People surprised by Red Hat's announced adoption of CentOS haven't been paying attention. While Red Hat early on felt competitive pressure from its clone, more recently it has tended to smile and nod at CentOS, considering it a valued member of the Red Hat Enterprise Linux (RHEL) communit and a "mindshare force multiplier" of the RHEL brand, as Ryan Paul notes and as Red Hat CEO Jim Whitehurst told ReadWrite months ago.
By officially embracing CentOS, Red Hat has co-opted what could have been a sales distraction.
CentOS The Innovator?
Indeed, the only surprising thing in the joint press release is the length to which Red Hat and CentOS go to make us forget that CentOS is a RHEL clone, nothing more. For example, Karanbir Singh, lead developer of CentOS, exults that "Now...we are able to count Red Hat among the active contributors to the CentOS Project," conveniently omitting the fact that Red Hat for years has contributed roughly all of the CentOS code.
It's called "RHEL." Perhaps the CentOS folks have heard of it?
The differences are negligible enough that former OpenStack executive Andrew Clay Shafer quips that all CentOS has really done is change the splash screen, not the underlying code. Still, on the FAQ for the community merger, Red Hat highlights early and often the additional testing and certification it invests in RHEL and will not give to CentOS.
The fact that Red Hat is bringing the popular RHEL clone into the fold doesn't mean it's giving up on its business model. To recap, that is to charge risk-averse enterprises to run a safe, secure and, frankly, boring enterprise-class Linux. CIOs don't want to have to think about the quality and pedigree of their server operating system.
Keep Your Friends Close And Your Enemies Closer
The problem is that not all enterprises are equally risk averse. CentOS is by far the most popular non-paid Linux distribution for enterprise use. While Red Hat lists several hundred case studies of enterprises like Goldman Sachs that depend upon RHEL for mission-critical deployments, there are many others that want the safety of running a close approximate of the RHEL standard but don't want to pay RHEL prices.
Or as Kenn White, principal scientist at Social and Scientific Systems, puts it: "Let's be honest—the main driver of CentOS adoption is avoiding $350/box/yr [for a] Red Hat license."
Red Hat could have kept trying to fight CentOS gravity, as it did before by "shipping Linux kernel source code in a big tarball with the patches already applied, making it more difficult to build Linux distributions from the RHEL source," as Ars Technica describes in its history of Red Hat. Or Red Hat could do what it did, which was to embrace the parasite, or "adjacent ecosystem," as Redmonk analyst Stephen O'Grady terms it.
The former would have been akin to Red Hat constantly trying to plug holes in a leaking dam. The latter is a recognition that there's more to be gained by enlarging the official Red Hat ecosystem than trying to bottle it up. As Whitehurst told ReadWrite back in August:
CentOS is one of the reasons that the RHEL ecosystem is the default. It helps to give us an ubiquity that RHEL might otherwise not have if we forced everyone to pay to use Linux. So, in a micro sense we lose some revenue, but in a broader sense, CentOS plays a very valuable role in helping to make Red Hat the de facto Linux.In other words, this was a move of unadulterated pragmatism, not some beatific vision of "pooled innovation" touted by Red Hat in its press release.
The Real Battle: Cloud
And maybe, just maybe, this is a recognition that rooming with CentOS won't lose Red Hat any money in server sales and could position it for success in the cloud. Former Ubuntu strategist Simon Wardley calls Red Hat's decision a "smart move" and notes that it makes Red Hat very competitive with Ubuntu for guest OS deployments on AWS and other public clouds.
Competitive for developers, maybe, but not for dollars.
Very little money is being made on operating systems in the cloud. No one necessarily wants to pay for an OS when the cloud itself is the OS. Red Hat's strategy is to offer an easy on-ramp to the cloud through OpenStack and OpenShift, and also offers a range of related services like KVM/RHEV (virtualization), JBoss (middleware) and more to help enterprises at any stage of cloud adoption. By broadening its RHEL on-ramp to also include CentOS, Red Hat has effectively embraced a far larger community than subscriber-only RHEL could provide.
Could Red Hat have continued to foster this strategy with CentOS at arm's length? Probably. But it's much more effective to be able to incorporate the clone into the host, scorching the earth for all would-be Linux competition that might want to attack Red Hat's weakest point: developer community.