Typical Two‑Layer Spine‑Leaf Topology and Its Scalability
The article explains the classic two‑layer spine‑leaf (Clos) architecture, describing how each leaf switch connects to every spine switch, how oversubscription is handled by adding spines or leaves, and why the design offers predictable latency and easy scalability for data‑center networks.
Figure 4 shows a typical two‑layer spine and leaf topology.
Figure 4. Typical spine and leaf topology.
In this two‑layer Clos architecture, each lower‑layer switch (leaf) is connected to every top‑layer switch (spine) in a full‑mesh topology. The leaf layer consists of access switches that connect to servers and other devices, while the spine layer forms the network backbone interconnecting all leaf switches. Every leaf switch connects to each spine switch, and paths are chosen randomly to distribute traffic evenly across the top switches; a failure of a single spine switch only slightly reduces overall data‑center performance.
If link oversubscription occurs—i.e., generated traffic exceeds the aggregate capacity of active links—the capacity can be expanded simply by adding an additional spine switch and extending the uplinks to each leaf, thereby increasing inter‑layer bandwidth and reducing oversubscription. When port capacity becomes a bottleneck, new leaf switches can be added to each spine and the network configuration updated. This scalability streamlines the IT department’s network expansion process, and if lower‑layer switches and their uplinks are not oversubscribed, a non‑blocking architecture can be achieved.
For spine‑leaf architectures, regardless of which leaf a server connects to, its traffic must traverse the same number of devices to reach another server (unless the destination server resides on the same leaf). This keeps latency predictable, as traffic only needs to hop to one spine switch and then to another leaf switch to reach its destination.
Original source: Cisco White Paper
Article: http://jiagoushi.pro/node/1032
Discussion: Please join the Knowledge Planet "Chief Architect Circle" or add the WeChat alias "intelligenttimes".
Follow the public WeChat account "Chief Architect Think Tank" and join the related groups for architecture, cloud computing, big data, data science, IoT, AI, security, full‑stack development, DevOps, digital transformation, and product transformation.
Architects Research Society
A daily treasure trove for architects, expanding your view and depth. We share enterprise, business, application, data, technology, and security architecture, discuss frameworks, planning, governance, standards, and implementation, and explore emerging styles such as microservices, event‑driven, micro‑frontend, big data, data warehousing, IoT, and AI architecture.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.