As I finish my Zeus latte, I hear a distant murmur. It echoes and amplifies, bouncing from one marble table to another. “Canvas is down?” “How will I submit my work?” “Wait, all of my lectures were on there.” A domino effect of dumbfounded expressions circle the room.
On Oct. 20, Amazon Web Services experienced an outage in their US-East-1 region due to a Domain Name System issue that impacted DynamoDB, their native database system. This issue had cascading effects into AWS cloud computing system, EC2, and their Network Balancing Load. As a result, Canvas, which is hosted on AWS, was disabled and cut off access to the academic management system for students and professors alike. This outage didn’t just disrupt coursework, though. It exposed the dangerous reality of reliance on digital infrastructure. Canvas is a useful tool, but an Ivy League School should not have stunted academics from one temporary system crash. Cornell students work so hard through our courses, but if one bug in a system update can potentially erase all of that work in seconds, how is that progress tangible?
AWS is the global leader for cloud services and their respective data centers. In a report by Reuters, the outage impacted several sectors. Uber was impacted, which caused a temporary blip in accessible transportation. Coinbase and Robinhood’s disruptions led to financial disruptions. Social media platforms like Snapchat and Reddit felt the outage effects, and even entertainment services (Netflix, Epic Games) were impacted. As an AI Engineering intern last summer, I learned that AWS is actually incredibly versatile for startups. Emerging startups such as Databricks pool their services through AWS. AWS certification is the ABCs of software engineering in 2025. This ubiquity is their company’s strength and our collective weakness. They are accessible to numerous clients across many sectors, but these digital products are at the same risk for any outage, which is how the outage from Oct. 20 affected so many companies and services. While AWS does not control the entire internet, it’s one of four leaders in global cloud services among Microsoft Azure, Google Cloud Platform and Meta Platforms.
We think of the internet and its many ornaments as tangible services, but they do not exist in a proverbial “cloud;” they are physically implemented across networks of data centers. AWS owns many of their data centers or occupies them on a long-term lease. The companies leading this market are large enough that they can afford the hardware and cooling systems necessary to keep these centers efficient, but they centralize internet traffic impacting every sector of the economy: they are functionally an oligopoly of the digital world.
AWS also provides services for the United States Federal Government, Department of Defense, state and local government, and intelligence agencies. This broadens the issue from a campus halt on due dates to one of national security. Data center blips or outages are inevitable alongside innovation, but the centralization of these networks through the data centers of a few major companies create a jugular vein of cybersecurity. A more serious internal issue, targeted attack or natural disaster could disrupt global operations. Decentralizing these services mitigates all of their associated risks.
To be accessible to the internet, it is easiest for small businesses to work through AWS or one of the other global leaders. It is also the most reliable option for large businesses, agencies and governments to run web services internally and to the general population. But this does not excuse the concentrated control of their market. In the 19th century, railroads were changing the national landscape through transportation. Trains are still an essential service, but there is still competition in the market because of antitrust legislation.
Theoretically, this market concentration makes it easier to regulate their operations. However, the tech world moves much faster than federal agencies, and data centers will need to maintain agile operations. Governmental procedure, especially given their inefficiency in the weeks leading to the shutdown, simply would not accommodate the volume or demand of digital traffic when the companies bleed into so many economic sectors.
It might seem logistically ideal that major tech companies finance these data centers. Maybe it is easier to regulate the few rather than a more competitive market. These companies may lead the market, but the outage shows that they are not being properly checked. If regulation would be better through these market leaders, the crash could have been cautioned or prevented — there could have been some kind of procedure or penalty in place. As it stands, these corporations hold our data, government and world in their hands, and they need to let go. Government incentivization for smaller companies to scale data center operations is a practical necessity and would ultimately protect our digital infrastructure.
The Canvas outage showed much of Cornell’s student body the severity of having all of your digital eggs in one basket. It should also make you realize the foundation of the digital infrastructure. Without some type of deconcentration of that world, it could break at a moment’s notice.
Zak Kheder '26 is an Opinion Columnist studying Electrical & Computer Engineering in the College of Engineering. RenAIssance Man explores questions of intellectualism and what it means to be a student in the age of information and technology. He can be reached at zkheder@cornellsun.com.
