Buying storage is a very important decision for your organization. Not only is it a huge financial investment, but it is a basket in which you’re putting many of your proverbial eggs – it must work properly or there will be egg on your face.
As you wade through your many enterprise storage options, there are certain factors you’ll inevitably evaluate, like capacity, performance, features, scalability, renewal costs, etc. You have to buy something that will work for you and continue to for at least 3-5 years. You don’t want to be stuck with a lemon.
While you compare IOPS, capacity and features, you may start to feel that you’re comparing apples and oranges. And as it turns out, you very well might be. (I’m not sorry for all the food metaphors.) To level the playing field, here are some ways your storage vendors may be misleading and some questions you can ask to get to the real information.
Not sharing the true usable capacity
There are several factors that affect this, including the head-scratching issue of the Tibi- vs Terabyte. What you see is NOT what you get in most cases when your capacity is being spec’d out. Software/firmware and other things take up space. Your RAID configuration also impacts this. So before all is said and done and you’ve committed to three years of “Pure” joy or “Smart”Net, do the real math.
Unrealistic deduplication figures
A 7-to-1 dedupe ratio isn’t really what we’d call a badge of pride. That probably means you’ve got some ugly data in there (and you very well could be waging a daily war with your DBAs about it, we get it). Additionally, there are just some things that can’t be or shouldn’t be deduplicated. Ask what can and can’t first, then compare that against the real-life planned workloads of your storage.
Misleading performance metrics
Everyone says their SANs positively scream with speed and they have the IOP measurements to prove it. Just like with most things, dig into their performance testing design a little bit more. Was it actually tested in a realistic scenario? Can you tell if it can move a lot of data quickly at low latencies?
Just like when your evening news says new science shows that drinking beer makes you live longer, you may want to doublecheck the merits of that study (wishful thinking is a powerful thing).
Hidden support or upgrade costs
Have you asked what your renewal costs look like? Do you have a required forklift upgrade to scale out? Find out what years four and five might look like, too – will you be able to get the most out of your hardware or will it become completely cost prohibitive to renew? If a subscription sounds too good to be true, run your own numbers and see if it really does make financial sense.