BERLIN/LONDON

Recent large-scale outages in services such as Cloudflare and Amazon Web Services (AWS) have spurred concerns over the global internet infrastructure.

The internet has become an essential part of everyday life and the global economy. Its ability to provide uninterrupted communications and transactions makes it irreplaceable.

Technical issues with service providers have recently shaken confidence in the global internet infrastructure.

Cloudflare, a US-based company that provides protection services to many websites worldwide, pushed a change on its platform, resulting in the loss of access to platforms such as X, OpenAI, and others on Nov. 18.

One provider’s error caused an outage on a global scale, revealing the structural dependencies of the Internet.

Web pages are accessed via the Domain Name System (DNS), which functions as an internet phonebook, with websites such as google.com being reached via the web browser by matching their domain name with their IP address. Users are affected by any disruptions to this system.

The rise of cloud computing is one of the causes of present-day large outages. Until the early 2000s, companies typically maintained their own servers in their offices, and digital content was largely stored on local, physical storage—outages in this era were more concentrated on individual companies rather than the general public.

Cloud computing became widespread with Amazon, when the e-commerce giant sought to bring a solution to spending excessive time solving the same problems related to computing infrastructure and data storage—the firm set up a shared infrastructure but then seized this new opportunity to apply it to a large part of the Internet.

Most internet users and businesses now rely on cloud computing in some way. Microsoft and Google also launched similar services in addition to AWS—these three players dominate the global cloud computing business.

Apart from software, physical issues like overheating or worn hardware can also cause widespread disruptions. Experts say the key step for businesses is to prepare for possible scenarios before disruptions—this preparation can involve backing up data or maintaining alternative, in-house servers for critical applications.

Massive outages become especially visible in ‘an always-on world’

Alan Woodward, a professor of cybersecurity and computer science at the University of Surrey in the UK, told Anadolu that the internet's structure was not designed for the type of critical and global applications it now supports, as it has been prone to misconfigurations and reliant on trust-based network protocols since its commercialization in the 1990s.

“(The internet) assumes, for example, (that) everyone is telling the truth about global routing in BGP (Border Gateway Protocol),” he said.

“We’ve seen some extensions to these shaky foundations to mitigate some of these issues, but it’s not wholly solved the weakness—having said that, the companies that provide the major services for network protection and global content distribution and storage (like Cloudflare, AWS, Azure, Google, etc.) achieve some remarkable figures for minimal outages,” he added.

Woodward stated that companies and governments relying on large infrastructure providers have basically no choice but to use them, while governments that work with these companies monitor the infrastructure to ensure that it supports critical national services and the highest standards of cybersecurity.

He noted that some errors or changes can result in unintended consequences, such was the case in the recent Cloudflare and AWS outages.

“Nobody foresaw the cascade of failures that followed from one relatively small change at Cloudflare,” he said.

“Even with testing, it’s not always possible to predict what will happen on the live systems, as they are far more complex and larger than the test systems can ever realistically be. I’m afraid this infrastructure is getting ever more complex and interdependent, and the only thing you can predict is that it will happen again,” he added.