I'm a writer in cybersecurity area and I also work for SafeLine, an open source WAF.
Introduction
TCP/IP, short for Transmission Contro...
For further actions, you may consider blocking this person and/or reporting abuse
Seriously? If you are going to write an article about a highly technical subject, at least ensure you understand the topic first! There are so many errors and misunderstandings in this.
TCP does not divide data into smaller "packets" or anything else, you are confusing that with IP. TCP manages streams of data, and mediates the flow using "windows". But these are not fragments in the same way that packets are, and are much much larger in any case.
IP creates packets. And beneath that, whatever the data link layer is, eg Ethernet, creates "frames" for instance.
Next, your description of the layers is incomplete and completely wrong in many respects. There are actually 7 layers in ISO. The top 3 layers (session, application and presentation) are completely unrelated to TCP/IP.. So saying layer 1 is application is wrong because it's the wrong way around and isn't part of TCP/IP in the first place. IP is layer 3, TCP is layer 4. Application is layer 6.
Thus the examples given on HTTP and FTP are not related to TCP/IP at all, they are 2-3 layers higher.
It's not called the Internet Layer it's called the Network Layer of which Internet Protocol is one example.
And "network access layer" isn't even a thing at all! There are 2 lower layers, per my comment above, layer 2 is called the Data Link Layer of which Ethernet is an example.
And below that is the Physical Layer, of which Cat6e, Token Ring, fibre etc are examples.
So many fundamental errors
Make sense
@cj9f61b66f9 Hey, other than RFC docs, do you suggest any resource to take a look to get the big picture?
I would never recommend RFC docs, lol! Maybe the wiki for TCP/IP? Or even OSI. It really depends on what you are interested in understanding. To some extent it's more enlightening to understand why these OSI layer exist in the first place (generally, rather than the TCP specific example)
The answer is, network comms is very complex, so they have broken down each function into layers so that each layer can be dealt with separately and not have to worry about the other layers, communicating with them in well defined ways, each does it's own job and keeps it simple. This means that you can assemble many combinations of protocols TCP doesn't have to use IP under it but that has become the standard. Under IP you could use Ethernet, or 802.1 wifi, or even bridge to cable and ATM, or DSL (unusual). Because they each stick to their own job, they can all be layered together in different combinations.
The bottom layer is Physical and it is more about radio, light or electrical standards on how to put a binary digit onto a transmittal medium. There's no addresses or routing or concept even of point to point. It raw get a bit onto the medium. After that, it doesn't care what happens. (Eg UTP twisted pair Cat6e cables)
Next layer is Data Link - this is entirely about taking the bits delivered by the lower layer, grouping them into a defined "frame" with some header information that says where the frame is going to and from (MAC addresses). But that's pure point to point, local. From me to you. The only thing it can handle is getting a frame of bits from you to a small number of directly connected, local other devices. There is no concept of the "rest of the world" or multiple networks, routing or remote networks.. They are completely invisible at this layer. (Ethernet is the common example)
Next layer is the Network Layer, this creates "packets" of bits, and places those inside the payload of the "frame"in the lower layer. This has more addressing information, and now allows for a global network, full of remote network segments, and an addressing scheme that allows for the notion of local and remote. When a device you are communicating with is local, the network simply uses the data link layer and delivers the packet locally. When the destination is "remote" meaning not on the local LAN, this layer knows to forward the packet to the local network router and asks it to forward it. That happens again and again until the packet reaches the destination network and then is delivered. (This is where IP sits)
Next is the Transport Layer - this is responsible for taking the bits, frames and packets, along with the ability to route packets globally that the lower layers provide, and converts these streams of packets into individual "virtual connections" so that many different processes on the same device can communicate with many different processes on the other device without the packets getting mixed up between processes. There can be hundreds of such connections (called sockets) that allow different conversations between the same devices not to get mixed up with cross talk between those conversations. The example would be TCP and it does lots more as well, like ensuring the communications are reliable and guaranteed, that the flow is controlled and doesn't exceed the speed the network can handle and so on.
Hope that helps
Thanks Man, for the detailed answer!
I think you're confusing OSI(not ISO as you mentioned) model with TCP/IP model.
Apologies, yes. Not sure if that was an autocorrect, or a genuine typo, missed in my fury lol
TCP/IP was introduced in earlier stage and If I'm not wrong it wasn't even by IEEE rather it was from DoD, and later OSI model introduced with more granular layers, But if you would ask me current model is not enough and require completly new framework. Or at-least requires expantion.
wow
I'm a writer in cybersecurity area and also work for SafeLine, an open source WAF. Understanding TCP/IP Protocol is crucial for cybersecurity engineers to protect web services.
Beginners yet very well detailed. Good one!
well written overview.
I wrote about this earlier too 😅.
I saw your post. It's much better. I should learn from you to post better articles. Thanks
I'm humbled 😊 . I'm not any better either though 😅 .
No, it's actually very inaccurate