Single virtual S3 endopoint for Multi-Site IBM Storage Deep Archive

Within a Storage Deep Archive architecture deployed across multiple geographically separate sites, S3 clients should be able to access the service through a single resilient logical endpoint, regardless of which physical site is active.

In the current configuration for geographically distributed environments, load balancers are not suggested, and each S3 endpoint is exposed through a different IP address and a different subnet.

We therefore ask to evaluate a product enhancement providing native support for a single geographic/logical S3 endpoint. This endpoint should automatically direct client traffic to the primary or secondary site based on the actual operational status of the S3 service.

The mechanism should include periodic asynchronous health checks, independent from S3 client queries, and should validate the application-level availability of the S3 service through lightweight operations (e.g., HEAD Bucket, HEAD Object or equivalent mechanisms implemented via specific dedicated health checker).

The requested feature should support active/passive scenarios, configurable site priority, automatic fault detection, controlled failback, maintenance mode, enterprise DNS integration and anti-flapping mechanisms. Support for advanced traffic steering policies, including endpoint weighting, source-subnet-based selection, geographic criteria, would also be valuable.