I will say this right up front. I am currently a college student in a class called Networking Fundamentals. Even though the class is very challenging, I find networking to be amazingly fascinating, enough so that I might consider changing my major. I am still trying to grasp the basics of the OSI model, but I am still a little bit shaky. I would like to have a firm grasp of the model, and I was wondering if anyone could help me with some analogies and general questions.

Let me share with you what I know, and please feel free to correct me if I am wrong.

1. I understand that the OSI protocol model is more like a theoretical guideline for networking.
2. I understand that each layer of the OSI has specific functions.
3. I know the seven layers are Application (Layer 7), Presentation (Layer 6), Session (Layer 5), Transport (Layer 4), Network (Layer 3), Data Link (Layer 2), and the Physical (Layer 1).
4. I know that some protocols begin in the Application layer and persist through the Transport or Network.

I know more about these layers, but I don't want to bore you with a long list.

Questions:

1. At what layer does data actually leave the original host to begin its journey?
2. Do packet-switching or circuit-switching take place within the OSI model (theoretically, of course)?
3. At each level, where--physically--do these functions take place (where is the data in each layer)?

I appreciate any assistance you might offer.

The 7 Layer OSI model is considered a reference model, the TCP/IP 'model" is an implementation of an actual protocol suite.

Are you doing a Cisco Networking Academy course called CCNA Exploration: Network Fundamentals? If you are then you will see that each chapter covers a layer of the OSI model/TCP/IP suite so your questions, and much more, will be answered. There's nothing stopping you from working ahead of your class schedule now that you have access to the curriculum. Each chapter has embedded Packet Tracer activities that will guide you through the operation of the OSI model; and has other activities and an end of chapter quiz that you can do any time.

If you are not enrolled in that course, then to generally address your questions:
1. Data, in the form of electrical, optical or radio signals, that represent frames of bits, leave a host at the Physical Layer when transmitted onto the media.
2. Packet-switching occurs at Layer 3 - the Network (Internet) Layer; circuit-switching ocurs at Layer 1 - the Physical Layer; and there's nothing theoritical about that!
3. Essentially Layer 7, 6, 5, 4, 3 and the upper sub-layer of Layer 2 are implemented in software and the lower sub-layer of Layer 2 and Layer 1 are implemented in hardware. Network client software (web browsers, email clients, messaging apps, VPN clients, video apps, your Windows Explorer, etc) and Server (web, file, email, messaging, etc) software that implement Application Layer protocols run at Layers 7, 6, and 5. The host Operating System contains code that implements the TCP/IP suite of protocols and takes over at some fuzzy point in Layer 5, and then Layers 4, 3 and the upper-sub layer of Layer 2. The bottom half of Layer 2 and Layer 1 are implemented in hardware, such as in your Ethernet Network Interface card. Your NIC has the electronics, including some processing and memory functions, to send and receive the electrical signals that is your data. The OS uses the NIC driver software to give data to, and take data from, that particular make and model of NIC. The OS then passes the data from and to the Application Layer via Layer 4 using the Transport Layer protocol (UDP or TCP) specified by the Application Layer protocol and coded into the application software by the programmer.

As you continue your study you will be able to expand and clarify these principles for yourself.

And finally, remember the OSI model is a useful reference, it is a means to understanding networking, it is not an end in itself.

Aubrey

_________________
The illiterate of the 21st century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn. Alvin Toffler, "Future Shock" 1970

Wow, thank you for the response. No, I'm not taking CCNA. I am simply working towards my Bachelor's in Web Development, but I'm considering a change to a Networking major (or perhaps pursuing a minor in Networking). I do realize the OSI is only a reference model to help people understand networking; that is why I am working so hard to grasp it.

I thought that packet-switching took place while the data was en route. I am a bit confused. Could you walk me through it layer by layer, if you have the time, that is? We could use an email as an example (I learn best with analogies and examples). If one person sends an email (using an email client like Outlook) to another, that would start in the application layer, which is supposed to be where network services are integrated with the OS. What exactly takes place at the application layer?

It does.

The OSI model principles not only apply to the two end devices (the sender and the receiver), it is implemented by all the intermediate devices between the sender and receiver.

So when the data frame leaves the source device as Layer 1 electrical signals and before those signals are received at the destination they will encounter at least one of the following intermediate devices:
(1) Repeaters - these are not intelligent, operate only at Layer 1 and repeat/regenerate the signals and pass them on - typically used on long distance links these days if required.
(2) Hubs - do the same as repeaters, operate only at Layer 1, but also connect a number of devices together at a central point (hub) - hardly ever used these days.
(3) Switches (or more precisely Layer 2 switches) - have replaced hubs - these receive the data signals (at Layer 1) but then examine the frame of bits at Layer 2 and can make forwarding decisions based on the Layer 2 destination address (MAC address) so the frame doesn't have to be sent (as data bits on Layer 1) to all devices connected to the switch just to the destination; though there are circumstances when the frame is flooded to all devices. Also a swtich only makes forwarding decisions in the context of the local network segment, if the frame is to go to another network, then a router is required.
(3) Routers - receive the data signals at Layer 1, takes the frame of bits (at Layer 2) and then extracts the Layer 3 packet. It then examines the Layer 3 destination network IP address. Then using its routing table it determines where to send the packet. The packet is re-encapsulated into a new frame and then forwarded as Layer 1 signals onto the link that it has selected. This is packet switching - each packet is independently forwarded by the router in accordance with the routing information that it has at that moment - hence packets in the same data stream could take different pathways to the destitnation. This makes for flexible, resilient data communications.

So the key thing is - the OSI model, to varying degrees, applies to all devices on the data network, not just the end devices.

Aubrey

_________________
The illiterate of the 21st century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn. Alvin Toffler, "Future Shock" 1970

Very good question, wouldn't we all like to know that .. but frankly .. *shrug* .. I'm an infrastructure guy these days, I wrote a program once, so I'm over applications

Seriously, you can tie yourself up and become totally confused if you try to dig out intricate details before having a grounding in the big picture stuff. Also remember, no one knows everything about networking (despite what they might claim) so there are some things that you just have to let go.

For example, the extent of my Active Directory knowledge is that I can spell AD, but I'm comfortable with that, it means others can solve those problems.

Aubrey

_________________
The illiterate of the 21st century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn. Alvin Toffler, "Future Shock" 1970

I'll take a stab at this one, the application layer is the user interface layer. So when you double click on the outlook program icon, the mouse generates some CPU interrupts, the CPU processes these interrupts and is directed to load a program, the program written in visual basic makes is easy for people to send email to each other, the program also establishes a communication session between the PC and the Exchange server.

So then you click on new, a cpu interrupt is generated, the CPU processes the interrupt and is directed to present a form to the screen, you type on the keyboard again, generating CPU interrupts that when processed puts characters on the screen in the form fields. you navigate your mouse to different form fields, again generating CPU interrupts that when processed by the CPU, moves your cursor to different areas of the screen so you can type, generating more cpu interrupts that when processed put letters on the screen in the various form fields.

when you click on send, another CPU interrupt is generated which processes some behind the scene subroutine that creates the message, adds some headers, makes the form disappear, and sends it to the Exchange server over the established session between the pc and server.

so that's pretty much what happens at the application layer when you send an email.

_________________
"If you're good at anticipating the human mind. It leaves nothing to chance."
-Jigsaw

Wow. Thank you both very much for your prompt answers.

Aubrey - That cleared up a whole bunch of confusion for me. I was trying to figure out how packet-switching took place before data even left the source. One more question:

Do the functions at layers 4 - 7 happen only in the source and destination host, or are there other devices that use those functions?

ristau5741 - Thank you. That helped to clarify things as well. I am starting to gain a new understanding.

The application layer is huge. How it operates is dependent on what function you're referring to.

For example, HTTP operates at the Application Layer, how it operates is determined by the protocol itself (Or rather, the implementation of it), which you can read up on by reading its RFC.

Different protocols operate... well... differently. So there is no "one way" that the Application layer operates - it all depends on what protocol you're talking about.

A great excercise is to take HTTP for example, read the RFC to learn about the protocol, and then do packet captures in Wireshark while accessing HTTP websites. You'll notice that what you see in the RFC is usually what you get in Wireshark (If you're interested in Networking - LEARN WIRESHARK!). You'll see the HTTP GETs, POSTS, etc. as well as the different HTTP codes... like 301s for redirects, 401 for proxy redirects (IIRC), etc. You'll see the 3-way TCP handshake, followed by a HTTP GET, and usually returned with an HTTP 200 OK and the content you asked for. It's quite entertaining.

That's how I self-taught myself about WCCP (Proxy-specific) when I had no clue how it operated beforehand - I barely even knew what its purpose was at the time.

Just keep in mind, that the RFC doesn't ALWAYS apply - that's where you hear people say things like "Well, not everybody follows the RFC", or "It'd be nice if they followed the standard.", etc. etc. etc.

_________________
Regards,

Steven King
San Diego Cisco User Group - http://www.sdcug.com
"The only time something is impossible is when you think it is." - Kevin Corbin, CCIE #11577

In general for normal data communication streams you can say that Layers 4 -7 only apply at the source and destination devices. But the terms source and destination devices include many different types, including DHCP, DNS and proxy servers which are transparent to the user.

Also switches and routers these days can run web interfaces for configuration purposes, so while they are being configured in this manner they are hosts operating at all 7 layers, but once they revert to their main role of forwarding frames or packets then they only process the data to Layer 2 or Layer 3 respectively.

Aubrey

_________________
The illiterate of the 21st century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn. Alvin Toffler, "Future Shock" 1970

there are layer 4 switches now, these switch traffic at the port level (TCP/UDP) and can prioritize traffic based on application and the ports it uses.

_________________
"If you're good at anticipating the human mind. It leaves nothing to chance."
-Jigsaw

I don't like calling a switch a layer4 switch. The layer a device is at, is the way in which it forwards traffic. i.e. layer3 devices forward traffic based on the layer3 address. Layer2 devices forward traffic based on the layer2 address.

Routers can even look at layer7 (class-map match url as an example) - but I would not call it a layer7 device.

The fact that a device can peek into the upper layers makes no difference.

It can get blurred though. You can have a policy-based route-map which says that traffic with tcp source port 3389 goes a certain way. However the device is still forwarding that traffic with it's layer3 address, and hence it's still layer3 to me

_________________
www.mellowd.co.uk/ccie/

+1

Thank you for this, this is a good way of thinking and I didn't like calling switches layer 4 devices either. Your train of thought makes alot more sense in my opinion.

_________________
Regards,

Steven King
San Diego Cisco User Group - http://www.sdcug.com
"The only time something is impossible is when you think it is." - Kevin Corbin, CCIE #11577

I appreciate all the help and advice I am receiving. Thank you all so much. I understand the OSI model so much better now.

I am sure that there are exceptions for the operation of OSI functions, but my class is a Bachelor-level course about the fundamentals of networking. That means I am beginning to grasp the basics of networking. I have a lot of self-taught knowledge through personal experience, but I never knew the terms for these operations. I called the router "The center of the galaxy" where I could plug in all my computers.

Wait till he learns about layer 2 1/2. heh heh heh - evil laugh

_________________
"If you're good at anticipating the human mind. It leaves nothing to chance."
-Jigsaw

that's just cruel!

_________________
"If you're good at anticipating the human mind. It leaves nothing to chance."
-Jigsaw

Am I seeing things, or did you just quote yourself?

_________________
Regards,

Steven King
San Diego Cisco User Group - http://www.sdcug.com
"The only time something is impossible is when you think it is." - Kevin Corbin, CCIE #11577

Two and a half??? *gulps*