What it reveals about the internet’s backbone

Three hours after the outage was confirmed, Amazon said that the cloud service had returned to normal operations. However, the company also said that some AWS services had a backlog of messages that would take a few hours to process.

What led to the AWS outage?

While the company is yet to share a detailed post-mortem report on the outage, AWS said that it traced the source of the problem to the “DynamoDB endpoint in the US-East-1 Region.”

DynamoDB is a centralised database service used by many internet-based services to track user information, store key data, and manage operations. However, AWS has attributed the outage to a domain name system (DNS) error. DNS works as a kind of phonebook of the internet as it translates internet addresses into machine-readable IP addresses that connect browsers and apps with websites and underlying web services.

DNS errors are not uncommon and can be addressed quickly. But since so many sites and services rely on AWS, a DNS error can lead to widespread disruption. The system’s failure to correctly resolve which server to connect to resulted in cascading failures that took down large swathes of the internet.

The outage on Monday was the third major failure tied to AWS’s US-East-1 data centre cluster in the past five years. It not only underscores how dependent businesses have become on the tech giants that deliver third-party cloud computing services but also serves as a reminder of how fragile the internet’s backbone can be even if the outage is resolved quickly.

How did AWS become the internet’s backbone?

The cloud computing market is said to have a brittle foundation because it is dominated by three hyperscalers, namely AWS, Microsoft Azure, and Google Cloud, who reportedly control two-thirds of the global market.

During the COVID-19 pandemic, several companies sought to digitise operations and move to the public cloud in order to tap into a wide range of services. This led to a further increase in revenues for cloud service providers like AWS, Azure, and Google Cloud that already dominate the global market. Earlier this year, Gartner forecast a 21.4 per cent jump in worldwide end-user spending on public cloud services to more than $723 billion.

Reliance on cloud giants like AWS has further strengthened cybersecurity and stability around the world by establishing common guardrails and best practices for all customers. Yet, even with their generally resilient infrastructure, AWS and its competitors remain vulnerable to large-scale disruptions because the internet is a complex web of overlapping services.

Past outages on this scale have stemmed from various issues ranging from faulty updates and accidental injection of bad code to incompatible third-party software changes. For instance, a faulty software update by cybersecurity company CrowdStrike affected devices running Microsoft Windows and caused massive disruptions around the world in 2024.

Why is concentration of cloud services risky?

Experts have repeatedly pointed out that competition in the cloud computing market has been stifled by tactics such as bundling services, opaque pricing schemes, and high switching costs that keep customers locked in.

AWS, Google Cloud, and Microsoft Azure have been known to impose egress fees, dubbed as the ‘cloud tax’, on customers who want to migrate to another cloud provider, or download and move their data elsewhere. These practices have created steep entry barriers for potential challengers, allowing big tech companies to maintain their dominance over critical parts of the internet’s infrastructure.

The concentration of the internet’s core infrastructure in the hands of just a few major cloud providers has created critical single points of failure. A three-to-six-day outage at one of the top three cloud providers could cause $15 billion in damages in the US, according to a 2018 study by insurance giant Lloyd. The risk management processes of these big tech companies are also opaque.

Can AI drive competition in the cloud services industry?

While large cloud computing providers still rule the global market, the artificial intelligence (AI) boom has opened the door to a new generation of players looking to challenge the status-quo.

For instance, Ubicloud is a startup that is looking to build an open-source alternative to AWS. It was part of Y Combinator’s W24 Batch and received $16 million in seed funding. Companies like CoreWeave, Together AI and Lambda Labs have also capitalised on the surge in demand for computing power driven by the AI boom, with some of them looking to offer distributed compute capacity.

Regulators could also play a role in levelling the playing field as big tech companies such as Meta, Google, Apple, and Amazon are already facing antitrust scrutiny in the US and European Union.