Every AREDN® node is capable of automatically joining an ad hoc mesh network which is operating with the same SSID, channel, and bandwidth. New nodes will each explore their surroundings by broadcasting their identity and listening for their neighbors’ responses. Once nodes identify others within radio range, they share this information so that each node has a picture of the network topology. Periodic updates adjust the network routes based on changes in signal quality or loss of a link, allowing the network to adapt to changing conditions. Since there can be several possible routes between nodes, and since network disruptions typically effect only part of the network, a mesh topology can provide redundancy for network links.
Every AREDN® node within radio range of other nodes will be able to participate in the network to extend its reach, provide route redundancy, or host services needed on the network at large. This simple mesh topology may serve its purpose perfectly for a short-term network deployed in support of a local event, or even for more permanent communication between nodes which are always within radio range. However, as mentioned in the previous chapter, the most important consideration for you network design is, “What is the purpose for this particular network?” The specific requirements of your mission should drive the design of your data network.
Types of Topologies
Although AREDN® nodes are capable of forming a simple mesh network, it is more common for operators to use different topologies in order to accomplish their data communication goals in growing networks. Typical network designs include Point-to-Point, Hub-and-Spoke, Tree or hybrid topologies.
- Point-to-Point Topology
Point-to-Point topologies are best suited for moving data between the far endpoints, potentially using one or more intermediate nodes in order to traverse different types of terrain or to overcome obstacles in the network path.
- Hub-and-Spoke Topology
Hub-and-Spoke topologies work well in situations where the data communication to outlying nodes should be coordinated or funneled through a central location. Even if a remote node becomes unreachable, the rest of the network can continue to operate; but if the central node goes offline, the network will not function.
- Tree Topology
A tree topology can be used to segment or partition network traffic, keeping specific data within a localized area while also allowing for links to remote parts of the network. The tree topology uses a parent-child hierarchy to structure the paths that data can take. This design can be easily scaled up or down to meet the specific requirements of the mission, but it does create “single points of failure”. If nodes go offline within the hierarchy then entire branches of the tree can become unreachable.
Once several local or regional networks have been created, there may be a need for communication between these “mesh islands.” Often node owners have used direct Internet tunnel connections to accomplish this. However, this has the effect of merging the mesh islands into a single network with all of the routing traffic traversing all of the member networks. Many of the legacy nodes with older hardware/firmware are unable to handle the increased load.
A more efficient solution is to use a Supernode network to provide access across mesh islands, without sharing all of the local routing traffic across the linked networks. A Supernode is a specialized, dedicated node whose sole purpose is to link with other Supernodes and to shield each local network from the aggregate routing traffic. Mikrotik hAP ac2 hardware is recommended for Supernodes, along with an Internet connection that provides robust bandwidth .
A Supernode network is a high-level mesh network —
super meaning “above or higher.” The Supernode network sits above the isolated mesh networks and provides connectivity without increasing the routing load on the local networks.
A new solution for Supernode networks is currently being tested, and more information will be forthcoming in future documentation.