AWS Restoration Delays in Bahrain and UAE Zones
Amazon reports it cannot restore access to its cloud facility in Bahrain or one of its three data hosting zones in the United Arab Emirates (UAE) six months after drone strikes.
Explanation
According to Amazon, the damage to its infrastructure "exceeded what our regional and multi-AZ services are designed to withstand." Multi-zone redundancy, a feature cloud customers pay for, did not prevent the outages.
Availability Zones, separate facilities with independent power and cooling systems, are crucial for cloud provider reliability. Amazon’s strikes affected multiple Availability Zones within the same region, surpassing the designed resilience.
Timeline of Events
- Early March: Two of Amazon’s UAE data centers were directly hit, and the Bahrain site was damaged by nearby strikes.
- Later March: Additional disruption occurred in Bahrain.
- End of April: Around 60 services remained down in the region.
- May: Amazon initially estimated restoration would take several months, advising customers to migrate services and backup data elsewhere.
Current Status
Amazon states it has helped Bahrain customers move their operations but hasn’t restored the affected regions. The company hasn’t disclosed if any customer data was permanently lost or if damaged buildings can be repaired.
Implications
The ongoing outage challenges the assumption that cloud infrastructure is inherently resilient. It raises questions about provider reliability, insurance coverage, and regulatory considerations in regions relying on cheap cloud services.