Choosing a Network Switch for Your Homelab: Speed, Managed vs Unmanaged, and VLAN Basics

A network switch is what connects your homelab nodes together. Without it, your Proxmox nodes can’t form a cluster, can’t live-migrate containers between each other, and can’t share storage. For most homelab guides, the switch is an afterthought — “just plug everything in.” This post explains what you’re actually buying and what the differences mean for your specific workloads.

What a switch does (and how it differs from a router)

A router connects different networks together and makes decisions about where traffic goes between them (your home LAN, the internet, VLANs). It assigns IP addresses via DHCP.

A switch connects devices on the same network together. It operates at Layer 2 (MAC addresses), not Layer 3 (IP addresses). When a Proxmox node sends traffic to another Proxmox node on the same switch, that traffic never touches your router at all — the switch forwards it directly.

In the diagram above, traffic between pvelab01 and pvelab02 (Proxmox cluster heartbeats, live migration, Corosync) flows switch → switch without going through the router. This is important: your router’s throughput and latency don’t matter for inter-node cluster traffic.

Speed: 1 GbE vs 2.5 GbE vs 10 GbE

What the speeds mean in practice

For most homelab workloads, 1 GbE is completely adequate:

• LXC containers running web services, databases, monitoring: generate 1–10 MB/s of traffic each
• A 4-node Proxmox cluster’s Corosync heartbeat traffic: under 1 MB/s total
• Eight 1 GbE ports can theoretically handle 8 × 115 MB/s = 920 MB/s of aggregate traffic simultaneously

When 1 GbE becomes a bottleneck:

• You’re doing frequent live migrations of large VM disks (live migration uses the management network)
• You’re running Ceph distributed storage (which requires high-bandwidth inter-node traffic)
• You have a NAS and regularly transfer multi-gigabyte files to/from multiple nodes simultaneously

For the workload in these guides — Proxmox with LXC containers, Ollama inference, monitoring — 1 GbE is more than sufficient. The CPU is the bottleneck, not the network.

Managed vs. unmanaged vs. smart switches

This is the decision most homelab guides gloss over.

Unmanaged switch

Plug in and it works. No configuration. No web interface. Traffic goes where MAC addresses say it goes.

Pros: Cheap ($15–$30), zero setup, never needs firmware updates.
Cons: No VLANs, no port mirroring, no traffic monitoring, no link aggregation.
Best for: A single-node homelab, a simple home network, desktop peripherals.

Smart switch (also called “web managed” or “easy smart”)

Has a web UI. Supports VLANs, QoS priorities, and port monitoring. Limited compared to full managed switches but vastly more useful than unmanaged.

Examples: TP-Link TL-SG108E, Netgear GS308E, QNAP QSW-308S
Pros: Affordable ($30–$60), supports VLANs (critical for advanced Proxmox setups), has a web UI
Cons: Limited CLI, sometimes quirky web interfaces
Best for: Most homelabs — this is the right choice for a 4-node Proxmox cluster

Fully managed switch

Enterprise-grade. Full CLI (like Cisco IOS or similar), SNMP monitoring, LACP link aggregation, spanning tree, port security. The kind of switch in data centers.

Examples: Cisco Catalyst 2960, HP Aruba 2930F, Juniper EX2300
Pros: Full feature set, enterprise reliability, CLI-driven configuration
Cons: Expensive ($200–$2,000+), steep learning curve, complex configuration
Best for: Advanced homelabs where you specifically want to practice enterprise networking skills

What a typical 8-port gigabit switch looks like

Recommendation

For starting out: TP-Link TL-SG108 unmanaged 8-port — $20, just works.

For a proper homelab cluster: TP-Link TL-SG108E smart 8-port — $35, has VLANs and a web UI, sufficient for 4 nodes.

For learning enterprise networking: Cisco Catalyst 2960-X (refurbished) — $80–$150, full Cisco IOS CLI, the same switch you’d find in a corporate environment.

What VLANs are and why they matter for Proxmox

A VLAN (Virtual LAN) segments a physical network into multiple logical networks using the same cables and ports. Traffic on VLAN 10 and traffic on VLAN 20 are completely isolated from each other even though they share the same switch and cables.

Why you’d use VLANs in a Proxmox cluster:

For a 4-node homelab on a gigabit switch, VLANs are optional — your total traffic fits easily on one network. But if you want to learn the concepts (and use them in a future job interview), setting up a VLAN-aware Proxmox bridge is good practice.

Task 1: Set up your switch for a Proxmox cluster

# Replace IPs with your actual planned IPs
ping 10.0.0.71   # pvelab02
ping 10.0.0.72   # pvelab03
ping 10.0.0.73   # pvelab04

# Check the link speed negotiated
ethtool eth0 | grep Speed
# Should show: Speed: 1000Mb/s

Task 2: Configure a smart switch with VLANs (optional, TP-Link TL-SG108E)

Sizing your switch

For a 4-node cluster: an 8-port switch is the right size. You use ports: router uplink (1), node 1 (1), node 2 (1), node 3 (1), node 4 (1), laptop (1) = 6 ports, 2 spare.

Cable recommendations

For short runs inside a rack or shelf:

• Monoprice 0.5ft Cat6 patch cables — $8–$12 for a 10-pack, keeps things tidy

For connecting to a wall jack or longer runs:

• Amazon Basics Cat6 patch cable 3ft — reliable, inexpensive, widely available

Avoid Cat5. It’s technically gigabit-capable but isn’t guaranteed to the same spec as Cat5e or Cat6. Cat6 costs the same and is better in every way.

Next steps

With your hardware connected and your nodes talking to each other, the next step is installing Proxmox VE on each machine.

→ Continue to: Installing Proxmox VE: Step-by-Step from USB to First Login

This post contains Amazon affiliate links (tag: buildahomelab-20). I earn a small commission on qualifying purchases at no extra cost to you.