| Feature | Hub | Switch |
|---|---|---|
| OSI layer | Layer 1 – Physical | Layer 2 – Data Link |
| What it does | Repeats signals to every port | Forwards frames only to the intended device |
| Uses MAC addresses? | ❌ No | ✅ Yes |
| Traffic | Everyone receives the traffic | Usually only the destination receives it |
| Collisions | One shared collision domain | Each port has its own collision domain |
| Performance | Slow and inefficient | Faster and more efficient |
| Security | Poorer — traffic is visible to all connected devices | Better — traffic is normally sent only where needed |
| Modern use | Very rare | Extremely common |
A simple example
Imagine PC-A wants to send data to PC-C.
With a hub:
PC-A → Hub → PC-B + PC-C + PC-D
The hub doesn’t know who the data is for. It simply repeats the signal to everyone. Each computer has to check the data and decide whether it is meant for them.
With a switch:
PC-A → Switch → PC-C
The switch learns the MAC addresses of devices connected to its ports. When PC-A sends a frame destined for PC-C, the switch can forward that frame specifically to PC-C.
The big idea
A hub broadcasts everything. A switch makes decisions.
That’s why switches dramatically reduce unnecessary traffic and collisions and provide much better network performance.
One small clarification to the transcript: modern Ethernet switches don’t literally “eliminate” all collisions in every possible configuration, but with full-duplex switched Ethernet, collisions are effectively eliminated because each link can transmit and receive independently.
Easy way to remember it:
🔌 Hub = Layer 1 = repeat everything
🔀 Switch = Layer 2 = look at MAC addresses and forward intelligently
Why do companies still use hubs?
Companies almost never use traditional Ethernet hubs in normal production networks anymore. Switches have replaced them because they are faster, more efficient, and much easier to manage.
However, there are a few situations where you might still encounter a hub:
- Legacy equipment: Some older networks or industrial systems were built around hubs and haven’t been upgraded.
- Specialized troubleshooting: A hub can be useful when a network engineer wants to see traffic from multiple devices with a packet analyzer. Because a hub repeats traffic to every port, a monitoring device can observe it easily.
- Training and education: Hubs are useful for demonstrating concepts such as collisions, collision domains, Ethernet frames, and Layer 1 networking.
- Specialized/industrial environments: Some older industrial or embedded systems may still use hub-like devices because replacing certified or long-lived equipment can be expensive or risky.
- Very simple temporary setups: Occasionally, someone may use an old hub for a specific testing scenario—but this is uncommon.
Why don’t companies normally use them?
Imagine a company with 50 computers.
With a hub, when computer A sends data to computer B:
A → Hub → 49 other devices also see the traffic
With a switch:
A → Switch → B
The switch makes a decision based on the destination MAC address, so unnecessary traffic isn’t sent to every device.
There’s also an important historical reason: old Ethernet networks using hubs operated in half-duplex and had to deal with collisions. Modern switched Ethernet normally operates full-duplex, so collisions aren’t an issue on the individual links.
The easiest way to remember it
Hub: “I don’t know where this goes, so I’ll send it everywhere.”
Switch: “I know where this device is, so I’ll send it directly there.”
So if you’re studying networking today, think of a hub as primarily a historical/educational device, while the switch is the standard device used to connect computers on a modern LAN.
Geef een reactie