A couple of days ago I published an article where I compared Pure Storage to EMC VNX. It was not meant as a direct comparison between the two products, but the original idea was to position Pure Storage as a general purpose next generation midrange array. The comparison was made because EMC VNX is generally considered a classic, and successful, example of “old generation” midrange array. That’s it.

Some tweets and comments made on twitter got me thinking, so I’ve decided to add some of my thoughts to that earlier post.

Needs come before solutions

Comparing new products to old products does not make sense and neither does comparing products based on similar technologies, if not done in the right context.

sarahPorsche-911_Carrera_CabrioletFor example, comparing Pure to Violin because they are both All Flash Arrays, makes no sense at all.
Violin is as fast as hell but has no features, on the other hand Pure has plenty of features but latency and 4K blocks performance are lower and less predictable.
Without questioning the quality of either one, they are both good in certain situations and worse in others.
If you need to run a highly transactional DB on a Unix server you’ll probably stick to a product like Violin, but if you need a storage for VMware (especially VDI) Pure will do it much better.

So, just for clarity, it’s senseless comparing a two-seater sports car with a MVP just because they have 4 wheels and an engine. Both are good, it only depends on your needs: if you are a wealthy 50 year old divorced man, your needs will probably differ from that of a soccer mom’s ! 🙂

But we will compare apples to oranges again (and again)

The problem, as always, then moves over to the commercial/sales side of things. If you know what you need, nothing will stop you from buying a better solution at a similar price.
That has happened all along. And this is what is happening with solutions like All Flash Arrays. If the total cost of the solution is comparable, then why not buy the better one?

At the end of the day if you weigh out all the aspects (not only $/GB raw) you’ll probably find that the next generation solution is much more cost effective than the older one. (I’m trying to generalize the concept but again, I’m thinking about the VNX and Pure, sorry)

And, just to answer another question that has popped up often in recent days: “why don’t you compare Pure to XtremIO?”.
It’s only because EMC, in the field, resellers are the first to propose the VNX product instead of XtemIO to medium sized customers (and, BTW, the same is happening for other vendors!)
Most end users probably don’t know about the existence of XtremIO (as is for some EMC resellers apparently!). Customers have needs, they ask for a solution and they often receive technically non comparable solutions (and prices).
I do consulting for a living and it’s my job to advise customers about these products/solutions.
And you know what? the more complex their needs the more are the wannabe-solutions different (but you already know that!)

Do we really need to care about technology?

If I were an end user I wouldn’t care less about technology. It’s not worth it.
I would care only about TCA and TCO. If a vendor can give me the best TCA/TCO, with a product that fits my needs, I wouldn’t ask for more… would you?

Of course, better technology and its implementation is the way to achieve that, but it’s not my problem after all. Isn’t that the vendor’s problem?

Why it is important

GMO AppleorangeGetting back to the storage, at the end of the day, the comparison is not really important per se. The most sensible thing to do is to make the right comparison in the right context.
Comparing products because they are based on the same technology does not make sense, and the same goes for the contrary.
On the flip side, I will continue comparing different products if they are aimed at solving the same need/problem… or even if they were supposed to.