A pair of big CNDs had outages earlier this week that showed, as Frost & Sullivan analyst Dan Rayburn pointed out in a blog post yesterday, that "even though all the CDNs always talk about the redundancy built into their networks, ALL networks have outages at one time or another."
Akamai told Rayburn it experienced a DNS-related outage Monday due to "a configuration change to take some DNS servers out of service." The change propagated to the production network, causing servers to repeatedly restart. Six minutes after the problem began, engineers began working on it and had the network back in service.
Amazon also had a problem Monday, causing part of its EC2 service to go down, affecting multiple customers, among others, Netflix. The company said it was down about 45 minutes and said they had been investigating connectivity issues at the time of the outage.
- see this post