Back to the archive

The encyclopedia · Engineering & Operations · Technical decision · 2022

A routine upgrade at Rogers took down Canada's networks for 19 hours — 911 went with it

On July 8, 2022 a routine upgrade at Rogers wiped out phone and internet for 12 million Canadians for 19 hours — taking Interac payments and 911 down with it.

Rogers Communications · 2022-07

What happened

On July 8, 2022, at about 4:30 a.m. EDT, Rogers Communications' network began failing across Canada. By 4:58 a.m. the core network had crashed: more than 12 million customers lost wireless and wireline service — mobile, home internet, corporate and institutional lines — and about a quarter of Canada's observable internet connectivity disappeared. Most services were restored within 15 to 19 hours, but full restoration took until 7:00 a.m. on July 9, and some customers stayed unconnected for days.

The independent assessment the CRTC commissioned, released in July 2024, traced the failure to a routine change. During phase six of a seven-phase upgrade of its IP core, Rogers staff removed the Access Control List filter from the distribution routers' configuration. With the filter gone, one distribution router released more than 900,000 route advertisements into the core — against roughly 10,000 normally — and the core crashed within minutes. A risk-assessment algorithm downgraded the phase to 'Low' risk after five clean phases, which the CRTC said contravenes industry norms.

The redundancy that should have saved Rogers was not independent. Wireless and wireline ran on one common IP core, so when the core failed, both went down together. Worse, Rogers' management network ran on that same core: staff at the network operations centre lost remote access, had no backup connectivity from an alternative provider, and could not read the failed routers' error logs — it took about 14 hours to pinpoint the cause.

The outage rippled through the economy: Interac debit and e-Transfer failed nationwide, banks reported ATM and online problems, Tim Hortons went cash-only, and police in Toronto and Ottawa said some callers could not reach 911. Rogers estimated it would refund about C$150 million — five days of credit for its roughly 13 million customers — and shares fell 4.61% on the first trading day after. The CRTC opened an examination, and the outage injected new regulatory risk into Rogers' pending Shaw acquisition.

Why it happened

  • An algorithm downgraded phase six of the core upgrade to 'Low' risk after five clean phases — the CRTC called that a contravention of industry norms, and the filter was removed on that basis.
  • No overload protection: with the Access Control List filter gone, one distribution router flooded more than 900,000 routes into a core built for 10,000, and the core crashed within minutes.
  • Wireless and wireline shared one IP core, and the management network rode on the network it managed — so Rogers' own engineers lost the tools to fix the outage for about 14 hours.
  • The network operations centre had no backup connectivity from an alternative provider — redundancy that exists only on paper is the same as none.
What it costC$150M in credits, 4.6% stock drop, Interac and 911 downcostly

The lesson

One routine change took down a national network: no overload protection on the core, and a management network that rode on the network it managed. Redundancy helps only when the backup is independent.

Aftermath

The CRTC opened an examination on July 12, 2022, and commissioned an independent assessment by Xona Partners, published in July 2024. Rogers committed C$10 billion to network investment and to separating its wireless and wireline systems — the network split alone cost $261 million. The outage also injected regulatory risk into Rogers' pending Shaw acquisition, which was approved in March 2023 with 21 conditions. Rogers told the CRTC it had fully implemented every Xona-recommended measure by mid-2026.

Sources

spotted an error? The club wants to know.

Comments · 0

    Sign in to join the comments.

    More like this

    Somewhere, someone solved the problem this company failed at. 2nd Opinion →