Log in to ask questions, share your expertise, or stay connected to content you value. Don’t have a login? Learn how to become a member.
In this installment of the 128T Data Model Series, we want to talk about the concept of neighborhoods in the 128T Networking Platform. Neighborhoods A Neighborhood is really just a name for a Layer 3 network that one or more of your 128T devices is attached to. But in this simple definition rests an incredibly subtle, but important concept in the 128T way of thinking; neighborhoods have three important influences on your network. One: Automatic Connections Between 128Ts First, as we were designing the software, we wanted each 128T router to be able to automatically identify and associate with other nearby 128T brethren. Of course, in the routing world, nearby devices are called neighbors. So what better name for a collection of nearby, neighboring routers than a neighborhood? That's really the crux of the first role of neighborhoods in your network: a neighborhood is used to describe the intention of connectivity between devices. All routers connect to more than one network. In the 128T data model, each network interface on your router can be assigned to a neighborhood. When two connected routers each have at least one interface in a common neighborhood, the conductor can construct and distribute configuration to those two routers - forming an introduction, essentially - causing them to peer with one another. This conductor function is governed by the topology type assigned to a router's neighborhood when you create it. Topology type can be one of three enumerated values:
As you'd probably surmise, spokes connect to hubs, and everything connects to mesh. Hubs don't connect to other hubs and spokes don't connect to other spokes. In a typical branch office/data center deployment, you'd assign your branches to be spokes and your data center to be a hub. This will have the effect of having your branches connect to your data center, but not have your branches connect to one another. The real power of this topology type is evident when you add add another data center as a hub: all of the branches - whether it's one or one thousand - will automatically create a peering relationship with it when you commit your configuration. Add a new branch? Same thing...place it in the same neighborhood as its counterparts and it'll peer with the data centers with no fuss. So that's the first main function of these Layer 3 networks we call neighborhoods: automatically building topologies of the 128T routers that are attached to them. Two: Multi-Tenant Network-Interfaces The second function of neighborhoods operates in conjunction with tenants, and is used to define tenant membership. As we described in the Video: Tenancy in 128T, your 128T can identify a source of traffic in one of several ways: