What is the best compression or dedupe you have had per job/VM?

Honestly, I’ve never really checked on this.  And, I have a lot of environments to monitor that one.  Not sure if this is something that Veeam ONE can monitor or not, but it’s on my radar to deploy Veeam ONE and attach it to my Service Provider Console to see what it can do there, but I haven’t had a chance to do so.

And yeah, you’re going to get crap dedupe and compression on file servers with data that can’t be compressed any longer, as well as deduped.  Video, Images, Encrypted files, etc.

Somewhat related….last week I was in a PowerStore bootcamp with Dell.  I personally have a PowerStore 500T in production at my head office and Dell guarantees a 4:1 DRR (Data Reduction Ratio) due to having a separate chip on each controller that handles the DRR functions.  Not that the DRR Guarantee excludes file types that don’t dedupe and compress well, just like we’re talking here.  So I’m getting what I expected to be a 50TB effective SAN, about 25TB because I’m getting between 1.8:1 - 2:1 DRR.  Not the worst, not what I had hoped for (because I hadn’t scoped out just how many images we have - we do Electrical Construction and have a lot of Virtual Construction/CAD and building plan images), but not really bad either all things considered.

I did actually question Dell on this as I have a client that I just added about 45TB of disks to their SAN yesterday but as it’s a Compellent SC which is now deprecated, but it has 4 years of support yet, I want to start thinking in about 3 years about what needs to replace it as that client works in media as an advertising agency and has a ton of images, layouts, 4k video, and the like.  PowerStore (and to some degree, Unity XT) is the direct replacement to the SC series arrays, but I would expect to get about 1:1 for my DRR on their data (the SC currently is getting 1.18:1 DRR).  The PowerStore shines on data that it can DRR because the it can solidly pack away the data, except when you have data that doesn’t compress well. Dell did acknowledge that it’s not ideal to have this type of data but did still cite PowerStore as a possible solution, even though you’re not going to get a very good DRR….so I’m not sold on it for that use. 

It may be time to start looking at something like PowerScale (Isilon) which was vetted as a possible solution when we ended up settling on the SC they have now, but the major concerns were cost and backups.  Dell’s answer to backups on the Isilon were NDMP backups which are obviously not great, or setting up another Isilon somewhere else, replicating the data and using versioning.  Veeam NAS backup would be great, but when I’m talking about 100TB or so of data (today), that can be pretty expensive for your smaller businesses as well.