A widespread internet outage caused by a problem with Amazon Web Services (AWS) disrupted services globally this morning. Affected services included social media apps like Snapchat and Signal, other Amazon platforms, government services and some airlines. The problem appears to have stemmed from issues at a critical Northern Virginia hub for global internet infrastructure.

Lee McKnight is an associate professor in the Syracuse University School of Information Studies (iSchool) whose research specialty includes cybersecurity. He provides written comments that can be quoted directly. He is available for interviews as more information becomes available, and for future topics related to cybersecurity practices in the public and private sectors.

McKnight says:

“When AWS, the largest cloud computing service provider with one third of the market, sneezes, many internet-dependent services die until whatever issue caused the outage is resolved.

“For this outage and others, it appears to be the ‘DNS resolution of the DynamoDB API endpoint in US-EAST -1.’

“Translated to English, the Domain Name System—the global internet’s directory system for all websites worldwide—is a critical digital public infrastructure the world depends on. It could not be reached from AWS’s massive data centers in the Eastern U.S.

“In this case of today’s AWS outage, it does not appear to be a cyberattack on their APIs, but rather an internal issue, that is, a cloud software misconfiguration. Essentially, every time software is updated, to add a new feature or close a security hole that has been discovered, there is a need for quality assurance purposes to test its performance before releasing the update to all systems.

“In this case, AWS’s ‘s US-EAST-1 went down in the middle of the night for Eastern Time residents but took out services nationwide and worldwide that were drawing some data or service from AWS. While many are now restored, outage reports persist.

“I have no insider knowledge or exactly what happened, but a reasonable supposition is that a routine cloud software upgrade cascaded into today’s outage incident.”

To request interviews or get more information:

Daryl Lovell
Syracuse University Media Relations
M 315.380.0206
dalovell@syr.edu | @DarylLovell