Four hours of chaos a stark warning about the fragility of the internet
Millions of people were unable to use their favourite apps after an outage on Monday night in what was a troubling reminder.
Many popular internet services, from streaming platforms to messaging services and some banks, went offline for hours on Monday night (AEDT) due to an outage in Amazon’s crucial cloud network.
The disruption affected streaming platforms, including Amazon’s Prime Video service and Disney+, as well as Perplexity AI, the Fortnite game, Airbnb, Snapchat and Duolingo.
Mobile telephone services and messaging apps Signal and WhatsApp were also affected in Europe, according to Downdetector, a website monitoring internet problems.
Even some banks were impacted and pointed to Amazon Web Services (AWS) as the problem.
AWS handles nearly a third of the planet’s cloud infrastructure market, powering millions of apps and websites around the world.
AWS’s maintenance site said its engineers scrambled to fix the underlying DNS issue once they became aware of “increased error rates” hitting multiple services.
More than four hours later, though some problems persisted, AWS said it was on the path “towards full recovery” and most of its operations were “succeeding normally”.
The outage showed “how reliant we all are on the likes of Amazon, as well as Microsoft and Alphabet for many of the online services we more or less take for granted,” said financial analyst Michael Hewson.
“On an economic level it’s almost akin to putting all of your economic eggs in one basket.”
The outage highlighted global reliance on the technology, which offers businesses on-demand IT resources without heavy investment in expensive server farms.
In its most widespread form, the so-called “public” cloud relies on shared data centres, where clients hire IT capacity. The “private cloud”, on the other hand, consists of dedicated machines reserved for a single company.
The United States dominates the global market for cloud computing. AWS held 30 per cent of the market share in the second quarter of this year, followed by Microsoft Azure (20 per cent) and Google Cloud (13 per cent), according to the Synergy Research Group.
These three firms have been dubbed “The Big Three”.
Other players, such as the US firms Oracle, IBM, Salesforce or Akamai, as well as Chinese giants Alibaba, Tencent and Huawei, share parts of a rapidly growing market, with significant barriers to entry.
There are three distinct major models of cloud services. In all cases, providers handle the installation of services, storage drive and network connections in data centres across the world, which consume huge amounts of energy daily.
The most popular model - Software as a Service or SaaS - allows the direct use of online apps such as Gmail, Zoom, Teams or Slack.
In 2023, 96 per cent of EU cloud client companies bought at least one SaaS service In the case of Infrastructure as a Service or IaaS, businesses hire resources with only minimal pre-configured settings to install their own software.
The third - Platform as a Service of PaaS - is an intermediate model, where the client can outsource not only the management of the infrastructure but also part of the software that runs their apps.
Local providers have seen their share of the market shrink in recent years and captured by foreign giants known as “hyperscalers”.
Hyperscalers are able to finance the construction of ever more data centres, even if the return on investment takes time.
Significant amounts are at stake as the construction of large data centres can exceed a billion dollars.
“This major online outage underscores a stark reality: Business operations associated with one critical vendor in a region can cascade into global instability,” said independent cyber analyst Rimesh Patel, quoted by the Science Media Centre, a UK charity focused on science and engineering.
“It highlights how interdependent our digital infrastructure has become,” said Alan Woodward, a University of Surrey professor quoted by the same centre.
In July 2024, another global online outage occurred when a US cybersecurity company, CrowdStrike, issued a faulty update to its software used by airports, hospitals and many organisations.
According to Microsoft, some 8.5 million devices were affected, resulting in a systems crash and users being confronted with a “blue screen of death”.
- AFP
Merciless response to Roblox’s big move
Grovelling apology as 20pc of internet fails
