Storage Spaces Direct vs Traditional SAN Cluster

[CoffeeTime]

Hi People!

 

Hope someone can help me out with their thoughts & experiences here. We're in the process of replacing our existing Dell Compellent SAN solution and I've got a few active tenders with vastly different suggestions; and pricing!

 

The first is the traditional SAN setup with an Infratrend SAN (Spinny Disks), two servers, two failover switches so I know what to expect performance wise. Comes in more expensive though by about 30%.

 

The second is is a Storage Spaces Direct solution, two servers for failover, 3x4TB 7.2 Spinny Disks & 2x 2TB Cache Enterprise SSD disks which is a completely alien solution for me. I've tried running a server VM on a single large capacity 7.2k disk in the past and it ran like a dog but Dell reps are telling me the cached SSD drives eliminate this bottleneck by caching actively used data.

 

Will be running 10-15 VMs in total on the new solution can anyone offer me any insight/opinions? Would you stick with the traditional SAN solution or would this new Storage Spaces cached solution work just fine?

 

Many thanks in advance!

Edited July 27, 2020 by CoffeeTime

[Norphy]

The S2D solution would be fine, on the proviso that you got faster spinny disks. 3x4TB? That's likely to be either parity (RAID 5 equivalent) or mirroring with a hotspare (RAID 1 equivalent) so performance would be pretty horrible for anything that's sitting on the spinning rust. 2TB of flash (Again, RAID equivalency would be RAID 1 or 0 so mirroring for resiliency) isn't that much so it's likely it'd get filled pretty quickly. I'd want 10000 or 15000 RPM disks in there, even if it means having to have more of them.

[mavhc]

Unless you're doing archive storage on 8-16TB drives, just stick to ssd, it's cheap enough now.

 

Most important question is how are the 2 servers linked so they can replicate the data?