Just to keep everyone on their toes, Amazon Web Services’ Content Delivery Network (CDN) experienced some glitches on Thanksgiving Eve, according to various reports all citing the AWS status page. According to that page, users experienced “elevated error rates when making DNS queries for CloudFront distributions: between 4:12 p.m. and 6:02 p.m. PST” on Wednesday, November 26.
The problem was global and has been noticed as far away as Australia, where posters to the AusNOG mailing list reported problems lasting for up to two hours. Chat on the list suggests Amazon has re-routed traffic over different submarine cables to address the issue.
[company]Amazon[/company] cloud customers, including online merchants and content sites, rely on CloudFront or other CDNs to place commonly viewed content on servers around the world — basically putting it closer to users so that it comes up faster.
Things appeared to be back to normal by early evening November 26 Pacific time. Had the glitches persisted into Thanksgiving Day specials or (God forbid) Black Friday, we’d have heard far more about this issue.
The CloudFront issue comes a week after Microsoft Azure experienced problems of its own.
Why you need to know what Facebook’s doing with networking
[company]Facebook[/company], which has already taken on web-scale compute power with the Open Compute Project, is starting to talk more about what it’s doing on the networking front. Najam Ahmad, the company’s director of network engineering, talked to us last week more about why Facebook needed to re-scope its networks. Check out the podcast below.