When I talk to our customers about the advantages of a Hybrid storage array, I frequently come across many questions that can be categorized as either buying criteria or use criteria. A couple quick definitions are in order to begin. Use criteria are the most important of the two because this is what you are actually going to “use” your array for. Buying criteria are what the vendor is going to use to try to persuade you to “buy” the array, but you will likely never use. Let’s look at a few examples.
To illustrate a point, compare the use criteria on the left with their matching buying criteria on the right. I’m poking a little fun here at some of these features, but the message is the same.
[one_half padding=”0 5px 0 5px”]
Use Criteria
Ability to perform snapshots and clones
Multiple simultaneous connectivity ports and protocols
Multiple disk layout options
Deduplication and compression across the entire array
Healthy business and strong company foundations
[/one_half]
[one_half_last padding=”0 5px 0 5px”]
Buying Criteria
Ability to perform 68,000,125 snapshots and 2,750,000 clones
EVERY connectivity option ever created including sneakernet
Obscure RAID layouts such as RAID3
Deduplication at .025 block size and LHarc compression algorithm
No one ever got fired for buying “X”
[/one_half_last]
When you are choosing an array, ask your competing vendors legitimate questions based on the actual use case you are buying for. How do their controllers work? Are they active/active? What does their version of active/active mean? How much memory does it contain? Is that memory shared between the OS and the cache? Can they support all of those protocols they mentioned at the same time? Within the same array? With the same software code? What types of components do they use? In the world of SSD, not all flash is the same. Are they using SLC, MLC? If MLC is it cMLC or eMLC? What are the environmental characteristics of the system? How much power will it consume? How much heat will it generate? How much of your rack will it occupy? What does the software set look like? Does it support deduplication? Compression? Both? Snaps? Clones? vCenter integration? Are all of these features licensed or included?
There are of course so many questions to be asked and the point is to be sure you tailor that conversation to match what you are using instead of what your vendor is selling and you are buying. Be honest with yourself and to the organization.
Digital IRL 