Taneja Group Analysis
I’ve been immersed in “Open” for the last two weeks here in Boston, attending both Red Hat Summit 2017 and then OpenStack Summit. There are quite a few things worth paying attention to, especially if you are an enterprise IT shop still wondering how your inevitable cloud (and services) transformation is really going to play out, including accelerating application migration to containers and the rise of platform Management as a Service.
First, let’s talk about what’s new with Red Hat (RH). As we’ve noted before, RH is an impressive company. They’ve so far nailed a successful open source business model. And they’ve successfully navigated through the ever churning waters of technology and market evolution that have left many (most?) open source companies floundering.
For the last few years I’ve been wondering where RH goes beyond RHEL and JBOSS. RHV (KVM) did compete for virtualization dollars, but really it mostly helped commoditize the hypervisor (meaning less money to be had in this space). And OpenStack has seemed, if not exactly dead in the water, at least looking like a science project suitable for the only most dedicated adherents (you know – big SP’s and Telcos).
BTW - 1st Rule of Prognostication - Never pronounce anything that hasn't yet come into its own as "dead" as most of the eventually best things start out with complex implementations - they just need to slowly climb the knee of their fated maturity/adoption curve...
But Red Hat has now developed a more cohesive, end-to-end vision that promises great things. With the rise and embracing of containers (and Kubernetes) in their Openshift platform, they now have a nicely complete span of solutions from RHEL (Linux) to RHV (KVM) to OpenStack (cloud) to OpenShift (container/devops) (although as a I’m still on the fence as to whether OpenStack is easy enough yet for rank and file enterprises to really deploy and manage at-scale).
Clearly, RH OpenShift is their latest big thing – it’s basically a container-oriented PaaS that layers on RH OpenStack (private) or on any of the big public clouds to provide a seamless application host layer. Openshift.io is RH’s own cloud service, but it’s all open source so you can run it anywhere you want. As we tweeted at the show (follow me at @smworldbigdata), RH intends OpenShift to commoditize cloud infrastructure (make it a level playing field) just like RHEL helped commoditize all those old school proprietary server stacks.
I’m not sure that cloud providers are all going to be happy about the leveling of their playing field as it removes potential differentiation, but at least they are all playing. As one of the keynote demos at OpenStack Summit (by CoreOS and CockroachDB – yes, it’s hard to kill), 15 or so world-famous cloud providers each provisioned an OpenStack cluster, deployed Kubernetes, and then launched containerized database nodes and joined them live together into one cross-provider, globally distributed SQL database – in about 10 minutes.
(*Note – Yet, previously on stage a simpler demo of DIY/MIY OpenStack Cinder with NetApp SolidFire fell flat on its face.. Hmm.)
RH is claiming strongly that “containers are linux” since the container host is the kernel side of Linux and a container is the user space side of Linux. They claim the whole container space is just Linux split in half. And therefore they are the only fully capable/scalable source for container hosting (OpenShift) because they are Linux experts. Well, maybe. I think containers are going to be much more than Linux (or depend much less on Linux depending on your point of view). Still, Linux expertise does give them running room in this space.
It does remain to be seen who will dominate the enterprise management of what is fast becoming a hybrid, multi-cloud, cross-platform, global, increasingly app/services focused, and incredibly dynamic IT span of responsibility. VMware has been setting crosshairs on cloud management now for a couple of years, and gets a boost spearheaded by such “lucky” acquisitions as NSX (you can’t tell me that they really thought back then that a security “network” add-on would supplant their server virtualization core focus).
Which brings up the other hot trend we note at these shows around “Remote Management as a Service” (what we’ve simply called “Management As A Service” or IT MaaS). I’ll be writing even more about this in upcoming reports, but as we’ve previously posted, if IT can subscribe to “remote” SaaS delivered expert infrastructure and platform management, while keeping the actual hardware (and their data!) onsite, they can then focus on adding value to the business at a higher level – working on applications and services that advance the business.
We’ve long admired Platform9’s innovation and leadership here, going way beyond glorified “call home” to actively host and provide MaaS, and now we see indications that everyone is talking about or actually getting into the management services business.
Like when IT finally recognized the eventual costs and constraints (and risks) of DIY – Do It Yourself, the time has come its seems to get away from MIY – Manage It Yourself.