What is the history of the Internet's world wide initial disaster occurrence incident?

ByFurtherfield Gallery

On October 27, 1980, the world's first wide area fault occurred on the Internet, and it seems that the connection was blocked for 4 hours on a global scale. At the time it was operated "ARPANETIt is a packet communication network called the world's first wide area disaster occurred in IT mediaMotherboardIt is open to the public.

Happy Anniversary to the Early Internet's First Network - Wide Crash | Motherboard
http://motherboard.vice.com/read/happy-anniversary-to-the-early-internets-first-network-wide-crash

The underlying cause of the world's first wide-area failure occurred in the networkInterface message processor(IMP). IMP is a device that switches processing of network traffic, and the basic mechanism is similar to that of modern routers. In short, ARPANET's IMP had the role of rewriting the communication exchanged between the PC and the network to the ARPANET standard.

ByAndrew "FastLizard 4" Adams

When a failure occurs, one of IMP's networks "IMP 29There was a problem caused by hardware failure. IMP 29 should have passed a status message with time stamp to another node, IMP 50. However, due to a failure, an unstable status message containing a problematic timestamp was delivered to IMP 50 is.

All the nodes including the IMP 29 and the IMP 50 have the role of sending the status message every other minute but despite receiving the unstable status message by the IMP 50, all the nodes have the highest priority and process the status messages As a result, all the nodes repeatedly send and receive the status message of the problem is caused.

ByBrian J. Matis

Because all nodes are processing the unstable status messages received by the IMP 50, there is no time to hit the processing of other status messages, and the status messages will be stored and the whole network is faulted is.

The world's first wide area network failure occurred as a result of one technical obstacle, and the failure cascaded to the other part. In order to recover, it was necessary to restart all IMPs one by one, which caused a wide area failure for 4 hours.