Understanding EOR, MOR, and TOR Data Center Network Architectures

Data‑center server rooms typically adopt a rectangular structure, arranging 10‑20 cabinets back‑to‑back into pairs called PODs to improve cooling efficiency.

What is EOR?

EOR (Edge‑of‑Rack) is the most traditional method for integrating access switches; the switch is installed in a cabinet at the end of a row, and servers connect to it via horizontal cabling, making it the most common architecture.

What is MOR?

MOR (Middle‑of‑Rack) improves on EOR by placing the top‑of‑rack cabinet in the middle of each row, reducing cable length between server and network cabinets, simplifying cable management, and allowing custom pre‑connected systems.

What is TOR (Top‑of‑Rack)?

TOR, also known as the top‑of‑rack wiring method, extends EOR/MOR by mounting one or two access switches on the top of each server cabinet; servers connect via jumpers, and the uplink ports connect to aggregation or core switches in the EOR/MOR network cabinets, reducing the number of cables per cabinet to 4‑6.

Differences among TOR, EOR, and MOR

As data‑center density and cloud‑computing demands increase, the number of network ports grows, making TOR more suitable for rapid expansion, while EOR remains the most common and MOR offers a balance with shorter cable runs and easier management.

For further reading, the article provides links to various market research reports and e‑book collections on cloud computing, PaaS, SaaS, and hybrid‑cloud solutions.