The OSI model is confusing me as hell. Help me understand it!

[iamgenius]

I just decided to seriously prepare for the comptia network + exam. I did take a look at it some years ago, but didn't continue. I have some experience when it comes to networking as I have built and troubleshooted networks before. I thought it will be easier this time because my knowledge increased. However, I just got shocked the first time I read through the chapter concerning the OSI model !!! It just doesn't add up. I couldn't understand it very well, and I hate lying to myself and I also hate memorizing something that I don't understand.

When you categorize something, you put similar stuff under the same category. I don't see how these things put under the data link or transport layer are similar!? And what is really the different between the session layer and the transport layer for example. Maybe I'm little hasty, and things will be more clear if I go on, but I think hearing different ways of explaining it from so many different people will help me understand.

And of course, it is not something to skip................it was created for a reason.


Thanks.

 

[seepy83]

My best suggestion is to look for other learning material (a different book, youtube videos, blogs, etc) on the OSI model. Depending on the way it's written, I'm not surprised that reading one author's explanation of the OSI model could leave someone clueless as to why OSI exists, what it does, how it works, etc.

There is a ton of material available on this topic. You just need to find someone that explains it in a way you understand. Once you get the basic concepts, re-read the chapter in the book you have now and it will probably make more sense.

And you're right, OSI is absolutely not something to skip. It is the foundation of networking.

 

[SecurityTheatre]

I suggest looking at real traffic. I finally understood it when I realized how real traffic went.

For example, looking at a connection to a server, I'm using the following layers.

Imagine the request to https://google.com.


Layer 1 Physical (The wire - 100-TX via cat 6 cable for example)

Layer 2 Data Link (On-the-wire Comms - 802.3 Ethernet Spec - MAC addresses, CAM tables, etc)

Layer 3 Network (Network-to-Network comms - using IP - also could be IPX, ATM, etc - IP addresses, routes, DHCP, etc)

Layer 4 Transport ( Endpoint-to-Endpoint comms - using TCP - could also be UDP, ICMP, etc - port numbers, transmission windows, re-transmissions, etc)
Layer 5 Session (merged into Layer 4 with TCP/UDP in normal TCP/IP comms, since TCP/UDP and most IP protocols handle sessions and transports as one thing)

Layer 6 Presentation (application wrappers like SSL when using HTTPS - not always used in TCP/IP)

Layer 7 Application (Application protocol - HTTP in the case of a website, could be FTP, telnet, etc)

 

[iamgenius]

My best suggestion is to look for other learning material (a different book, youtube videos, blogs, etc) on the OSI model. Depending on the way it's written, I'm not surprised that reading one author's explanation of the OSI model could leave someone clueless as to why OSI exists, what it does, how it works, etc.

There is a ton of material available on this topic. You just need to find someone that explains it in a way you understand. Once you get the basic concepts, re-read the chapter in the book you have now and it will probably make more sense.

And you're right, OSI is absolutely not something to skip. It is the foundation of networking.

Alright, I'll try. I think I'm just angry at myself because I was really clueless. I will come back after sometime and tell you guys how it went. I'd love to explain it in my own words if I fully understand it.

 

[iamgenius]

I suggest looking at real traffic. I finally understood it when I realized how real traffic went.

For example, looking at a connection to a server, I'm using the following layers.

Imagine the request to https://google.com.


Layer 1 Physical (The wire - 100-TX via cat 6 cable for example)

Layer 2 Data Link (On-the-wire Comms - 802.3 Ethernet Spec - MAC addresses, CAM tables, etc)

Layer 3 Network (Network-to-Network comms - using IP - also could be IPX, ATM, etc - IP addresses, routes, DHCP, etc)

Layer 4 Transport ( Endpoint-to-Endpoint comms - using TCP - could also be UDP, ICMP, etc - port numbers, transmission windows, re-transmissions, etc)
Layer 5 Session (merged into Layer 4 with TCP/UDP in normal TCP/IP comms, since TCP/UDP and most IP protocols handle sessions and transports as one thing)

Layer 6 Presentation (application wrappers like SSL when using HTTPS - not always used in TCP/IP)

Layer 7 Application (Application protocol - HTTP in the case of a website, could be FTP, telnet, etc)

Your way helps a little. Actually, the only layers I kind of understood before are Physical and application layers. Now, I know that my understanding of these two layers weren't off by much. I still have some way to go though!


Forgot to add that I currently carry:
Comptia strata
Comptia A+
Comptia project+

 

[SecurityTheatre]

Your way helps a little. Actually, the only layers I kind of understood before are Physical and application layers. Now, I know that my understanding of these two layers weren't off by much. I still have some way to go though!

The difference between session and transport are subtle and made confusing by the fact that common protocols we use today implement both together.

The transport layer ensures that communications that are sent are properly received, for example, with TCP, there are ACK packets that are sent - and each of these packets is identified by unique frame numbers - this is all part of the transport layer. Also included in this are network "ports", which identify different types of applications on the LAN, although the concept of "ports" is not strictly required by OSI, they are helpful when talking about and identifying traffic on the network.

The session layer involves grouping individual packets into "conversations" or "sessions" and TCP also handles this through the use of the SYN-SYNACK-ACK handshake and the FIN-FINACK handshake for opening and terminating connections, coordinating "sessions" of communications with a single server/port combination.

In this way, the two are merged into a single protocol. A transport like UDP is very "thin" in these layers as it provides very little control and almost no session information (being "connectionless") and simply specifies sequence numbers and then the higher layers (application, etc) must determine where there is missing data, or if the missing data is even important and whether it must be retransmitted.

Also, this image (from Microsoft) may help simplify things, if you are familiar with TCP/IP.