Asymmetric innovation: legacy vs cloud

by Robin Harris on Wednesday, 9 March, 2016

Here’s a good question: why are cloud vendors able to innovate so much faster than legacy vendors? AWS brags about the hundreds of new features and services they implement every year – and their accelerating pace.

But here’s a better question: why don’t enterprise buyers accept the same level of innovation from legacy vendors as they do from cloud vendors?

The legacy innovation trap
Legacy vendors have always faced innovation competition from startups. That’s why buying startups is a favored way to bring innovative ideas and people into large companies.

But the cloud is something different: they aren’t for sale. They’re replacing legacy vendors.

Several currents are sweeping up the unhappy legatees:

  • The market is shrinking. Demoralizing margin and R&D spend pressure.
  • Years of FUD marketing have trained customers to fear storage innovation.
  • Vendors have sizable bureaucracies devoted to slowing innovation vetting innovation. At Sun, for example, even the most minimal product change took 6 months to work through all the hoops. God help you if had an original idea!
  • The geo-scale of the cloud vendors and their uptime and data recovery records have overcome the fear of cloud innovation.
  • Much of the best talent wants to work at cloud vendors, not IT companies.
  • And, of course, the cloud folks have total control over their infrastructures, from the details of server design and config, to the OS and all apps.

The last point explains why they’ve been able to beat many enterprise datacenters for uptime and availability.

The StorageMojo take
The optimistic take on current legacy vendor troubles is that the new technologies and products they’ve brought to market just need to be given time to grow. Eventually they’ll take up the slack.

EMC’s DSSD and XtremeIO, and NetApp’s SolidFire buy exemplify this thinking. Sorry, no.

Why? The array market is in a permanent secular decline. High performance flash owns the OLTP business, and is, in turn facing competition from in-memory databases and, soon, even larger non-volatile memory systems.

What a few vendors – such as HDS, Nutanix, Spectra and HP – realize is that there is another way to compete with the cloud: price. It is possible to beat the cloud on price with scale-out commodity object storage systems.

Right now they’re focused on archiving, but I expect that these object storage system will increase in performance over the next decade to the point where they will be practically indistinguishable from current NAS boxes – only cheaper and much more scalable.

One thing that would help businesses see the advantages of object storage would be to ditch the name. How about “Super-scale NAS”? Its worked for Ethernet and USB.

Courteous comments welcome, of course.

{ 3 comments… read them below or add one }

Bill Strugger March 10, 2016 at 5:22 am

Legacy vendors have generations of Install Base that they have to continue to serve. Innovation is easier if you can start with a blank white board.

Jerry Leichter March 12, 2016 at 6:23 am

Installed base is the traditional explanation/excuse. It works when you’re looking at startups bringing out a new technology to compete with existing large-base old-technology vendors. It especially works in the classic disruption model where the startups target small, low-end customers who the existing vendors barely even notice, and the move up scale later.

It doesn’t work at all to explain AWS and Google and their like. These are massive vendors who already have a great deal of installed base – and all the things that go with it, like massive locked-down investments in hardware and software and technologies. And yet they *still* manage to innovate rapidly – and their installed base of customers is willing to change along with them.

There are many factors involved, and I won’t pretend I have a full explanation. An obvious element is capital investment: In the old world, you’ve paid your vendor for a bunch of equipment and you’re damn well going to squeeze every bit of value out of your sunk cost. In the cloud world, it’s the cloud world, the capital risk is all on the cloud vendor – you pay for what you use as you use it, and can (in terms of investment) walk away free and clear at any time. And on the other side of the transaction, the cloud vendors are willing to spend big, and keep spending big, on hardware. Those that can’t or won’t are left behind by the competition. (It’s also why the underlying vendors of “computes” and “cloud storage” are all giants. No one’s out there starting an AWS competitor in their garage.)

Another factor is the emergence of higher-level standards. At one time you couldn’t move from on architecture to another without rewriting all your (assembler) code. Then you were locked in by various system-unique programming language. Then by operating systems. Today, the only thing unique to you is your business logic, expressed in a programming language that’s available everywhere. The database stuff, the Web interface, failover, scaling, the glue that holds all the pieces together – that’s all standardized (to one degree or another). You can move with relatively small investment.

*And the cloud vendors know this.* It allows them to change tons of stuff internally without affecting any of their customers – and it also allows them to decide to improve (aka change) customers actually *do* depend on because the cost for those customers to adjust is reasonable.

The net effect of all this is a system in which the cloud providers, while huge, not only want to and can but *must* act like startups. Rather than viewing their installed base as a drag on their ability to change anything, they see it as willing to adjust and *wanting* to adjust. It’s not that users of traditional computing *wanted* stasis; they (or their customers, in any case) *wanted* flexibility and responsiveness. But the cost structure and technological basis (and the management styles that those produced) made that very difficult to attain. Many of the risks and costs of such flexibility can now be pushed on to the cloud vendors – and they, in turn, have figured out how to deal with them.

— Jerry

Robin Harris March 13, 2016 at 9:33 am

Jerry, excellent points. I’m hoping to delve deeper into the unwitting complicity of the legacy IT vendors in making the cloud vendors successful.


Leave a Comment

Previous post:

Next post: