edgeContinuum is an edge computing platform for OpenStack infrastructure spread across many locations: an agent at every facility reconciles desired state locally, so every site is driven from one control plane yet keeps running on its own when the link goes down.
Most platforms manage remote sites by remote control: a central brain pushes commands over the network, and the day the link drops, nothing at the site can change, heal or restart. At edge scale that design fails weekly. edgeContinuum inverts it: the control plane holds the desired state for every facility, and an agent at each site does the actual reconciliation locally. The center decides what should be true; the edge makes it true, and keeps making it true, connected or not.
Each edge site runs its own agent against the local infrastructure. The agent pulls the site’s desired state and reconciles it on site: provisioning, healing, restarts. Deploy one agent or hundreds, one per facility; every new site joins the same console, the same API, the same tenancy model.
When a site loses its uplink, nothing dramatic happens: workloads keep running, the local agent keeps enforcing desired state, healing and restarting what fails, and local management stays available. When connectivity returns, the site resyncs with the control plane and picks up whatever changed. The same property, taken to its extreme, powers our air-gapped deployments.
Shape your facilities into regions and availability zones and operate them as one platform: curate the catalog once, roll it out everywhere, watch health and usage across the fleet from a single console. Adding facility number two hundred looks exactly like adding the second: connect the infrastructure, the agent takes it from there.
Managed, conformant clusters at any site, hosted control planes for small footprints.
The same managed services as your data center, at every facility.
Marketplace products rolled out across sites and fleets in one action.
Organizations, projects and quotas apply at the edge like anywhere else.
An edge computing platform only earns the name if the sites are real infrastructure, not thin remote nodes. Each edge site is your own compute, storage and networking, sitting in the facility where the data is produced, driven today through OpenStack. edgeContinuum puts its control plane on top and turns that estate into one platform. Connect the infrastructure at a plant, a POP or a regional data center, shape it into a region or availability zone, and it appears in the same console as everywhere else, which is the management layer OpenStack never shipped.
The services teams consume at the edge are the services they already know. Managed Kubernetes runs at the site, with hosted control planes when the footprint is too small to spend three nodes on etcd. PostgreSQL, virtual machines and marketplace applications are provisioned through the same self-service flows, under the same organizations, projects and quotas. Usage is tracked per tenant per site, so a service provider sees what each customer consumes at each location, and an enterprise sees which facility is eating its capacity.
And because reconciliation is local, an edge site needs no permanent low-latency path back to the center. The agent holds the site’s desired state and enforces it against the local infrastructure APIs; the WAN carries intent and telemetry, not the control loop. That is why a site on a flaky uplink behaves like a site on dark fibre, and why the same architecture extends all the way to air-gapped facilities.
Keep exploring
Book a call and we’ll walk your edge topology: sites, regions, agents and what an outage looks like on the platform. Or start with one site on the free trial.
