[ad_1]
And that’s not all! CDNs do not store content closer to the devices they want. They also help direct the internet. “It’s like orchestrating traffic flow in a massive road system,” says Ramesh Sitaraman, a computer scientist at Amherst University in Massachusetts, who helped create the first major CDN as the principal architect of Akamai. “If some internet links fail or pile up, CDN algorithms quickly find another way to the destination.”
So when a CDN goes down, you can start to see how it can take up a lot of parts of the internet. Although it doesn’t explain how the impacts on Tuesday were so far-reaching, especially when there are multiple releases in those systems. Or at least, it should be.
Consolidated CDNs
Again, it’s not clear what happened in Fastly. “We have identified a service configuration that has caused disruptions in our POPs around the world and we have disabled that configuration,” a company spokesman said in a statement. “Our global network is reconnecting.”
“Service configuration” means anything; the only certainty is that, in any case, it had a great influence. According to Fastly’s incident report, all continents other than Antarctica felt the impact. Even after Fastly fixed the underlying issue, he warned that users could still see a lower “cache hit ratio” – how often you can find the content you’re looking for stored on a server – and “increase the original load.” which refers to the process of returning to the source of non-cached items. In other words, the closets are still pretty bare.
It is surprising that a disruption should occur, as CDNs are typically designed to deal with these storms. “In principle, there’s a lot of redundancy,” says Sitaraman, who generally talks about CDNs. “If one server fails, the other servers can take over the load. If an entire data center fails, the load can be moved to other data centers. If things work perfectly, you may experience a lot of network outages, data center problems, and server failures; The CDN’s resistance mechanisms would ensure that users never see degradation. “
When things go wrong, Sitaraman says it is usually associated with a software error or configuration error that is pushed to multiple servers at the same time.
Even then, sites and services that use CDNs typically have their own releases. Or at least, they should. Actually, this morning to find out what kind of diversity they have in the speed of response of various diversified services, says Medina. It took Amazon about 20 minutes to get it up and running again, as it could divert traffic to other CDN providers. When anyone relying solely on Fastly had no automation for the system to accommodate the outage, he had to wait.
“The disruption was the result of a monoculture,” says Roland Dobbins, chief engineer of the security company Netscout Arbor. It suggests that all organizations with a significant online presence should have multiple CDN providers to specifically avoid such situations.
Their options, however, are increasingly limited. As Amazon, Google, and Microsoft have largely highlighted the cloud, three of CDN’s providers — Cloudflare, Akamai, and Fastly — dominate the online content flow. “There is a high concentration of use within a few service providers,” says Medina. “When one of these three providers has a problem, it’s usually not something that lasts very long, but it has a big impact on the Internet.”
That’s a big part, Medina says, of why interruptions of this kind have occurred more often and why they will continue to get worse. Baseball needs a cut man; intersections need traffic police. The less you have to trust, the more connections you will lose and the bigger the crash.
More great KABEKO stories
[ad_2]
Source link
