Hybrid cloud options with auto tiering?

[Exterous]

A bit of a long shot but does anyone have any experience or know of anything that does this? Looking for an appliance or VM that can sit in front of on-prem and cloud storage that would auto tier old data to infrequently accessed cloud storage (Ex S3-IA or glacier) once certain time frames have been met and that is seamless to the end user. Isilon has something like this but is a bit pricey.

 

[dave_the_nerd]

I think most of that stuff is expensive. (Those sorts of features would be pretty enterprise-ey.) But there are a lot of options - just google for "Hybrid cloud NAS".

Depending on your l33t sysadmin skillz and what you're willing to support, it probably wouldn't be too painful to whip one up yourself, either. (Just mount the cloud storage on your file server, find any files that haven't accessed recently, and move/symlink them to cloud storage. If they get accessed, add them to a queue to move back to local disk.)

I've done some of option 2 - except our "cloud" is just in another building.

 

[Exterous]

I think most of that stuff is expensive. (Those sorts of features would be pretty enterprise-ey.) But there are a lot of options - just google for "Hybrid cloud NAS".

Depending on your l33t sysadmin skillz and what you're willing to support, it probably wouldn't be too painful to whip one up yourself, either. (Just mount the cloud storage on your file server, find any files that haven't accessed recently, and move/symlink them to cloud storage. If they get accessed, add them to a queue to move back to local disk.)

I've done some of option 2 - except our "cloud" is just in another building.

We have - plus lots of calls with various vendors. The auto-tiering seems to be messing a lot of people up. The last killer is the 'seamless to the end user'. Its both NFS and CIFS shares that we need to appear to remain in the same space so the end user sees their data in the exact same location its always appeared to them in and retain AD based permissions - regardless of whether its on a SAN, S3, S3-IA etc.

The other issue is scale. 2-3PB of data would be a lot of migratory data for the existing team to support given the static nature it has now