The existential battle facing legacy storage vendors is about business models, not technology or features. Market forces – cloud providers and SMBs – are trying to turn a high margin business into a much lower margin business.
We have already seen this happen with the servers. Many large minicomputer companies enjoyed 60 to 70% gross margins for years.
But then the PC – the new minicomputer – came along with much lower gross margins and much higher volumes. As CPUs, networks and software improved, the PC evolved into the server business we have today.
In the process, many technically superior products– such as Decnet and MVS – were pushed aside. Their functions were commoditized, mass-produced and plummeted in price – and gross margin.
Storage is an application
In the last 10 years storage companies have moved most of their functions to software. They’ve made storage an application.
In the 1970s and through the mid-80s several companies built large businesses based on word processing. They packaged software and hardware with support and training and found a ready market in offices worldwide.
Then PC word processing software started catching up with the proprietary word processing systems. Early packages such as Wordstar were hopelessly primitive compared to high-end Wang systems, but they evolved and in a few short years Wang was dead.
The StorageMojo take
Server technology is evolving faster than storage controllers can. The traditional storage engineering model of tightly specified controllers running well-tested firmware is getting less viable each year.
Scratch most storage array controllers today and you’ll find a barely disguised x86 server decorated with low-volume/high-cost redundancy bits. Scale-out storage with many redundant storage servers can be just as reliable and performant as traditional storage.
Fortunately for legacy vendors, most customers don’t know that. Many mid and upper level IT managers are comfortable with the devil they know and resist change.
But as AWS, Google and Azure are showing, large-scale cloud infrastructures can handle most enterprise jobs today. And they keep getting better.
The upshot is that most storage capacity will become a 30% gross margin business in the next 10 years. There will always be specialized applications that require high-margin gear, but that won’t be the bulk of the market.
Look at servers today. That is storage in 10 years.
Courteous comments welcome, of course. Because of the critical nature of storage, software only vendors will have to step up their quality to achieve wide penetration. How do you think they can best do that?
Good piece Robin. While I agree that legacy storage vendors are facing a battle I’m not sure it’s only one of business models. The cloud is certainly the change agent for much of what IT does, but I still see technology, primarily the microprocessor, as the enabler of this change. Here’s a short piece with more details – thanks for starting this conversation.
http://storageswiss.com/2014/12/05/storage-evolves-thanks-to-processor-not-business-model/comment-page-1/#comment-44938
Absolutely agree – been struggling to get people to grapple with this issue for a while. We can’t afford to keep the data we are generating on enterprise disk arrays. I’m keen to try (experiment) with alternatives – but the usual response I get is “who’s going to support it?”
Support…. I suspect a lot of outsourced “support” is now worse than internal support once was. The idea is attractive – buy in expertise from an external group who support one product across many customers – lower overheads, better knowledge base, etc. I suspect the next step is that those support companies find it very attractive to reduce costs (less people, less expertise) and take on more products until the level of specific expertise drops to a similar level to that which used to exist in house…. Only now they don’t have any business knowledge of your business either….
Some of the things that storage vendors can do to maintain their margins:
1) Make the storage mgmt. software on linux user-friendly. Ever seen what linux sysadmins have to do to make Dell Equallogic work? Not fun. Might as well go Google “pizza box” or Blackblaze – less painful actually.
2) Look at how Cisco is falling down in networking and do the opposite in storage. Cisco has failed cloud providers in mgmt. software, 10 GB port pricing and bandwidth fan-out loss. Storage vendors should look at their offerings with a critical eye.
3) Just adding a Swift API would add 5 years to most storage product lines and get the whole company thinking “Cloud,” similar to Microsoft’s Internet eye-opener in ~2000.
James B.
Curiosity. Mentioning DECNet and MVS together seems an odd combination. Did you mean DECNet and VMS or were you trying to include DEC and IBM together in a group of pushed aside technically superior providers?
Charlie,
Yes, I meant DECnet and MVS.
Say what you will about IBM’s tactics, MVS became a very solid OS by virtue of being beaten on – and fixed – for decades. People will pay a steep price when they’re locked in, but when an almost free alternative comes along many will jump on it.
That’s what Ken Olsen, DEC’s founder, failed to see: much cheaper and less reliable alternatives that are mostly “good enough” suck the value out of your high-priced, feature-rich, and more reliable proprietary products. It’s a lesson most legacy vendors are learning now.
Robin