Like it or not, organizations are moving more and more workloads to the edge. The benefits of moving workloads closer to the users that depend on them are undeniable. So are the downsides. However, with Red Hat Enterprise Linux (RHEL) 8.3 we've added several features to help admins and organizations cope with the challenges of edge computing by simplifying image creation, reducing update sizes and fending off unnecessary downtime.

If we're being honest, edge environments are a system administrator's nightmare. The conditions we depend on (and almost take for granted) in a datacenter or cloud environment -- like plentiful bandwidth, 24x7 hands-on support and redundant hardware -- aren't part of our edge environments. Edge sites may only have one server to handle workloads, with limited bandwidth and no trained IT staff (or any staff!) to help troubleshoot if an update goes awry. 

Providing stability and deployment flexibility

Despite those challenges, you still need to keep those systems up and running. In RHEL 8.3 we've added several features that are specifically designed to help you build, deploy and maintain edge systems. 

Generate images for the edge quickly and with flexibility

With RHEL 8.3 we've added support for rpm-ostree to build RHEL images for edge deployments. You can think of rpm-ostree as "git for operating systems." Instead of thinking of an installation or update as a transaction consisting of many RPMs, each update is a single atomic operation and that unlocks a lot of possibilities for edge systems.

Building images and updates for the edge is quick and fast. The images are optimized for edge architecture, and customizable for specific deployments. But wait, there's more!

Efficient over-the-air (OTA) updates

We know that edge systems may be located in environments with limited bandwidth or even intermittent connectivity. In those situations, you want to limit the amount of data you have to push to update a system. 

When using traditional package updates, admins would run a transaction that would update, add or remove packages while a system was running, restart services and possibly reboot if any of the packages required it. 

Using these new images, we have the ability to send out delta updates that are optimized for intermittent or low-bandwidth connections and contain only a fraction of the size of traditional RPM-based updates. These give organizations efficient over-the-air (OTA) updates that are applied to a staged filesystem (not the running filesystem) tree. 

With RHEL 8.3 and new images, admins can stage updates (consuming less data) and then apply all the updates on reboot. Choose the best time for a maintenance window and apply the updates on your terms to maximize uptime. 

Because these are image-based updates rather than a collection of packages, if you push out an update to 10 edge sites they will each receive the same update and won't be subject to package drift from site to site.

Intelligent rollbacks to avoid downtime

If you've worked in IT for a while, you've probably experienced an update that broke an application or rendered a system unusable in some way. That's a headache if you're sitting next to the machine, and a real problem if the machine happens to be several cities away where there's no IT staff to provide support.

With RHEL 8.3, we are introducing intelligent rollbacks that will allow you to run a health check on an edge server and roll back to the last known-good update if problems are detected. This gives you the ability to check critical services, applications and system health to prevent unnecessary downtime. If a test fails, roll back to the state prior to the update and resume services quickly.

The same technology that lets us stage updates and hold them for a convenient maintenance window allows us to roll back to the prior image when necessary. 

The confidence to take on the edge

Organizations already depend on RHEL to run workloads across on-premise datacenters and the public cloud. By adapting RHEL to meet the demands of the edge environment, we make it possible to extend your tools and expertise from the comfort of the datacenter to the reaches of the edge.

More on Red Hat, RHEL, OpenShift and the edge


About the author

Joe Brockmeier is the editorial director of the Red Hat Blog. He also acts as Vice President of Marketing & Publicity for the Apache Software Foundation.

Read full bio