The road ahead for the Red Hat OpenStack Platform

If you didn't have a chance to attend our Road Ahead session at Red Hat Summit 2019 (or you did, but want a refresher) you'll want to read on for a quick update. We'll cover where Red Hat OpenStack Platform is today, where we're planning to go tomorrow, and the longer-term plan for Red Hat OpenStack Platform support all the way to 2025. 

A strategic part of our portfolio

Red Hat OpenStack Platform is a strategic part of Red Hat's vision for open hybrid cloud. It's the on-prem foundation that can help organizations bridge the gap between today's existing workloads and emerging workloads. In fact, it just earned the 2019 CODiE award for “Best Software Defined Infrastructure.”

One of those emerging workloads, and more on the rest in a moment, is Red Hat OpenShift. 

How OpenShift and OpenStack fit together

Red Hat OpenShift Container Platform is our enterprise Kubernetes platform to build and deliver better applications, faster. As such, it's an important workload for Red Hat OpenStack and is used by some of the more than 1,000 customers running OpenShift to offer them faster deployment, autoscaling, and faster disaster recovery. 

While we think OpenShift is the best answer for container native workloads, we know that many of our customers are going to have traditional workloads that are better suited for OpenStack today. We want to help those organizations with a platform that's suited to supporting today's workloads and tomorrow's. We think OpenStack underlying OpenShift is the way to manage both without having to compromise. 

OpenStack provides agility and the ability to scale a range of workloads from bare metal to virtualized  as well as manage and scale OpenShift too. OpenShift can bring the cloud-native open hybrid cloud functionality that customers need, which can help them bridge workloads from their private cloud to OpenShift on the public cloud. To make this easier for our customers to deploy, we just released a new reference architecture (you can read more about it in its announcement blog). 

Use cases for Red Hat OpenStack Platform

In addition to running OpenShift, Red Hat OpenStack Platform is well-suited to some of today's popular workloads like Network Function Virtualization (NFV), artificial intelligence and machine learning (AI/ML), high performance computing (HPC), edge computing, and for service providers. 

Service providers have looked to OpenStack to provide management and scalability while still giving direct access to hardware acceleration to boost applications. Reducing latency and managing large volumes of data are important for providers who offer services that are very sensitive to delayed response. Who wants to watch choppy video on their cell phone, or have to wait ages to upload pictures to social media? 

Our NFV specifics extension provides support for vCPU pinning, NUMA-aware scheduling, and accelerated packet processing so that organizations can take advantage of their hardware's capabilities. 

While the public cloud trend has been centralizing services, we see service providers and other organizations looking to edge computing to help bring services closer to their customers. Again, latency matters for many use cases, and organizations offering Internet of Things (IoT) services, robotics, and telco services can benefit from having edge computing services closer to the customer.

By putting processing power at the "edge" of the network, applications can have lower latency and have compute closer to the data source. 

There are many edges, also known as distributed computing, hybrid edge computing, datacenter in a box, and my favorite "fog." (You know, a closer cloud.) What the edge looks like varies by organization and use case. We've designed, and continue to improve, OpenStack to help organizations, like telecommunications providers, scale out to the edge with the same management tools, security policies and automation, and functionality they use for their central private cloud.

Finally, Red Hat is also continuing to focus on helping bring capabilities to bear for new enterprise workloads and use cases. For example, in the OpenStack Stein release we have helped contribute around virtual GPU (vGPU) support for AI/ML workloads like time series forecasting. This can enable business applications like sales prediction tools and manufacturing planning services. 

Linux is the foundation of Red Hat OpenStack Platform

All of our work on OpenStack upstream and Red Hat OpenStack Platform would be moot without a solid platform underneath it. That's Red Hat Enterprise Linux, and with the release of Red Hat Enterprise Linux 8 we're more excited than ever about the possibilities with an intelligent OS as the consistent foundation for open hybrid cloud. 

We'll support RHEL 8 as a guest in Red Hat OpenStack Platform 9, 10, 13, and 14. With the coming OpenStack Platform 15 release, we’ll introduce RHEL 8 as the platform that enables better kernel tracing and monitoring, an enhanced TCP stack, better support for AI/ML workloads with GPU/vGPU support, and improved load balancers. 

Though we abstract things away from the operating system for users and developers, the operating system underneath it all continues to be important and a key focus for Red Hat.

The road ahead

Let's look at some of the things coming in the near and longer term for Red Hat OpenStack Platform.  Bear in mind that nobody knows what the future holds, but we want to be able to share the same roadmap and planning info we offered to customers at Red Hat Summit. In fact, we hosted a live webinar including a customer Q&A (accessible on demand until June 2020) during which we discuss much that was covered at Red Hat Summit.

Our focus for the next few releases of OpenStack are:

• Best on-prem Infrastructure-as-a-Service to run containers (OpenShift)

• Scalability

• Edge computing

• Day two management

Granted no major surprises there, but let's dig in a bit. 

Making Red Hat OpenStack Platform easier to manage

Coming in Red Hat OpenStack 15

Red Hat OpenStack Platform 15 is being based on the upstream OpenStack Stein release. It's going to be a short-life release, meaning that we will support 15 for a total of one year. 

OpenStack Platform 15 is our first OpenStack release which will be based on Red Hat Enterprise Linux 8 control nodes and support for RHEL 8 guests, and is planned to include OpenShift 4.2 integration. We'll also be drawing other important features from OpenStack Stein, like vGPU support. vGPU support allows applications to have managed access to hardware acceleration enabling near-real time decision-making and machine learning. Whether it’s security at a sporting event or quality-control in manufacturing, vGPU support helps make a range of use cases possible.

With OpenStack Platform 15, we’ll be offering a technology preview of our Service Assurance Framework that provides at-scale monitoring of an OpenStack deployment.

Coming in Red Hat OpenStack 16

Our OpenStack 16 release will be a long-life release based on the upstream OpenStack Train release, with support through 2025. Some of the features that we're looking at for that release, expected in early 2020, include: 

• Live migration for pinned VMs

• SR-IOV warm migration

• Full IPv6 support and the Octavia OVN driver

• Dynamic storage provisioning for OpenShift

We also are looking to enhanced day 2 management for the OpenStack Platform 16 release, that includes automated backup and restore, logging and event management. 

We're also planning future OpenStack Platform releases as long life releases. As OpenStack matures we have decided to move to a new cadence for releases that reflects the maturity of the platform and ability of customers to consume and deploy new releases. 

We believe our customers are best-served now with long-life releases that allow them to consume OpenStack in a cadence that's appropriate to mature and important workloads. 

As we continue working on these releases, keep an eye on the OpenStack channel on the Red Hat Blog, where we'll be talking about new and improved features in more detail.