Let's just focus on LUN provisioned from HDP/HDT pool.
I would like to ask, which kind of configuration deserve for better LUN performance? Assuming within same pool:
- Let's say 500GB: Single LUN with 500GB.
- Let's say 500GB: Multiple LUNs, let's say 10x50GB; then combine in OS level to form 500GB capacity.
From Storage View - Without Considering Read/Write Patterns:
In VSP, (2) is utilizing more processors than (1); can I say (2) is better? What is the case for HUS?
Even though (1) and (2) are sharing the same pool, but (2) having more "chunk" than (1) and thus more parallel processing are possible; can I say (2) is better?
Whole I/O Process - Sequence Read/Write
Concatenate 10x50GB in OS: (1) is better?
Stripe 10x50GB in OS: Overhead by disk management software, so (1) is still better?
Whole I/O Process - Random Read/Write
Concatenate 10x50GB in OS: (1) and (2) could be no difference?
Stripe 10x50GB in OS: (2) is better, because more processing parallel processing from both disk management software and storage?
I understand that carry out a test is always more practical and better, but I would like to listen theoretical answers first.