Presentation: Understanding Progressive Collapse: How To Avoid A Cascading Failure

Sam Newman, an independent consultant specializing in cloud, microservices architecture, and continuous delivery, discusses the concept of progressive collapse in civil engineering and its application to distributed systems. Progressive collapse refers to a situation where a small failure in a system results in a significant collapse, which can be compared to a cascading failure.

Using real-world examples, Newman highlights the importance of understanding and preventing progressive collapse in digital systems. He shares the story of Ronan Point, a tower block in Canning Town, England, that partially collapsed in 1968. Only four people died that day, a miracle considering the scale of the collapse, which occurred in the early morning when most residents were still asleep.

The collapse began when a faulty gas nut caused a gas explosion in Mrs. Ivy Hodge's kitchen, which she survived by being knocked unconscious. The explosion blew out the outer wall, a load-bearing structure, causing the four floors above to collapse and shear off her bedroom. This is an example of progressive collapse, where a small failure triggers a larger collapse through a chain reaction.

Newman explains that the building methods used for Ronan Point were rushed due to post-war housing needs and the aftermath of the Blitz in London. Despite being built quickly, the tower block should have been designed with proper load-bearing structures to prevent progressive collapse.

Newman also discusses two other examples of progressive collapses: the US-east-1 AWS outage in October of the previous year and a website selling used cars that he was involved with. These examples illustrate how understanding progressive collapse can help software leaders strengthen components, isolate failures, and reduce interconnections to prevent catastrophic cascades in digital systems.