Leaf-Spine FAQs

Leaf switches are woven into the spine so that your network has a broad access layer with multiple network connection locations for the servers. In most builds, you’ll see that spine switches have a significantly higher port density in order to form a strong core architecture.

This type of layered connection that builds upon core switches is what makes this deployment type a top-of-rack (TOR) deployment, even though this architecture is often virtualized.

The leaf-spine network deployment topology is somewhat more predictable and consistent than other topologies, because all devices on this network will have be the same number of network segments away, creating predictable delay and latency times for information traveling through your network.

The spine-switch backbone performs all of the routing itself, and is able to deliver with speed and consistency thanks to having a connection to every leaf switch.

Part of the predictability of this topology comes from a Layer 3 routing paradigm that interconnects the two-layer design. By relying on software defined networking technologies with a focus on “East-West” traffic, this network can dynamically route traffic to the best path based on network changes and traffic spikes.

This directional traffic is data that moves within the data center itself and is able to remove the past limitations of the Spanning Tree Protocol (STP) network deployment topologies.

The chief advantage of leaf-spine network deployment topologies that live in the ToR is their ability to use all of your connections at the same time in a way that keeps the network stable — that means no loops. You’re doing this more efficiently because the virtualization of the network removes many of the STP limitations and requirements for configuring these devices and properly setting priorities. STPs often face inefficient paths and work harder because it is very easy to set improper device priorities and manage those afterwards.

Leaf-spine networks also have an infrastructure and growth benefit over STP deployments. The software-defined networking support for leaf-spine allows this topology to better support adding hardware and new network capacities.

You can easily and automatically expand capacity with predefined routes whenever you experience traffic demand or an oversubscription of links. Spine switches are straightforward to add and extend to each and every leaf switch. To improve port availability, leaf switches are installed and quickly supported by adding the already defined network configuration to the switch.