I am trying to clarify what "effective capacity" means as far as EMC spec sheet is concerned.
I have 2 x 20TB xbricks so i see 30TB capacity is XMS. Two questions about the screenshot above
1) Under Usable Capacity what is 33.3 ?
2) Effective Capacity is 182 TB (assuming 6x savings). So what does it mean, can i have 360TB of effective capacity on the same cluster if i have 12x savings (say i have a VDI farm)
1) For all fields, you're seeing the usable capacity in <base10/marketing TB> / <base2/engineering TB>. In marketing capacity, it is assumed that there are 1000 bytes in a kbyte, 1000 kbytes in a mbyte, etc. In engineering capacity (aka "reality"), there are 1024 bytes in a kbyte, etc. Also see https://en.wikipedia.org/wiki/Kibibyte. Usable capacity means the capacity you get after XDP (data protection), but before data reduction services like thin provisioning, inline compression/dedupe, and virtual copies.
2) Effective capacity is showing you the actual capacity you'll be able to store, assuming a 6:1 overall data reduction ratio. So yes, if your actual data reduction ratio was 12:1, you'd get twice the effective capacity.
To add to Sean's response above - 182TB mentioned in effective capacity assumes a 6:1 data efficiency. That would be the logical capacity that your applications would be able to store on the XtremIO array.
However, there is an upper limit to the logical capacity that the array can support. That has to do with the fact that we do need to store the metadata for all that logical capacity in the memory and we have a finite amount of memory in the array. The actual amount of maximum logical capacity would depend on the dedupe and the compression ratio that the application workload is driving - but we do guarantee that if you were seeing a 6:1 data efficiency - the array will be able to support that resulting logical capacity. It does NOT mean that if you are seeing a 600:1 data efficiency ratio, you would be able to store 100 x 182TB of logical capacity. You would run out of memory on the storage controllers before you reach that logical capacity.
Thank you Sean and Avi.
So Avi , what is that logical limit for a cluster with 2 xbricks ? Can you provide documentation for my reference. ?
The maximum logical capacity is specified in the release notes (see Scalability section). That number is per X-Brick. If you have a dual X-Brick - multiply that number by two.