engineers also lost remote access and reliable ways to contact one another.
  • A published Xona assessment and subsequent CRTC correspondence document corrective work. The question is how those controls perform, not whether an assessment exists.

A backup connection sounds like something a telecoms company sells to its customers. On 8 July 2022, Rogers needed one itself. Its network outage also disrupted the routes staff used to reach equipment and communicate with colleagues. Fixing a router is harder when the failure has taken away your access to it.

That predicament is described in the independent assessment by Xona Partners, published by CRTC in July 2024. It belongs at the centre of this story. An earlier version of this article instead said the public record lacked a named assessment or evidence of corrective work. That was wrong.

Correction, 26 October 2026: This article replaces that conclusion with the published assessment and follow-up record. It also removes the unrelated CRTC 2022-261 tariff order and corrects the claim that the mutual-assistance agreement's parties and effective date were unavailable. The original URL, publication date and membership access are retained.

The way back must survive

The failure began with a configuration change, not an excess of customers making calls. Removing a routing filter allowed more route information into the core routers than they could handle. The network carrying mobile and fixed-line traffic failed; the means of managing it were caught in the same disruption.

Xona's public, redacted report records corrective measures including routing limits, a separate management network and outside connectivity for critical sites. These changes tackle different problems. A limit helps prevent overload. Separate management access lets engineers investigate a failure that has already happened. Buying one does not make the other unnecessary.

The report also distinguishes that repair access from Rogers' larger decision to separate its mobile and fixed-line core networks. Fault containment, cost and the difficulty of operating more infrastructure all enter that decision. Two boxes on an architecture diagram are not, by themselves, a demonstration that one will survive the other's failure.

There is a record to follow

In its 4 July 2024 letter, CRTC said the measures had addressed the outage's cause, based on Xona's findings. It also required follow-up reporting. A December 2025 letter recognised progress and requested another update; the public docket lists a July 2026 filing. That is a continuing accountability record, not an empty one. It is not a warranty against every future outage, either.

For a business choosing connectivity, the practical lesson is more specific than “demand resilience”. Ask how the provider reaches broken equipment, how its incident team communicates, and which failures its backup arrangements can withstand. A recovery plan earns its name when the people using it can still get to work after the main network stops.