Hacker News Reader

Looks very similar to Facebook's outage. A combination of BGP and DNS problems that started with a configuration change.

What evidence are you seeing that makes you think this is happening in the Rogers' case? I haven't seen any diagnostics or analysis by anyone.

Rogers did just abruptly withdraw all their routes, then reannounced some, then withdrew them again. It's not dissimilar at least.

You probably saw it up thread but since rogers isn't talking this is all we really have: https://blog.cloudflare.com/cloudflares-view-of-the-rogers-c...

https://radar.cloudflare.com/asn/812?date_filter=last_24_hou... look at the IPV4 vs IPV6 traffic, when there is IPV4, there is no IPV6 and vice versa. Issues announcing IPV6 BGP prefixes and IPV4 at the same time?