Lost Password

Switches vs Hubs: What’s the Difference

In the heart of a physical network, every network host on a local network relies on interconnectivity to send and receive data. This describes the fundamental of switching; but, many remain confused on the differences between network switches vs network hubs. In this article, we’ll look at these differences and how network switching has evolved in general.

What are Network Hubs?

netgear en104 network hub

The previous way network switching worked back in the late 1990’s was through the basic network hub (some people may call a hub a “multiport repeater”) at the center of a LAN to connect network devices to one another and transfer the data.

For each network host, the data is sent first to the hub and then the hub sends out a broadcast frame (this would be a destination MAC address of all F’s or FFFF.FFFF.FFFF) in a looped fashion around the LAN to all interconnected network devices until the hub finds the destination host the data will go to.

Network hubs in general were only seen in smaller and very-slow speed local area networks due to the fact of broadcasting multiple data traffic to all devices connected to the network ports on the hub. The more devices means the more data being broadcasted at once and the slower data will transmit due to the interface speed network hubs have.

Hubs Operate at “Half-Duplex”

half duplex switches vs hubs

Continuing on the topic of “ethernet/interface speed,” network hubs use the half duplex interface speed vs switches to accommodate data transfer coming and going.

Operating at half-duplex allows the network hub to only work in one direction at a time meaning that the network hub can receive or transmit; but cannot do both functions at the same time otherwise the network will hang up.

The half-duplex operation gives one network device to talk with the network hub while every other device is placed in “listening” mode of the network traffic while connected to the hub while waiting for their turn to talk to the hub.

Hubs Use CSMA/CD Protocol

In addition to using half-duplex interface mode, network hubs and all devices are part of a collision domain in which frames sent from one interface would collide with frames sent from another interface.

Due to this, Carrier Sense Multiple Access with Collision Detection (CSMA/CD) is enabled by default to enforce devices to listen in to detect the collision until no other devices are sending frames of data before sending frames out.

Using CSMA/CD, the receiving network device would send a jamming signal to inform other devices including the hub when a collision occurs to stop the data transfer sequence which is why half-duplex connections are slow to begin with.

Hubs Operate at the Physical Layer

The other major detail that adds to the meaning of a network hub being a “repeater” is that it simply repeats any and all signals being received from the Ethernet cables interconnected to it.

Because of this characteristic, network hubs operate at Layer 1 in the OSI model or the physical layer to determine how data is transferred between network devices.

What are Network Switches?

tp link tl-sg1024d ethernet switch

Used currently today in all modern local area networks, a network switch in general provides a direct connection to network hosts within the same local area network or LAN…a similar definition compared to a network hub above

Network switches aggregate connections between different sites and branches to make the forwarding decisions based on the single benefit of providing efficiency of determining source and destination computers when transferring data rather than communicating to all connected devices at once.

The major function of switches is the capability of monitoring data traffic on all ports and collect MAC address information to build up a MAC address table for devices connected to it’s ports in order to learn more about the devices

In today’s modern switches, there are at least 24 ports or more (similar to a hub) to switch in data connections from any port to another port to make a bridge. Switches can also be called “multiport bridges.” because of this fact.

Switches Isolate Collision Domains

Unlike hubs, a network switch breaks up the local area network into segments to isolate all data traffic from one part of the network from other networking devices.

By isolating collision domains, the main benefit to achieve for the network switch is there would be an improvement in data transfer speeds and better throughput (less latency) all because the local area network bandwidth is not being shared.

Switches Operate at “Full Duplex”

full duplex mode switches vs hubs

The other factor behind isolation collision domains for network switches as it differs vs. hubs is the CSMA/CD protocol will be turned off by default to prevent the switch from running in half duplex.

Because of this, modern day network switches always run in full duplex interface mode meaning that it can transmit and receive frames of data at the same time along with each network device connected to it.

Unlike half duplex, network devices running in full duplex including the network switch does not have to wait for another network device to finish receiving data before continuing. That is why the full duplex mode is considered the simultaneous and continuous interface mode.

Switches Operate at the Data Link Layer

Unlike network hubs that operate at Layer 1, network switches operate at Layer 2 or the Data Link layer of the OSI Model.

Network switches instead use layer 2 addresses in this particular case of using MAC addresses to configure the switch to learn more about each network device.

By keeping up with MAC addresses, the switches create a MAC address table to determine which network devices are connected to the switch as it plays a factor in a node-to-node connection.

Lastly, switches operating in layer 2 sequence to determine which network devices are receiving and which devices are transmitting frames of data.

Switches Create CAM Tables

The final major difference between switches vs hubs is that switches do create Content Addressable Memory (CAM) tables whereas hubs do not create such a table whatsoever.

Switches create CAM tables for the sole purpose of learning MAC addresses from the connected devices by extracting the MAC address information from the Ethernet packet headers that is transmitted.

Besides identifying MAC address information, the CAM table in a switch has port map information of the device connected to determine which port on the switch it is connected to and what type of device it is connected to the switch (…is it a PC, router, printer, another switch, etc.).

Share This Post

Related Posts


    Leave a Reply

    Thanks for submitting your comment!
    Seraphinite AcceleratorOptimized by Seraphinite Accelerator
    Turns on site high speed to be attractive for people and search engines.