Looking forward to read about the postmortem analysis of this outage, and whether this operator had single points of failure.
I wonder if it will be anything like the Rogers outage in Canada. A bad software maintenance update made their infrastructure malfunction and so many people were offline from a few days to a week.
Yeah sounds like it was something similar. CloudFlare showed spikes in BGP announcements (the Border Gateway Protocol - basically how each network node knows it's neighbours and the route through that neighbourhood) when the system went down. So Optus stopped telling the world it existed and everything updated around it, shutting it out from the internet.
It may just be a coincidence but my work voip, which is non Optus, went down for an hour two days ago. My csm software is in the cloud and uses an sms gateway, with a different company and provider. It went down twice yesterday for a few hours. I don’t recall the last time that happened. I wonder if there was some underlying telecoms issue that undermined them all at different times.
That page about Roger's outage is an interesting read. Interac deserves a (lesser) part of the blame, such a critical service shouldn't depend on a single network provider. Good practices include redundancy in different locations and backed by different network and electricity provider, or at least not share the same local power grid.
We'd thankfully just disconnected from Optus (not easy, we had a commercial fibre line) a few days before this outage. They've been getting progressively less reliable and outages had almost become routine.
Looking forward to read about the postmortem analysis of this outage, and whether this operator had single points of failure.
I wonder if it will be anything like the Rogers outage in Canada. A bad software maintenance update made their infrastructure malfunction and so many people were offline from a few days to a week.
Yeah sounds like it was something similar. CloudFlare showed spikes in BGP announcements (the Border Gateway Protocol - basically how each network node knows it's neighbours and the route through that neighbourhood) when the system went down. So Optus stopped telling the world it existed and everything updated around it, shutting it out from the internet.
It may just be a coincidence but my work voip, which is non Optus, went down for an hour two days ago. My csm software is in the cloud and uses an sms gateway, with a different company and provider. It went down twice yesterday for a few hours. I don’t recall the last time that happened. I wonder if there was some underlying telecoms issue that undermined them all at different times.
That page about Roger's outage is an interesting read. Interac deserves a (lesser) part of the blame, such a critical service shouldn't depend on a single network provider. Good practices include redundancy in different locations and backed by different network and electricity provider, or at least not share the same local power grid.
We'd thankfully just disconnected from Optus (not easy, we had a commercial fibre line) a few days before this outage. They've been getting progressively less reliable and outages had almost become routine.