Amazon.com's cloud computing division reported a failure in one of the availability zones at its data center in Northern Virginia on Thursday. Concurrently, derivatives exchange CME Group Inc and cryptocurrency exchange Coinbase both indicated issues with their trading platforms. It remains unclear whether the problems at AWS and CME are connected. However, Coinbase stated that the issues its platform experienced were caused by the AWS outage. AWS explained that the failure originated from increased internal temperatures within a single data center. With additional cooling system capacity coming online, initial signs of recovery have now been observed. As part of the restoration efforts, the cloud platform stated it has redirected traffic away from the affected "availability zone" for most services. An "availability zone" consists of one or more interconnected physical data centers designed to operate independently within an AWS region. Coinbase mentioned that some users might experience performance degradation due to the AWS problem but added that customer funds remain secure. Coinbase said it is working to restore trading functionality in the short term. The world's largest derivatives exchange, CME Group, stated in an update on its official website that it has completed necessary maintenance work and users can now log in to its CME Direct trading platform. CME did not specify the exact cause of the technical failure and delays.
This latest outage follows a major AWS failure in October of last year, which caused chaos for thousands of websites and popular global applications, including Snapchat and Reddit. That incident was the largest internet disruption since the 2024 CrowdStrike failure that paralyzed hospital, bank, and airport technology systems, highlighting the fragility of globally interconnected technology. The following month, CME Group experienced one of its longest-lasting outages in years, halting global futures markets for several hours and suspending trading in stocks, bonds, commodities, and currencies. At the time, CME attributed the failure to a cooling system breakdown at a data center operated by CyrusOne; CyrusOne stated that the failure at its Chicago-area facility affected services for clients, including CME.
Reviewing major incidents in recent years, the triggers for technical failures are becoming increasingly complex. The 2020 full-day shutdown of the Tokyo Stock Exchange due to a single hardware failure revealed the inflexibility of traditional backup system switching logic. More alarmingly, climate change is evolving into a financial risk. The recent AWS outage was not caused by a cyberattack but by a pure cooling system failure. As extreme global heat events become more frequent, the physical redundancy designs of data centers, these major electricity consumers, are being pushed to their limits. When coolant cannot dissipate server heat, liquidity worth hundreds of billions can be physically "circuit-breakered" in an instant.
Over the past decade, financial institutions have aggressively migrated their architectures to public clouds in pursuit of ultra-fast trading and cost optimization. However, the frequent occurrence of outages indicates that "multi-cloud strategies" and physical redundancy are no longer optional but mandatory for compliance.
Comments