Amid rumors that Red Hat is now or plans to strong-arm customers to use its own flavor of OpenStack, history tells us one thing: They won’t be the only one trying to make the open source cloud technology their own.

My guess is that if OpenStack becomes really popular it will look less-and-less like an open source project and become semi-proprietary software with significant vendor lock-in. That means there will be cooperation on the foundation but the closer OpenStack gets to customer dollars, the more it will diverge.

Our Mike Wheatley reports on the rumor, which Red Hat has denied in a wishy-washy sort of way. My take is that if it’s not true now, it likely will be in the future. How hard Red Hat will push is the open question.

Steering customers their way

 .

There are legitimate and dodgy reasons why Red Hat, or any vendor, might steer OpenStack customers to their own stuff. The legit reason is support. It is difficult to impossible for vendors for assure that an OpenStack installation procured Chinese menu style (one from column “a”…) will function reliably. Vendors are also proud of their own work and their ability to build (hopefully) compelling value into their own products.

Of course, Red Hat wants its customers to use the version of OpenStack it knows best — its own. Presumably every OpenStack vendor will do this to some extent. This is where open source can really break down. (Android, I’m talking to you).

• Customer insight

Further, I don’t believe customers want a generic OpenStack installation as much as they want one that a) does cool things and b) actually works in their environment. They also want to save money, which may be why KVM is the overwhelmingly favorite hypervisor among OpenStack users, trouncing VMware technologies. It’s almost like the old shampoo commercials, “KVM does what theirs does, for a lot less.”

That and some other interesting statistics appeared in the results of an OpenStack user survey, released last week.

Perhaps the most telling finding is that Ubuntu is the preferred operating system, with CentOS in second place. That Red Hat isn’t number one suggests OpenStack customers are still in an experimental stage and may move to Red Hat as they “get serious” about cloud technology.

That’s a big opening for Ubuntu and ought to scare Red Hat. The issue won’t be so much converting current installations, but attracting new customers getting into OpenStack for the first time or expanding its usage in their companies.

Balancing lock-in with customers’ best interest

 .

That leads vendors, naturally, to “how do we create customer lock-in that customers will tolerate?” Now, I am not accusing anyone, but I’d be selecting a limited number of configurations for premium support while adding features of my own that are compelling but not compatible to OpenStack itself, perhaps as add-ons.

In deciding what to support, I’d want to cover what my current customers are doing and who their existing partners are, along with any companies that just can’t be avoided, unless there were compelling reasons for doing so.

So who would I lock out? VMware if I could, but I might try to partner with Microsoft or IBM where possible. Those partnerships could create the aura of openness without opening up to the whole world.

Schemes like this are what product managers and technical marketing people spend their days dreaming up and turning into PowerPoints for the C-suite.

Open source is a valid response to the bad old days when IBM controlled enterprise computing with little opposition. Open source and Linux have made it possible for vendor companies large and wanna-be large to play on a more open playing field. This has been a big win for customers.

Yet, the desire to lock-in customers remains strong. The trick for them is not getting too locked-in and getting best-of-breed technology and pricing concessions as part of the deal.

Let Red Hat, be Red Hat!

photo credit: slimmer_jimmer via photopin cc