In general, if you buy your network, storage and/or server kit from a big established vendor, you can count on getting support, both pre- and post-sales, from that vendor for the life of the product. Like everything else in life, that support, or more accurately, the peace of mind that support will be forthcoming, comes at a cost. Part of that cost is of course that solutions from Dell or NetApp will cost a more than the roughly equivalent product from a startup such as Tegile or a second-tier vendor like SuperMicro, although the big boys are usually more willing to discount.
That discount depends on how much business you do with the vendor, the size of the particular order and how well your salesman is doing at making his quota, especially towards the end of the quarter. The other factor in how big a discount your vendor will give you is how sure it is that you’re going to buy from it. In my opinion, the biggest single mistake an end-user organization can make in purchasing is to declare itself to be an IBM, EMC or vendor X shop.
Once you’ve let your vendors know that you’re going to be loyal, they have less incentive to give you that extra 10% off on a big purchase. Even worse, you and the other vendor loyalists will be the only customers for the products in their line that are clearly more expensive or less capable than the competition, such as an object storage system that uses Fibre Channel arrays in the backend or the re-labeled KVM switch that’s double the actual manufacturer’s price.
The bigger additional cost of only buying from big established vendors is forgoing the innovation that startups bring to the market. The big boys took a couple of years to bring out all-flash arrays that were competitive with Pure Storage, Nimbus, SolidFire or Whiptail, and in fact Cisco and EMC had to buy Whiptail and XtremIO respectively to get into those markets. Today, hybrid storage systems such as those from Nimble, Tegile or Tintri still make better use of flash and have functionality the hybrid versions of old arrays from the big boys don’t.
The problem, as I know all too well, is that any time you buy from a startup, you run the risk that that startup will go belly up, like Starboard Storage. As a technologist with a preference for innovation, I’ve been on both sides of the startup game. I installed some of the first EqualLogic arrays and those clients were winners. On the other hand, I bought a network admission control system from Lockdown Networks and it went out of business before we even got the darned thing fully implemented.
[Read Howard Marks' analysis of an emerging software-defined storage technology that's getting a lot of attention in "Are Server SANs The Future Or Just Hype?"]
If you want innovation at minimal risk of your vendor going out of business before your equipment reaches a ripe old age, you need to look at the upstarts. These are startups that have been successful not only with their technology but also in selling the product. While identifying upstarts is more an art than a science, there are a few indicators I use.
The first is that the company has completed a successful IPO. As the fiasco at Violin Memory demonstrates, just going public isn’t enough to make it to upstart status. The company also has to demonstrate that it’s continuing to grow sales. Nimble Storage has clearly moved into the upstart category, having raised $176 million in its IPO. In Nimble's first post-IPO quarterly filing, it reported growing sales with losses of $13 million for the quarter. At that rate, the company will have enough cash on hand to cover losses for three years, almost guaranteeing that it’ll be around for the three-year life of an array in your data center.
The other way a company can make it to upstart status is to raise a large round of additional funding that includes all its existing VCs. This pre-IPO round is typically $75 to $125 million. Once again, Violin is the exception that proves this rule and for me to call a vendor an upstart based on funding, I’m also going to have to believe that its technology is better than what the big vendors offer. Violin’s problem was, in part, that it was selling raw speed while the market was moving to more full featured all-flash arrays from Pure Storage, Whiptail and most significantly, EMC’s XtremIO. Today, Pure and Tintri have made it to my upstart list based on VC confidence, announcements of growing sales, and innovative technology.
I would advise any technology buyer to pay close attention to upstarts such as Nimble, Pure and Tintri. They offer a unique combination of innovation and security.
Disclaimer: Pure Storage is a client of DeepStorage, LLC and Tintri has provided a storage system for DeepStorage labs.