A self-healing wireless mesh is a network of radio nodes in which each node can reach more than one neighbour. When a link fails or its quality drops, traffic takes another path automatically, and nobody has to go out and fix anything before the network works again.
How it differs from other wireless designs
A point-to-point link, or a point-to-multipoint network built around one hub, gives each site a single path. If that path fails, the site is cut off until someone repairs it. A mesh builds alternative paths into the design, so a failure becomes a change of route rather than an outage.
What self-healing means in practice
In an Anywhere network, the three steps below take less than a second, with no manual intervention and no single point of failure.
- Detect: nodes continuously measure the links to their neighbours
- Reroute: when a node, link or path fails or degrades, traffic moves to a healthy route
- Continue: cameras, sensors and users stay connected while the fault is found and fixed
How the path is chosen
Self-healing depends on choosing good paths all the time, not only when something breaks. Anywhere nodes pick the best route by latency, load and link quality, and can spread traffic between several sites or servers by least path cost. Several nodes can act as drop-off points to the wired network, which adds both capacity and resilience.
Why the number of hops matters
In a conventional mesh, a node with one radio has to share its airtime between receiving traffic and passing it on, so capacity falls with every hop and the network becomes unusable after a few relays. Anywhere nodes use dual radios and a patented topology and channel allocation scheme to avoid most of that loss, supporting more than 30 hops of backhaul.
Rings: the simplest self-healing design
A ring gives every node two paths, one in each direction. It is often the backbone of a larger mesh. Fung Kai Schools in Hong Kong runs a five-hop ring across the rooftops of four campuses, and Whitsunday Regional Council in Australia runs more than 100 HD cameras on a dual-ring network with several levels of redundancy.
Does it behave like a normal network?
To the systems connected to it, an Anywhere mesh behaves like a Layer 2 switch with its ports spread across the site. Cameras, recorders and video management systems connect as they would over Ethernet, VLANs pass through, and the whole network is managed through one IP address.
Where self-healing mesh is used
- City-wide CCTV, traffic and flood monitoring
- Ports, airports and industrial sites
- Train-to-ground and depot connectivity
- Cranes and other moving equipment
- Temporary sites and events
- Rural and long-distance backbones
Questions to ask a supplier
- How quickly does the network recover when a link fails?
- How many hops can it carry before throughput becomes a problem?
- Does it need a central controller, and what happens if that fails?
- Is it transparent to existing cameras, VLANs and video management systems?
- How is traffic encrypted on the wireless link and end to end?
- Are there licence fees that grow with the network?