So you’ve founded a new storage business; you’ve got a great idea and you want to disrupt the market? Good for you…but you want to maintain the same-old margins as the old crew?
So you build it around commodity hardware; you use the same commodity hardware as I can buy off the shelf; basically the same disks that I can buy off the shelf from PC World or order from my preferred Enterprise tin-shifter.
You tell me that you are lean and mean? You don’t have huge sales overheads, no huge marketing budget and no legacy code to maintain?
You tell me that it’s all about the software but you still want to clothe it in hardware.
And then you tell me it’s cheaper than the stuff that I buy from my current vendor? How much cheaper? 20%, 30%, 40%, 50%??
Then I do the calculations; your cost base and your BoM is much lower and you are actually making more money per terabyte than the big old company that you used to work for?
But hey, I’m still saving money, so that’s okay….
Of course, then I dig a bit more…I want support? Your support organisation is tiny; I do my due diligence, can you really hit your response times?
But you’ve got a really great feature? How great? I’ve not seen a single vendor come up with a feature that is so awesome and so unique that no-one manages to copy it…few which aren’t in a lab somewhere.
In a race to the bottom; you are still too greedy. You still believe that customers are stupid and will accept being ripped off.
If you were truly disruptive….you’d work out a way of articulating the value of your software without clothing it in hardware. You’d work with me on getting it onto commodity hardware and no I’m not talking about some no-name white-box; you’d work with me on getting it onto my preferred vendor’s kit; be it HP, Dell, Lenovo, Oracle or whoever else…
For hardware issues; I could utilise the economies of scale and the leverage I have with my tin-shifter; you wouldn’t have to set-up a maintenance function or sub-contract it to some third party who will inevitably let us both down.
And for software support; well you could concentrate on those…
You’d help me be truly disruptive…and ultimately we’d both be successful…
Hi Martin – Rob from Tegile here. I’ve been (and am now) at a few of the companies you are challenging here. Some have been wildly successful, some not. What you seem to be asking for is what Nexenta has tried, correct? We have several customers that have taken that bike out for a ride, and all have come to the conclusion that “roll your own” storage is really easy until something breaks – then they are screwed. They realize that all the testing and driver work that we do under the hood to make “our nebulous value add” strong enough to stick with the model so many of us are operating on.
Keep pressing though! The feedback you post here is heard and is materially used in our product planning discussions.
Rob
You see there is really a rather limited amount of variation now in the Intel stack…and already we are seeing some moves to vendor’s shipping virtual arrays on top of a hypervisor which can help to mitigate some of the driver differences. And then we start to see products like ScaleIO for example which do run on commodity…although their licensing costs are quite simply idiotic at the moment. And then RedHat could actually start to play properly…IBM could have a software moment…It’d be funny if the oldest elephant in the room did something quite so disruptive. Or Microsoft with SMB 3….
Perhaps offer a choice…you can have hardware appliance or you can run on a certified stack. You could be fairly proscriptive if you wanted..but I suspect you won’t…because you all are addicted to selling tin.
Perhaps what we want is some more useful Open Compute reference architectures that address primary storage rather than large-scale or cold archives. If we had that, then the value of one software implementation over another might more easily be measured.
Most hardware is pretty much identical now so the differentiation is gone.
However, Rob, as you say, your value is in things like component/driver testing or other validation. I’d say that still counts as software and still has massive value.
Ultimately, the problem is convincing end users to pay for software. Bundling the hardware does make the whole solution more palatable, but conversely it does stick in my gullet when I see a consumer SATA drive in an enterprise VMAX and know how much markup EMC are making.
Chris
Good discussion Martin. I think you’ve encapsulated many of the thoughts going through buyers minds right now when it comes to software/storage mix.
I think I can provide at least one unique insight here. A number of SolidFire’s early engineers and managers came from LeftHand Networks. As you may be aware, over the course of their life, LeftHand sold their system *every* possible way:
-As a storage appliance on branded hardware
-As software on qualified OEM hardware (HP, Dell)
-As software only via the channel to bundle on hardware
-As a VSA to run on any hardware
We asked these folks the lessons they learned from this approach, and here’s what they said:
1. The qualification time to test the software on multiple platforms is significant, and hugely slows down development cycles.
2. If you don’t do this testing (and just “assume” that any x86 box will work), you run into extremely complex reliability and performance issues that are difficult to troubleshoot. Customers just don’t want to accept those type of issues on a storage system – it’s too important.
3. Perceived cost savings (by letting customers buy their own hardware) aren’t actually there. Just because a customer buys the hardware doesn’t mean that a storage vendor’s software should/can/does cost any less. If you want to know why something like ScaleIO “seems” expensive, it’s because you’re being exposed to the real cost of the software that ships on complex storage systems. Don’t be surprised if VMWare’s vSAN “seems” expensive as well.
In the end, what we learned is that there are only two viable models for shipping a storage system today:
1. In an appliance model with fully qualified hardware (and ideally a single platform, to minimize qualification time)
or
2. In a VSA form factor (where you are qualifying the virtualization platform), although you can still have some complex performance issues depending on what hardware they run. This is a particularly keen problem in the flash space, where the difference between two similar looking drives (from a spec sheet) can be completely night-and-day in actual operation.
There is a perception in the storage world that vendors “mark up” their commodity hardware tremendously. Some vendors haven’t done any favors to this perception when they do things like sell replacement drives for 500% markup or $20 cables for $600 each, but I think you’ll find that really only exists in the large, legacy vendors. Most storage startups these days have simple, flat rate pricing.
Once you get past some of the line-item silliness of legacy vendors, in the appliance model where you are buying bundled hardware/software, the “mark up” is really just the price you are paying for the storage software – the real IP in the storage system. If you aren’t getting a corresponding value from that IP, over and above the actual collection of tin, then don’t give then vendor your business.
Be honest about it then…show the cost of your software versus the cost of the hardware. Break it down…
That’s a good challenge, and may be more common in the future (rumor has it Netapp may start doing that).
From my perspective, there are two main reasons that isn’t commonly done today.
First is accounting rules. When you separate software pricing from hardware you have to account for them differently than when they are bundled. In many cases, this may mean you can’t recognize all of the software revenue up-front and have to defer part of it or recognize it over a period of time. The rules on this are complex, but they’ve been known to bite companies (if you remember the Apple-charging-for-ipod updates issue, it was related to this). At worst, it can make you have to restate earnings if you get it wrong (which can be devastating), but even at best it can mean you can’t recognize your full revenue number, which is very problematic, particular for startup’s whose growth and valuation is based heavily on GAAP revenue.
The second reason is simply that customers have been taught to expect a steep discount on every line item. Breaking out the hardware means you either to have to mark it up silly (only to discount it back down) or convince them that they really shouldn’t get discounts on it (good luck). They also think they should get huge discounts on software, “because it doesn’t have any cost” (which couldn’t be further from the truth).
This complicates pricing discussions that should in the end be pretty simple for storage: For the functionality and capabilities I’m providing, what is a fair price (per GB, per IOP, or on whatever metric you want to gauge it). Having a single bundled price that represents the “entire” system makes pricing discussions much easier.
If anything, I think storage buyers appreciate simplicity and clarity (note: not the same as transparency) in the quoting and pricing process above all. While a typical vendor storage quote may have 30+ line items, between all the individual hardware components, cabling, software enablement, etc… a typical SolidFire quote has two line items:
1. SolidFire Node (type, qty)
2. Service & Support (1-5 years)
Love this post, Martin — but you knew I would!
I have my own opinions about all the newer storage hardware startups, but it’s great to see your perspective on all of this as well. You’ve chosen to focus on the practical aspects: cost, support, etc. which is all quite valid. I think where we can go farther is to use software-based storage to deliver new operational models as well, something I’ve blogged incessantly about …
I found all the “here’s why it’s hard” comments interesting, and there’s a certain validity to some of it. That being said, now being part of VMware, I think we’ve got a good approach to establishing and managing HCLs for VSAN, built on top of the extensive work VMware has already done for vSphere et. al. Time will tell.
Keep blogging!
— Chuck
At times I wonder that any of our compute stacks work at all you know because as far as some storage vendors are concerned they should be broken all the time. I think we’d all accept HCLs and if we go off piste, it’s at our own risk. With a lot of start-ups and larger vendors, I know exactly what components are being utilised; from the SuperMicro chassis to the Emulex card to the base Linux version…
Yes, there is some value in qualifying the components and integrating them. But I don’t think it’s quite as high as some like to pretend.
Noam from Zadara Storage here. I call red herring. While some vendors argue about whether we as an industry should be selling tin or floppies, others aren’t selling anything. They provide storage as a service, even on-premise (yes, private, on-premise managed arrays, with flexible, recurring payments). It’s happening sooner than later, even for the traditional box guys: http://blog.architecting.it/2013/10/04/what-is-emcs-project-nile/
Chris was speculating as to what Nile is. I think he is actually some way off in the speculation but I’m sure more will be revealed as the year progresses. The ‘as a service’ model does not fit all companies’ requirements or indeed their financial models. And all vendors are selling something…it’s a definition of being a vendor.
Michael Williams from CloudByte weighing in…
Martin, your customer viewpoint is well articulated! By the way, you are not the long voice for this line of thought. CloudByte’s storage software overcomes the elephant-in-the-room issue of having performance for everyone suffer due to a single ‘noisy neighbor’ customer or application. Our secret sauce – so far I haven’t seen it copied by other software vendors – is ensuring that every tenant or application gets dynamic, predictable performance from commodity hardware.
Our software runs on commodity storage (Dell, Super Micro, DataOn, …) – as you indicate is the goal. This isn’t theoretical – we include the actual Bill of Materials for Dell hardware in our Tiered Reference Architecture document, freely available at http://www.cloudbyte.com/downloads/resources/.
Through testing, we make sure that our software works on commodity hardware components that can be used in a broad range of configurations – much as VMware is able to do great things without having to bundle with hardware.
And yes, our customers are innovating new, disruptive services with our software – we’ll have more news on that in the near future.
Thanks again for your refreshing take on storage!