First, a mea culpa: I was one of the handful of people who closely followed the PaaS world a few years ago.
For those not in the know, PaaS (platform as a service) was all about providing a development and automation solution for application developers so that they didn't need to think about managing infrastructure themselves. The initial PaaS offerings were Heroku (acquired by Salesforce) and Engine Yard (once a high flier but now, at least by my assessment, languishing in obscurity).
Thereafter the world of PaaS moved on a little with two distinct camps, Red Hat OpenShift on one side (itself the product of an acquisition as Red Hat scooped up Makara) and Cloud Foundry, a product initially incubated within VMware but now part of the Pivotal world.
In the battle between those initiatives, OpenShift and Cloud Foundry, Cloud Foundry seemed to have all the momentum with Pivotal itself -- not to mention IBM, HPE and others jumping on board the open source project.
Fast-forward to today and PaaS has somewhat been subsumed into a far broader application automation story. Both Pivotal and Red Hat are strongly articulating their products as container automation/orchestration offerings and less about PaaS offerings as a distinct segment.
That move looks to have been concreted further with the launch of Red Hat's latest version of OpenShift, a product that is no longer called a PaaS but rather a container platform. Further to that, Red Hat strongly articulates the fact that it is a leading contributor to both the Docker and Kubernetes projects. Docker is, of course, the open-source initiative that (re)-popularized the usage of Linux Containers, while Kubernetes is, for its part, another open-source initiative that is a direct descendant of Borg, the systems and tools that Google uses to run its global infrastructure.
Anyway, moving on from the abundance of different open-source projects, this latest version of OpenShift claims to be both an enterprise-ready version of Kubernetes 1.4 and also the Docker container runtime. The inference is that Red Hat has capitulated to broader initiatives around application management platforms and accepted that it needs to play far more broadly in a container-aware world.
In terms of the particular features, Red Hat is trumpeting the fact that this latest version of OpenShift offers:
-
Next-level container storage with support for dynamic storage provisioning, allowing multiple storage types to be provisioned, and multi-tier storage exposure via quality-of-service labels in Kubernetes. Container-native storage, enabled by Red Hat Gluster Storage, enhances the user experience running stateful and stateless applications on Red Hat OpenShift Container Platform by simplifying the consumption and provisioning of application storage by developers. With Red Hat Gluster Storage, OpenShift customers get the added benefit of a software-defined, highly available and scalable storage solution that works seamlessly across on-premises and public cloud environments and one that is more cost-efficient than traditional hardware-based or cloud-only storage services.
-
Enhanced multi-tenancy through simplified management of projects, a feature powered by Kubernetes namespaces, in a single Kubernetes cluster. Multiple developer teams, applications and life-cycle environments can already run fully isolated and share resources on a single Kubernetes cluster in OpenShift Container Platform. Red Hat OpenShift Container Platform 3.4 adds the capacity to search for projects, project details, manage project membership and more via a streamlined web console, making it easier for users to work with multiple projects across dispersed teams. These multi-tenancy capabilities enable enterprise IT organizations to provide application development teams with their own cloud-like application environment to build and deploy customer-facing or internal applications using DevOps processes that are completely isolated from one another.
-
New hybrid cloud reference architectures for running Red Hat OpenShift Container Platform on OpenStack, VMware, Amazon Web Services (AWS), Google Cloud Engine and Microsoft Azure. These guides help walk a user through deploying a stable, fault-tolerant, production-grade environment that uses the power of Red Hat OpenShift Container Platform across public and private clouds, virtual machines and bare metal.
Indeed, Ashesh Badani, vice president and general manager of OpenShift at Red Hat, seems to wave a white flag when it comes to containers, albeit with a "we do it better" spin.
"While Linux containers represent an innovative future for enterprise applications, traditional and legacy applications remain critical to the modern business," he said. "Red Hat OpenShift Container Platform 3.4 meets the needs of these existing applications while providing the tools and services necessary to drive cloud-native application creation and deployment. The latest version of our flagship container application platform goes a step beyond simply creating and deploying applications by addressing the growing storage needs of both stateful and stateless applications across the hybrid cloud, allowing for coexistence of modern and future-forward workloads on a single, enterprise-ready platform."
Red Hat would seem to be struggling to find a compelling point of differentiation now that every man and his dog is defaulting to the "we make it easy to orchestrate containers" story. In its case, Red Hat suggests that the fact that its container portfolio spans private and fully managed cloud offering and thus supports more traditional applications, as well as cutting edge cloud-native ones, makes its offerings stand out. Whether you buy this line all comes down to your particular view on the world.
Either way, the world has very much moved on from the days when a select number of analysts, myself included, suggested that PaaS was the future of enterprise IT. It now seems that PaaS has been very much superseded by new ways of doing things -- from containers to serverless and that the very acronym looks set to be forgotten.
Ah well, there are plenty of other acronyms that can take its place, I guess.