I'd like to chime in that I'd like something like this as well, but for a different reason.
When using Google Cloud Storage as a CDN for images (all images are public-read), we've found that we can get a small performance boost by randomly distributing the images on a calling page across more than one domain. This is possible with Akamai, for instance:
(the above are fake)
The performance boost comes from browsers deciding to only download so many images in parallel from a given domain. By using multiple domains, you can trick the browser into pulling down more images in parallel.
I figure, with Google Cloud Storage I could simulate this with the two different domain options provided.
...as well as via the CNAME method:
...then, I suppose if one really wanted to, the HTTPS versions of the commondatastorage methods could be used to add two more, but this seems unnecessary and a potential performance loss.
So, it would be really great if we could specify the equivalent of "Bucket Aliases" which allow new bucket names to really resolve to the content found in the bucket it really points to.
Another possibility is if there is some arbitrary (and completely ignored) portion of the domain:
...etc...
However, I could still see this conflicting with unique bucket names no matter what naming convention is given, unless you disallow naming buckets that start with "variation" just as you disallow those that start with "goog", but I think the "Bucket Alias" concept might give better control.
I realize this could also be done by generating each bucket name and syncing content between them, but this seems like a huge waste of storage and syncing activity to simply be able to simulate a bucket alias concept.
Has there been any further thought to supporting similar to this, or did I miss a memo?