List prices mean zilch. Remember that when buying from upstarts
Be nice to vendors – even if they're the new kids on the block
Storagebod Storage industry blogger Howard Marks has written about telling storage startups apart from upstarts. Except, when you dig into his piece, it reads more like how to be a savvy tech buyer.
Marks' online post, titled ”Separating Storage Startups From Upstarts” got me thinking, as someone involved in making large purchasing decisions.
List price from any vendor is completely meaningless. Most vendors only seem to have list prices to comply with various corporate governance regimes. Of course, having a list price means that the procurement department can feel special when they’ve negotiated the price down to some stupidly low percentage of the original quote; in a world where 50 per net discounts are common, list prices are nonsense.
What is true is that a startup's list price will often be lower than traditional vendors' list prices; it’s got to be, even if it's just to start a conversation.
In my experience, the biggest mistake an end-user can make is not being willing to take any competitive bid; dual supplier-type arrangements are good but often can lead to complexity in an environment. If you can, split your infrastructure into domains – say, for example, buying all your block from one vendor and your file storage from another vendor. Or perhaps you have a tiering strategy that allows you to do something similar.
But loyalty can bring rewards as well. The word “partnership” is thrown around like a child's toy these days, but learning to work with your vendor is important. Knowing which buttons to press and learning how a vendor organisation works is often key to getting the most out of infrastructure procurement.
Marks' assertion about a three-year life of an array in a data centre doesn’t ring true for me, or many of my peers. Array lifespans of four to five years seems to be the minimum. If it were three years, we would generally be looking at an actual two-year useful life of an array; six months to get on, two years running and six months to get off. Many organisations are struggling with four years and as arrays get bigger, this is getting longer.
As for the pain of going through a technology refresh every three years? We’d be living in a constant sea of moving data whilst trying to do new things as well. So my advice to you is, plan for a five year refresh cycle…
My advice to any technology buyer is to pay close attention to the upstarts but also pay attention to your existing relationships; know what you want and what you need. Make sure that any vendor or potential vendor can do what they say; understand what they can do when there are problems. Test their commitment and flexibility.
Look very carefully at any new offering; is it a product or a feature? If it is a feature, is it one that is going to change your world substantially? Violin Memory arguably fell into the trap of being a feature: extreme performance is something that few really need.
And when dealing with a new company, understand where their sales culture has come from… if you had a bad experience with the sales drone's previous employer, there’s a fair chance that you might have a similar experience again. ®
Sponsored: Benefits from the lessons learned in HPC