0

A simplified example of how the OSI model works, from this page:

For example: Mr. Cooper wants to send Ms. Palmer an email. Mr. Cooper composes his message in an email application on his laptop and then hits ‘send’. His email application will pass his email message over to the application layer, which will pick a protocol (SMTP) and pass the data along to the presentation layer. The presentation layer will then compress the data and then it will hit the session layer, which will initialize the communication session.

The data will then hit the sender’s transportation layer where it will be segmented, then those segments will be broken up into packets at the network layer, which will be broken down even further into frames at the data link layer. The data link layer will then deliver those frames to the physical layer, which will convert the data into a bitstream of 1s and 0s and send it through a physical medium, such as a cable.

Now that's fine, but a browser will first initialize communication with a server by agreeing in a cipher and encryption system (the "hello"). Could we say that when we press send, what first happens is that there is the hello from the client to the server, and that's how the session starts? I think I'm quite confused so any help is welcome.

1 Answer 1

1

and pass the data along to the presentation layer is wrong, at least outside the theoretical OSI model. In practice, whatever exists of a presentation layer resides inside the application (you could regard MIME as part of the presentation layer). Of course, you could structure an email client in the described way but not necessarily so. We can't dive deeper into that here as protocols above the transport layer are explicitly off-topic.

The data will then hit the sender’s transportation layer where it will be segmented, then those segments will be broken up into packets at the network layer

Not quite. The segmentation in some transport-layer protocols already anticipates packetization. The network layer just encapsulates those segments and guides them to the destination.

Only if the network layer discovers that the upper layer protocol unit causes a packet to become too large for the underlying data link layer protocol (larger than the MTU), then it may choose to break it down further, e.g. fragmentation in IP (IPv6 can fragment on the source node only, not in-path). It may fragment but it may also reject.

packets at the network layer, which will be broken down even further into frames at the data link layer.

Like above, the data link layer encapsulates network-layer packets. It doesn't break them down any further, there's always a 1:1 relationship.

a browser will first initialize communication with a server by agreeing in a cipher and encryption system (the "hello")

It might, not all communication is encrypted. With TCP/IP, encryption happens on the application layer or on top of the transport layer (SSL/TLS).

Could we say that when we press send, what first happens is that there is the hello from the client to the server, and that's how the session starts?

TLS does establish a kind of session, but that isn't really what the OSI model defines (foremost, TLS misses user authentication).

I think it's best to visualize a session inside the application layer - for instance, many web services use cookies or URL tags for their concept of a session. Many other TCP-based protocols base their session concept on the TCP socket at hand - like SMTP. The session concept is often more obvious with UDP-based protocols because you can't just use anything from UDP.

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.