Amazon AWS Outage Reveals Tech Vulnerabilities and Provider Risks

TL;DR

A recent Amazon Web Services (AWS) outage caused significant global disruptions, affecting thousands of apps and websites including Snapchat, Roblox, and educational platforms like Canvas. The incident, traced to an internal monitoring subsystem, highlights the risks of relying heavily on a few major cloud providers and raises concerns about internet infrastructure resilience. Experts urge greater diversification in cloud computing.

Amazon Web Services Outage Disrupts Internet Services

A significant outage in Amazon's cloud services triggered widespread disruptions across various apps and websites, impacting thousands of users globally. The incident highlighted the reliance on a small number of tech companies for internet infrastructure. Downdetector reported over 2,000 companies affected, with 8.1 million user-reported issues, including 1.9 million in the US and 1 million in the UK. Amazon confirmed that its cloud services had "returned to normal operations" by Monday evening.

Affected Platforms and Services

The outage impacted a wide array of platforms and services. Affected platforms included Snapchat, Roblox, Signal, and Duolingo, along with Amazon-owned operations like its retail site and Ring doorbells. In the UK, Lloyds Bank, Halifax, and Bank of Scotland experienced issues, as well as the HM Revenue and Customs website. Other impacted platforms included Wordle, Coinbase, Slack, Pokémon Go, Epic Games, PlayStation Network, and Peloton.

Amazon's Response and Root Cause Analysis

Amazon Web Services (AWS) reported the initial problem around 8 a.m. UK time and indicated resolution progress by 10:30 a.m. UK time. However, the company acknowledged ongoing elevated errors later in the day. AWS stated it was implementing request limits to aid recovery. The root cause was traced to an internal subsystem responsible for monitoring load balancers. According to Steven Murdoch, a security engineering professor at University College London, the incident appeared to be accidental rather than malicious.

Technical Details of the Outage

The outage originated in the US-East-1 region. Cisco’s Thousand Eyes, which tracks internet outages, reported a surge in problems, many located in Virginia, where AWS has numerous data centers. AWS's online health dashboard pointed to issues with DynamoDB, its database system. Amazon clarified that the issue was not due to a cyber-attack but an internal problem with a subsystem monitoring its load balancers.

Impact on Education Sector

The AWS outage significantly impacted the education sector. Many college and K-12 students experienced difficulty accessing or submitting homework and course materials due to the outage affecting Canvas, a widely used educational platform. The University of California, Riverside, reported that students could not submit assignments, take quizzes, or access course materials. Ohio State University informed its 70,000 students that online course materials might be inaccessible.

Concerns and Expert Opinions

Experts emphasized the risks of relying on a small number of companies for internet infrastructure. Dr. Corinne Cath-Speth, head of digital at Article 19, stressed the urgent need for diversification in cloud computing. Cori Crider, executive director of the Future of Technology Institute, highlighted the vulnerability of critical infrastructure to US tech giants. Madeline Carr, professor at University College London, acknowledged the debate around the risks and benefits of relying on hyper-scaling companies.