Brother
Professional
- Messages
- 2,590
- Reaction score
- 533
- Points
- 113
Recently, the work of many of the major Internet services occurred failure, he touched upon, in particular, the cloud service Amazon, Reddit site strimingovuyu platform Twitch, Discord and messenger service Downdetector, which itself keeps track of failures on the Internet. The crash occurred at about 2 p.m. Moscow time (7 a.m. in the eastern United States) and lasted for about two hours.
The problem was initially associated with the large content delivery service Cloudflare, which is used by millions of sites. The company said it had lost about 15% of its global traffic, but its own systems were working fine: traffic did not reach it due to other people's errors - including Verizon, one of the largest telecommunications companies in the United States. Cloudflare posted on her blog a long post about the cause of the failure (and heavily criticized Verizon), in which she described the incident with the words "the Internet has had a small heart attack." Pennsylvania ISP DQE Communications configured incorrectly routing about 2% of the world's Internet through its own network and the network of a client - the metallurgical company Allegheny Technologies. The source of the error was in the configuration of the BGP optimizer, the main dynamic routing protocol of the modern Internet.
For some reason, the incorrect configuration was accepted and transferred to the whole world by the owner of the DQE gateway to the backbone Internet - Verizon. Allegheny is also a Verizon customer, and this may have triggered the error. Traffic destined for the internet giants went through a low-power network. This triggered the cascading denials of service that most affected customers of Cloudflare, a secure cloud host that includes the Overcast podcast app and the Discord messenger. Many hosted sites were unavailable for several hours, at the peak the hosting lost 10% of its traffic.
Cloudflare chief executive Matthew Prince said: “The employees at Verizon and Noction [the BGP optimizer developer] should be deeply ashamed of their negligence affecting Cloudflare and much of the Internet. That BGP is so unstable is absurd. Even more absurd is that Verizon is accepting routing without the simplest of filters."
The incident showed that the Internet protocol, which most users are unaware of, can directly affect their lives - and that it is very easy to use BGP for disruptive purposes. A similar routing error was the cause of the unprecedented crash WhatsApp, Instagram and Facebook in March, and Google crash in November 2018 caused a small Nigerian provider.
The technical director of RosKomSvoboda Stanislav Shakirov sees in the incident a demonstration of how the future "sovereign Runet" can work:
“After each such incident, I remind you: the Internet was not created for an untrusted environment in which there are woodpeckers,” writes Telegram-channel addmeto. - Several times a year a "woodpecker" arrives, which destroys the routing of the global network with several blows, this is how we are not protected. If you think there is an easy way to isolate yourself from this - no, until no one has invented it."
The problem was initially associated with the large content delivery service Cloudflare, which is used by millions of sites. The company said it had lost about 15% of its global traffic, but its own systems were working fine: traffic did not reach it due to other people's errors - including Verizon, one of the largest telecommunications companies in the United States. Cloudflare posted on her blog a long post about the cause of the failure (and heavily criticized Verizon), in which she described the incident with the words "the Internet has had a small heart attack." Pennsylvania ISP DQE Communications configured incorrectly routing about 2% of the world's Internet through its own network and the network of a client - the metallurgical company Allegheny Technologies. The source of the error was in the configuration of the BGP optimizer, the main dynamic routing protocol of the modern Internet.
For some reason, the incorrect configuration was accepted and transferred to the whole world by the owner of the DQE gateway to the backbone Internet - Verizon. Allegheny is also a Verizon customer, and this may have triggered the error. Traffic destined for the internet giants went through a low-power network. This triggered the cascading denials of service that most affected customers of Cloudflare, a secure cloud host that includes the Overcast podcast app and the Discord messenger. Many hosted sites were unavailable for several hours, at the peak the hosting lost 10% of its traffic.
Cloudflare chief executive Matthew Prince said: “The employees at Verizon and Noction [the BGP optimizer developer] should be deeply ashamed of their negligence affecting Cloudflare and much of the Internet. That BGP is so unstable is absurd. Even more absurd is that Verizon is accepting routing without the simplest of filters."
The incident showed that the Internet protocol, which most users are unaware of, can directly affect their lives - and that it is very easy to use BGP for disruptive purposes. A similar routing error was the cause of the unprecedented crash WhatsApp, Instagram and Facebook in March, and Google crash in November 2018 caused a small Nigerian provider.
The technical director of RosKomSvoboda Stanislav Shakirov sees in the incident a demonstration of how the future "sovereign Runet" can work:
“After each such incident, I remind you: the Internet was not created for an untrusted environment in which there are woodpeckers,” writes Telegram-channel addmeto. - Several times a year a "woodpecker" arrives, which destroys the routing of the global network with several blows, this is how we are not protected. If you think there is an easy way to isolate yourself from this - no, until no one has invented it."
