I'm currently learning about the OSI model and the TCP/IP models, and I was wondering: Why in the TCP/IP model the physical layer and the data link layer are combined?

I'll state how I understand the division between the two in the OSI model:

Physical layer

The physical layer is responsible for transmission and reception of raw bits between a device and a physical medium. So this layer is basically responsible for transforming digital 1's and 0's to signals depending on the physical medium such as electrical, optical or radio signals. On the receiving end, it is responsible for doing the opposite - translate the physical signals into digital ones. This is all done by hardware components - NICs, cables, etc. The specifications in this layer will include voltage levels, wave frequencies, etc.

Data link layer

Everywhere I read, the function of the data link layer was phrased as "the transfer of data between nodes on the same network segment" or something similar. That confuses me a little bit since the physical layer puts the bits on the wire and collects it from there, which is effectively transferring data between nodes on the same network (please help me cut this more clear if you can). Anyway, from what I understand (please confirm my understanding or correct me) is that this layer supersedes the function of the physical layer. While the function of the physical layer is merely to translate digital to whatever, the duty of this layer is to make sure those bits arrive to their destination. This may (but not must) include error checking and correction, Media Access Control - which means synchronizing the different devices that wants to use the medium, addressing - so that frames will go to the desired-by-the-network-layer destination, and maybe other things.

The data link layer core operation is to encapsulate network packets into frames which includes a checksum, source and destination address and some more data. These frames are used in order to achieve the above goals.

All this is done in software - software that partially runs in the OS as device driver of the NIC, and partially on the NIC itself. Protocols in this layer defines the structure of frames, how Media Access Control is managed, error detection and correction, etc.

So now to the question - if these are 2 separate functions, why is this division missing from the TCP/IP model? I can understand that in the TCP/IP model there's no session and presentation layers, since there is no such functionality on its own, its all done by the application layer. But here with physical and data link, surely these are 2 separate functions, aren't they?

  • 1
    Remember that these models are just that -- abstract models. IMO studying the models are good for a general understanding, but don't get too hung up in the details. Many protocols in actual use don't neatly fit into the models.
    – Ron Trunk
    Sep 12, 2021 at 21:04

2 Answers 2


[Data link layer] That confuses me a little bit since the physical layer puts the bits on the wire and collects it from there, which is effectively transferring data between nodes on the same network (please help me cut this more clear if you can).

The physical layer puts bits and bytes on the wire - between two nodes. The data link layer frames them and directs the frames to the local destination.

The data link layer is usually completely handled in hardware. A NIC covers physical and data link layer, so does a switch. (Frame assembly and interpretation used to be software as well but see below.)

The network layer used to be handled in software, due to complexity, but specialized hardware has long since learned to handle network and even transport layer in hardware, partially or entirely: multilayer switches, hardware router, offloading function in NICs, ...

why is this division missing from the TCP/IP model?

It's an important distinction to make if you are developing services within these layers, but the folks from ARPA (or IETF later on) defining the TCP/IP model simply didn't care. They abstracted the data link layer downward as the link layer. To be fair, the TCP/IP model predates the OSI model quite a bit. The ISO, developing the OSI model, aimed at standardizing the full stack, from top to bottom.

If you're studying the popular physical and data link layer protocols, you will see that they are even sublayered - for work distribution and modularity reasons.


The internet was a project to connect disparate networks into a coherent whole.

It was neither necessary nor desirable for the architects of the internet to worry about how exactly those underlying networks got the packets from A to B as long as they got them there.

By contrast OSI were trying to standardise the whole network stack from top to bottom.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.