Understand the OSI Model

0:00Hey friends and welcome to the HPE Aruba Networking Certified Associate course

0:05here on CBT Nuggets.

0:07So in this opening skill what we're going to be talking about is some

0:10foundational topics

0:11related to computer networking. More specifically we're going to be looking at

0:16something known

0:16as the OSI model. So let's discuss the seven layers within the OSI model. We'll

0:21discuss what

0:22each one actually does and the first thing we'll begin talking about is

0:25something known as the

0:27application presentation and session layers. So with that said let's just get

0:31to it.

0:00So friends to kick off this skill, what I want to do is to first talk to you

0:03about the concept of modeling and

0:06And no, I'm not talking about a catwalk or a runway type of modeling. What I'm

0:12talking about here is the ability to model

0:15How data is actually transported across a network?

0:18So say for example you on perhaps this machine here wanted to talk to this

0:22server over here. How does the data actually get from

0:26Your machine all the way through all of these different nodes end up in the

0:30server whereby the server can understand your request

0:33And ultimately relay it back with the correct response directly to your machine

0:38without any particular issues

0:40Well as you can imagine for this to actually happen

0:43There is a lot of complexity going on and in order to understand the complexity

0:48We want to break down each component into individual steps and this is what we

0:52're modeling

0:53We're modeling the actual steps taken when we send data from point a all the

0:58way over to point b

0:59Now the most well-known model that was ultimately developed way back in the

1:041970s as the OSI model

1:07Now all OSI stands for is open systems interconnect

1:11And basically what we do describe the connection from say for example

1:15This device here to another device within a network

1:17It has been logically broken down into seven distinct steps

1:22Now let's plant a flag before we go any further. So we have the application

1:26layer

1:26The presentation layer followed by the session layer the transport layer the

1:31network layer

1:32Then the data link layer and then finally the physical layer

1:36So what we're going to do in this particular skill is we're going to talk about

1:39each of these individual layers

1:41And describe what they actually bring to the table how they actually play a

1:45part in facilitating this connection from here

1:48All the way over to here now very very important point to note

1:51These seven layers they actually are not how things really operate in reality

1:56So upon hearing that you might think so if this is not how things are actually

2:01done

2:01Then why even bother doing this at all but quite honestly this is more of a

2:05helpful utility

2:06Whereby we can easily identify layers in isolation

2:10So if you want to troubleshoot a very particular issue say for example, you say

2:14hey

2:14I'm going to troubleshoot a transport layer issue

2:18That refers to a very precise portion of the communication very very different

2:22from say for example examining a data link issue

2:25Now the way things work in reality is actually more closely modeled via

2:29something known as the tcpip model

2:32Now the tcpip model is really quite similar but it happens to combine some of

2:37the layers within the osi model together

2:39So in the osi model we have the application layer the presentation layer as

2:43well as the session layer

2:45And the osi they're all subdivided very very clearly whereas in the tcpip model

2:50all of these different layers are simply just

2:52Aggregated and referred to as the one application layer. Okay. So what does

2:56this layer actually cover?

2:57Well, again

2:58We'll talk about all of these individual components as we progress throughout

3:01this particular skill

3:02Now the next layer in the osi model happens to be the transport layer and the

3:07good news is is in the tcpip model

3:09This one is still the exact same basically the model the exact same thing

3:14similarly

3:14With respect to the osi model we happen to have the network layer

3:18And this is kind of the same thing in the tcpip layer

3:22But it's known as a different name

3:24It's referred to as the internet layer and then finally when it comes to the

3:29last two layers in the osi model

3:31This is the data link layer as well as the physical layer

3:35These last two in the osi model are ultimately combined and just referenced as

3:39the network access layer in tcpip

3:42So ultimately in the tcpip model what we get is this the application transport

3:47internet and network access layer and this much more closely models

3:51How things are done in reality?

3:53But like I said the osi model is still very very important

3:56It's used all the time to reference particular problems within troubleshooting

3:59scenarios

4:00It gives us a model to much more precisely diagnose particular issues

4:04So it honestly is very important that we understand how each layer actually

4:09operates

4:10So to get things kicked off let's go and begin first talking about the

4:13application presentation and session layers

0:00Alright friends, so let's begin talking about the application layer within the

0:04OSI model.

0:05So the first thing to state about the application layer is that it's referenced

0:09as layer 7.

0:11This is the highest layer within the OSI model and you can kind of conceptual

0:15ise this

0:16as the way the end user, such as you or I, actually interacts with the network

0:21itself.

0:22So what exactly do I mean then? Well imagine you happen to be over here,

0:25let's say this is you on your little laptop or computer and you want to go and

0:30access a web

0:31server out on the internet. How on earth would you actually do that? Well what

0:34you could do is you

0:35could open up your web browser and quite simply you just type in the page that

0:40you want to go to

0:41and suddenly you happen to reach it. Now there is a lot going on in that single

0:45transaction there

0:46but the way we interacted with the service itself is via something known as

0:51HTTP which is the

0:53hypertext transfer protocol. More accurately though, you probably will use in

0:58modern terms HTTPS,

1:00this is the secure version of HTTP with encryption. Alright start the bus there

1:05, start the bus.

1:06This particular protocol, HTTPS, this is actually operating at layer 7 of the

1:12OSI model. So whether

1:13you happen to be using Firefox or you happen to be using Chrome or Edge, the

1:17same application

1:18layer protocols are ultimately in play, HTTP and HTTPS. But quite simply hyper

1:24text transfer protocol

1:26is not the only layer 7 protocol that we can use. Say that we as the end user

1:31want to access email

1:32services, okay? So what do we do? Well we can use things like Outlook or you

1:36can use things like

1:37Gmail and ultimately under the hoods they happen to use a whole bunch of

1:41different protocols. Now

1:43we don't really have to talk about these different protocols, certainly not

1:46quite just yet but

1:47just understand that these protocols are actually in place. So say for example

1:50we have SMTP which

1:52is the simple mail transfer protocol. This is how emails are actually sent

1:57between servers and the

1:58way we can actually receive our emails is going to be using things like IMAP4

2:03or maybe even POP3.

2:04Now once again, despite being very very different from hypertext transfer

2:09protocol,

2:09all of these protocols right here they are also operating at layer 7 and in a

2:15similar way if

2:16you want to maybe transfer files from a particular server to another server

2:20over the network once

2:21again. Well there are even more protocols we can use such as FTP which is file

2:26transfer protocol

2:27or we could use SFTP, the secure version of that protocol or similarly we could

2:32use

2:32SCP for secure copy. All of these protocols allow us to do just that copy files

2:38from one location

2:39to another and again these particular protocols are once again acting at the

2:44application layer.

2:45All of them essentially are our bridge as the end user to the underlying

2:49network that we want to use.

2:50Let's talk about the next layer. The next layer as you may imagine is going to

2:54be layer number six.

2:56Now this one here is referred to as the presentation layer. So what does this

3:01one do? Well let me tell

3:02you. So again let's say we are over here on our laptop here and we want to

3:07reach a web server over

3:09here. Let's imagine that this is google.com or something okay. We know via the

3:13application layer

3:14we can use HTTP or rather HTTPS as we would in the modern day and we can

3:18ultimately establish

3:19that connection and get the data back but the data itself has got to be

3:24formatted in a particular

3:25way and we've also got other considerations about how we want the data

3:30presented to us.

3:31Would we like say for example the data to come back in an encrypted form? Well

3:35if that is the case

3:36the data is going to be encrypted on the wire and then decrypted at the end

3:41points so that ensures

3:42that as the data is actually traversing the network it's not readable to any

3:47type of eavesdropper

3:48but it also means that the data is readable to you on your end station when you

3:52get it back on your

3:53browser. The handling of encryption and decryption well this is going to be

3:58handled via the presentation

3:59layer. It's how the data is going to be presented in a similar way. Perhaps we

4:04want to have the data

4:05compressed. How on earth is this going to be handled once again? This is just

4:09yet another task

4:10of the presentation layer in a similar way. What about the data actually being

4:15formatted? Say for

4:16example is the data going to come back in ASCII or maybe the data is coming

4:20back as a JPEG maybe

4:22the data is going to be presented as an mp3 or maybe even the data is coming

4:27back in some type of data

4:28format like JSON or maybe even XML. Now it doesn't matter if you know what JSON

4:33and XML are absolutely

4:35not but not focused on that at all. These are just file formats that are very

4:39popular when

4:40machines want to talk to other machines. That detail isn't actually what is

4:44important.

4:45The important part is that whether or not we're using JSON or XML or we're

4:48dealing with jpegs and

4:49images or ASCII and regular text. All of that stuff along with encryption and

4:54compression this

4:55is once again all handled via the presentation layer number six which then

4:59leads us to the next

5:01layer that I just want to talk to you about and you already know the number

5:04this is going to be

5:05layer number five. So this one here is known as the session layer and the name

5:10itself really is

5:12quite descriptive. It's going to allow us to manage connections between the two

5:16endpoints. So this

5:17device here talking to this device here. So in fact you know what let's maybe

5:20say you are here

5:21and I am over here okay. We want to connect over the network and have some type

5:26of video

5:27conferencing call okay. Now in order for this connection to actually begin we

5:31've got to do a

5:32few things. We've got to first establish the connection with one another using

5:36things like

5:36authentication. So we're actually talking to the correct person when we happen

5:40to be talking we

5:41want to ensure that the audio and the video is actually correctly in sync with

5:45one another

5:46and if we happen to get some type of glitch in the network and the connection

5:50temporarily drops

5:51we do not want to have to start the meeting all over again instead we would

5:55much rather just have

5:56the application itself just kind of figure that out. Everything I've just said

6:00here

6:01is ultimately controlled via the session layer layer number five. So that is

6:05our introduction

6:06into the application presentation and session layers layers five six and seven.

6:11Let's now move

6:11on and talk about the transport layer layer number four.

0:00So friends, the next layer of the OSI protocol, OSI protocol, OSI model, this

0:06one is known

0:07as Layer 4, otherwise referred to as the transport layer.

0:11The whole purpose of the transport layer is really about facilitating what is

0:15known as

0:16end-to-end communication between our devices.

0:19So really we're talking about how connections between devices are established,

0:23how they

0:24actually are maintained throughout their life cycle, as well as how we can tear

0:27down that

0:28connection once the data transfer is over.

0:31Now just for clarity and posterity, the data at this particular layer is

0:36referred to as

0:37segments.

0:38Alright, first things first, very, very important.

0:42Two well-known protocols happen to operate at this layer.

0:46The first one is known as TCP.

0:48The other one, this one is referred to as UDP.

0:51So the name for TCP or rather the acronym for TCP actually stands for the name

0:57Transmissions

0:58Control Protocol, really, really important protocol, very, very useful.

1:03The UDP protocol, this one is referred to or rather stands for User Datagram

1:08Protocol.

1:09Alright, so apart from having different names, what actually is the use case or

1:14the purpose

1:15of having two distinct protocols to work at the same layer?

1:19Well, as a high-level overview, this one here is the one we want to use when we

1:25want

1:26to establish a reliable connection.

1:28So when the data between sender and receiver, we want to ensure that the

1:31receiver gets all

1:32of that data, and if any data happens to be lost or corrupted, we want the

1:37protocol to

1:38notice this and to re-transmit that data so that ultimately it can be rebuilt

1:43at the

1:43end point and nothing is lost.

1:46When we want to do this, we're going to be using our friend, Transmission

1:50Control Protocol.

1:51Now you can maybe imagine that this is a little bit involved to ensure that all

1:55the

1:55data is always sent and if it's not, we re-transmit.

2:00The drawback of doing this is that ultimately we get a slower connection and

2:04this also actually

2:05adds overhead to the connection itself.

2:07So more data slows things down.

2:10This is the price we pay for reliable transmission.

2:14So you might be able to imagine what the benefits and drawbacks are for UDP.

2:19UDP is not reliable in the sense that it's not going to re-communicate any type

2:24of lost

2:25or corrupted data.

2:26UDP is kind of like a fire hose just going blaaaand just sending all the data

2:31all at

2:31once it doesn't care if any gets lost.

2:33It's just fire-hosing this data towards the end point destination.

2:37Now whilst this might sound a little bit reckless, it might sound sub-optimal,

2:41the use case here

2:42is that because UDP doesn't concern itself with the re-transmission of data, it

2:47's also

2:48going to be much more lightweight.

2:50So now we have a faster connection with less overhead.

2:54Now for clarity here, you as the end user, you're not going to be concerned

2:58with picking

2:59a particular protocol.

3:01You're not going to say hey I want to use TCP, hey I want to use UDP instead.

3:05This is going to be handled via the upper layers within the OSI model.

3:09The application layer is going to handle this for you transparently.

3:12Now okay, very very important.

3:15This part here really lets you know that you're working with the transport

3:19layer.

3:20Remember we're talking about utilising particular ports, we're talking about

3:25operating at the

3:26transport layer.

3:27Now just to be clear here, we are not talking about physical ports, you know

3:31like taking

3:32a cable and plugging it into the port of a computer.

3:36This is not the same thing at all.

3:37We are talking about logical ports, these are just end software and used to

3:41ultimately

3:42differentiate between different services.

3:45So let's imagine this, okay?

3:46Here I am over here, this can be my little personal PC within my local network.

3:52And let's go with all the way over here, this server, let's give it the IP

3:56address of let's

3:57maybe go with, so I don't know 192.168, let's go with 2.10.

4:02So I over here want to talk to this server over there.

4:06Now this server actually has multiple different services that it actually

4:10supports, so it's

4:11got web services.

4:13So it's going to allow connections over HTTP.

4:16It also allows for secure web access, so it also supports HTTPS, the secure

4:22version with

4:23encryption.

4:24Let's maybe say it also supports the ability to handle SSH connections which is

4:29often used

4:30for secure management of a server, as well as a whole bunch of others.

4:35We're not too concerned about the details of what it might be actually running.

4:38The thing I want to highlight here is that I over here want to connect to this

4:42server

4:42and I'm going to connect over this IP address, what is going to differentiate

4:48between if

4:48my connection to this server, if say for example I'm using a web browser to

4:53manage the connection,

4:54how do I differentiate whether I'm going to be connecting over HTTP or I'm

4:58going to wrap

4:59the connection and encryption and use HTTPS.

5:02Well, drum roll please, the way we're going to do this is via the use of our

5:06logical ports.

5:08The connection is not just specified with the IP address, we also add on a

5:13little colon,

5:14let me just remove this, and then specify the port number, so if you want an un

5:18encrypted

5:18connection, we specify port 80.

5:21And now, we get the connection to the server, it's unencrypted and anyone can

5:26read it.

5:27If we're not happy with this, what do we do?

5:29Specify the logical port of 443 over the transport layer once again, now the

5:34connection

5:35between these end devices is going to be fully encrypted and protected from any

5:39potential

5:40eavesdroppers.

5:41Now, one thing to state here is that when I'm specifying these port numbers,

5:45that is not

5:46the full story.

5:47In the case of say for example HTTP, whilst it does use port 80, it also uses

5:53TCP port

5:5480.

5:55Okay, this is very, very important.

5:57This is going to be a reliable connection.

5:59Any errors or any type of loss data is going to be retransmitted.

6:03This is the standard here.

6:05Similarly for using HTTPS, you can imagine it's going to be port 443, but also

6:10wanting

6:11that reliable connection, it's going to be TCP port 443.

6:15In fact, even the other example I was giving to you, SSH, this is going to be

6:20port 22.

6:21Once again, though, it also uses TCP.

6:23So you might be thinking, John, does everyone just use TCP?

6:26Is there ever a use case for UDP?

6:28And the answer for that is absolutely yes, very, very often.

6:32If you happen to be dealing with say, for example, real time data, you can

6:37imagine if

6:37you are, let's go with an a zoom meeting, okay, some type of teleconference.

6:42In this case here, we are going to be utilizing UDP for the transport.

6:47Here's the reason why.

6:48For a start, UDP is lightweight.

6:50It's fast.

6:51We want that for real time communication, but also think about it.

6:54In the event that we lose data, let's just draw it out.

6:57Once again, let's say here I am over here.

6:59Here you are over there.

7:00Let's say we are connecting over, let's say, Skype to our devices.

7:04If along the way, as we are streaming data to one another, let's say some of

7:07that data

7:08happens to be temporarily lost or corrupted.

7:11You know, something happens in the network.

7:13There's a little blip.

7:14Some data doesn't actually make it from you to me.

7:17So the net effect of that is I, as the end user, I am watching my screen, I see

7:21your face.

7:22I can hear you talking and maybe in the actual audio, I get a little bit of

7:27distortion in

7:28the sound.

7:29But you've all heard it before.

7:30The picture resolution maybe dips a little bit temporarily.

7:33The audio becomes a little less clear.

7:36But then the connection resolves itself.

7:37The HD comes back.

7:39The actual audio comes back, nice and clear.

7:41Nobody actually cares.

7:43What use therefore do you really have for retransmitting that lost data?

7:48Think about it.

7:49You and I are having a conversation in real time.

7:51Let's maybe say half of a word that you said gets cut out or clipped.

7:56The conversation is continuing on.

7:58Does it really make sense to retransmit that data and get the missing word five

8:03seconds

8:04later and then just insert into the communication.

8:07Doesn't make much sense at all.

8:08So in this case here UDP is actually the way we're going to handle this end to

8:13end communication

8:14between these two end devices.

8:17Speed and low overhead is all we're actually after.

8:20Now here's another example.

8:21Let's maybe say that I want to access this particular server.

8:26Okay, let's maybe say it's a web server.

8:28Now this web server is let's maybe say we just know the name of the actual

8:32address.

8:32So let's say it's called server one dot CBT nuggets dot com.

8:37Now the way we actually communicate is via IP addressing.

8:41But the problem is is that I don't know the IP address of this particular

8:44server.

8:45The way we actually get this is by taking the domain name we see right here and

8:49we resolve

8:50that to the servers actual IP address.

8:53The way we resolve it is via the main name system, a DNS request.

8:58Now we will talk about DNS in more detail later on.

9:00But for now just assume that let's go with this server right here.

9:04Okay, this one on my local network is my DNS server.

9:07It's just there to resolve these types of names to IP addresses.

9:11So effectively I make a connection to my own server here, the DNS and I ask you

9:17the question

9:17hey, what is IP address associated with this particular name?

9:22Now you see that little request I just made right there.

9:25This was also made over UDP and it's been made over UDP port 53.

9:30This is the well known port for DNS requests.

9:34So what the server ultimately does it responds back over the UDP connection

9:38gives my device

9:39the IP address information.

9:41Now I know the IP address I can therefore make the connection over TCP this

9:46time to the

9:46web server.

9:47Over let's go with port 443 so it's nice and encrypted and fully secure.

9:52Now just notice all the nuance here.

9:55Sometimes we're asking for a very particular request such as DNS.

9:59How we doing that?

10:00We're using port numbers, we're using the transport layer when we want to build

10:03the

10:03connection to the web server whilst we are also using other layers such as the

10:07network

10:07layer but also relying on that transport layer.

10:11In this case here utilizing port 443.

10:14So really when we're talking about the transport layer we're talking about port

10:17numbers we're

10:18talking about two very particular protocols UDP and TCP we know their

10:22advantages and disadvantages.

10:25One is reliable, one is very fast but that is something very very important

10:29about this

10:30reliability that we get with TCP.

10:32The way this is actually instantiated is via something known as a three way

10:37handshake.

10:37I know that sounds pretty pretty cool.

10:39So how about we talk about that in the very next video.

0:00Alrighty then, so let's talk about the three-way handshake.

0:03Now, first things first, to be clear, the three-way handshake is a function of

0:09TCP.

0:10This is not something we're going to be doing when establishing a connection

0:14over UDP.

0:16So let's talk about how the actual mechanics work then.

0:19So let's imagine we happen to have a machine on the right here, on the right,

0:23on the left here.

0:24Let's just call this PC1.

0:26Okay, super creative name, I know.

0:28And on the right hand side, we have PC number two.

0:32So here's the deal.

0:33PC1 wants to talk to PC2.

0:36We want to have a reliable connection over TCP.

0:39So we're going to have to establish that everyone is a okay with this

0:44connection coming through.

0:45The way we do this is by establishing the three-way handshake.

0:49So it consists of three particular parts, no real surprise there.

0:53The very first part is something known as a SIN message.

0:58And as a SIN with AY, it's basically a message for synchronization.

1:03So the first thing is we do is we send off the SIN message to what will be a

1:07receiver.

1:08And the receiver, if everything is AOK, we'll send their own response back.

1:13And this response is going to be something known as SIN.

1:16Now, the SIN portion might be self-explanatory.

1:20It's the exact same type of message as we just saw above.

1:23The difference though is we have this additional acknowledgments.

1:27Now, before we go on, any further, let's just close off this three-way

1:31handshake.

1:31Once the initiator receives the SIN ACK, what is then sent is the third and

1:38final component of the three-way handshake.

1:40And that's just simply an ACK message, basically an acknowledgement.

1:44So in a regular conversational style way, what is happening is PC1 says,

1:50"Hey, Mr. PC2, I would like to talk to you. Are you ready for this?"

1:53PC2 is up for a good old chat and PC2 says, "Yup, I hear your request and I'm

1:59ready for it."

2:00Would you also like to talk to me, hence its own SIN message?

2:04And PC1 goes, "You betcha. Let's begin this conversation. Here is your

2:08acknowledgement."

2:09So once this has happened, both sides are waiting and ready to participate.

2:14PC1 can now begin sending data to PC2 as they expect.

2:19Now, these are the three main components of the actual handshake,

2:23but there actually is something else that we're missing out here that is

2:27something known as a sequence number.

2:28And this is going to kind of play into the fact that TCP has reliability.

2:33Remember, if we happen to lose a packet or have a packet corrupted,

2:38the TCP protocol is going to be built in to handle that type of issue.

2:42So let me just describe it a little further with a little bit more detail.

2:46Once again, PC1 on the left, that's the sender, PC2 on the right, the receiver.

2:52So we already know the first message is going to be a SIN packet,

2:56but this packet is also going to include an initial sequence number.

3:01And this is just in numerical value chosen by PC1.

3:05Now, the value itself that happens to be chosen isn't really that important.

3:09Whether the number happens to be 1000 or the number happens to be 5 or the

3:13number happens to be 3520,

3:16the initial number is not what is important.

3:18What happens to that number though is, so let me describe it.

3:22So let's imagine in this packet here, we send the sender quest and we also have

3:27this sequence number.

3:28So let's just say this is sequence number.

3:31Let's go with 100.

3:32Nice simple number, okay.

3:33Now we know PC2 is going to respond with its own SIN Act.

3:37So let's just draw out the packet right here.

3:39This is SIN Act coming back.

3:41Now, similar, this particular device here is also going to choose a randomized

3:46numerical value for its sequence number.

3:49So let's just imagine that this one happens to use the sequence number of let's

3:53go with 5000.

3:54So we'll say sequence is 5000.

3:57But as a part of the Act message is going to send an numerical value associated

4:01with that Act and that is going to be

4:04the value of the initial sequence number that was sent and be simply incre

4:09mented by one.

4:10Here's what's happening.

4:11PC1 over here chooses a number at random.

4:14This is 100, no case here.

4:15That is then communicated to PC2 in the form of the initial SIN request.

4:21And in the response, we take that number and add one to it and the

4:24acknowledgement.

4:25Pretty much what we're saying here is that we are ready to receive the packet

4:30with sequence number 101 next,

4:33which is the next one in line that PC1 would send.

4:36We just incremented by one.

4:37But remember the acknowledgement is only one portion of the SIN Act.

4:42We've also sent our own sequence number right here.

4:45So can you predict what the response from PC1 and its acknowledgement message

4:50is actually going to be?

4:52Well, like we say, we know we're going to get the Act back right here.

4:55So in fact, let me just draw this here as the packet coming back at the Act

4:59message.

5:00Now, that actually is two numbers that are going to be associated here.

5:03There is going to be a sequence number.

5:05And what do you think that sequence number is going to be?

5:07It's going to be the exact same value as the previous Act right here because

5:12remember in this message,

5:14PC2 is saying, hey, the next thing I'm expecting, not actually have received,

5:18but I'm expecting will be sequence number 101.

5:22And this is exactly what we're getting here.

5:24Sequence number 101 is indeed the one that's been sent.

5:27Now, we also have to remember for sending an acknowledgement and what is it?

5:31But actually acknowledging we're saying, hey,

5:34I saw your sequence number of in this case, you're 5000.

5:38The next one I would expect would be 5000 and one.

5:42And then from this point on, once the connection has been established,

5:45we can then begin sending our data with these sequence numbers intact.

5:51All right, so let me just try to simplify this a little bit more so we can

5:54actually just identify

5:55how these sequence numbers help detect errors.

5:58OK, so PC1 on the left PC2 over on the right.

6:03PC1 begins by sending over a message.

6:06And let's for now just really focus on the sequence numbers.

6:09OK, so we send the packet over with sequence number of let's go with,

6:12I don't know, let's say number 12.

6:14OK, what is that act coming back from that then?

6:17Well, the act message to acknowledge the previous sequence number is going to

6:21take the number plus one.

6:23So we're going to send an act back of 13 pretty much saying, hey,

6:26I got your message with sequence number 12.

6:29The next one I expect from you will be 13.

6:32Now let's just say this particular device, PC1 sends a bunch of messages

6:36and for some reason they don't get through.

6:38OK, so PC1 then sends another message which finally does get through.

6:42However, it's got a sequence number of let's go with 16.

6:45Now this device here is going to notice, hey, you just sent me a message of

6:51sequence number 16.

6:53The number I was actually expecting was 13.

6:56Therefore we have a mismatch here and I can interpret that as at some point

7:01between this point here and that point there,

7:04there was some type of loss data.

7:06Hey, can you resend the information back because I'm expecting the packet with

7:10sequence number 13.

7:12And in that case here, this device here can begin resending these packets

7:16and everything is OK once again.

7:19This is the beauty and the reliability of how TCP actually operates.

0:00So the next layer we are going to be talking about is, you guessed it, it's

0:03going to be layer

0:04number three. This is the network layer and it's so, so important. We're really

0:09going to do a lot of

0:10focus on this layer as we progress throughout this particular course. So what

0:14is it all about?

0:15Well, you might already have a little bit of an inclination as to what we're

0:18trying to do here.

0:20The network layer, this is where we're going to do things such as routing. All

0:24right, but what

0:25does that even mean? Well, as you happen to be watching this, no doubt within

0:29your home you

0:30happen to have a router and you are probably familiar that this particular

0:35device let's you go to the

0:36internet. Now to be clear here, I'm not saying that routing only involves your

0:41visiting the internet,

0:42more specifically when we're talking about routing, we're talking about moving

0:47data across different

0:49networks. Now first thing to point out right here, see this term I happened to

0:53use, I said data

0:55this is not really the terminology we would use when we're talking about layer

0:58three,

0:59the data layer three is actually referred to as packets. So more accurately

1:03what we're doing when

1:04we're routing, we're moving packets across from one network over to another.

1:08Now when you happen to

1:10visit the internet, all the internet is is a whole bunch of connected networks,

1:14you know,

1:15all over the world, all over the globe. So let's imagine this is now your

1:18little home, okay?

1:19And this is your home router. So let's imagine this is your computer here. And

1:25this one can be

1:26your sister's computer, you know, you've got one computer in each bedroom maybe

1:30. Let's pretend

1:31that this here is not a server but rather it's a printer. So because this

1:35printer here is in the

1:37same network as you over here, if you want to print something on that printer.

1:42So in this case

1:43here, there actually is no routing involved. And the reason why there's no

1:47routing involved

1:48is because all of this communication you to the printer is all within that same

1:52local network.

1:53Whereas if you want to go and let's maybe say this is google.com, this is any

1:58separate network

1:59entirely, you ultimately have to send that packet to your router. And the

2:05router will figure out

2:06what to do with it. Most likely what it's going to do is just send it to your

2:10ISP, the people who

2:11you pay for your internet and then their routers will ultimately find the

2:14correct path to get you

2:16to the server that you want. Thus facilitating the connection that you're after

2:19. So that is the big

2:20big distinction when we want to talk to a different network. We're going to be

2:24operating at layer three

2:25with an OSI model. We're going to be routing and very, very important the way

2:30this is actually handled

2:31or the way this is actually facilitated is via the use of logical addressing.

2:36Now these logical

2:37addresses, but it might sound a little bit complex right now, you no doubt have

2:41heard of these at

2:42least in passing the logical addresses we use at layer three are known as IP

2:47addresses.

2:48This is how the connection actually happens. So your computer here will have an

2:52IP address.

2:53And if you want to talk to your printer in this case here that will have its

2:59own IP address,

3:00and as we'll get to learn because in the same network right here, the IP

3:05addresses of both the

3:06printer and your router are going to be within the same network range. What

3:09does that mean?

3:10Again, do not worry about that quite just yet. We'll talk about the details

3:14very, very soon.

3:15However, when we want to speak to a server out in Internet such as let's go

3:20with google.com once

3:21again, it is also going to have an IP address. But because it's not in this

3:25network, it's going to

3:26have an IP address and a different network range. And thus the routers here are

3:30going to go figuring

3:31out the best path how to reach that particular IP address. This is what we're

3:35talking about

3:36this particular layer. So the address portion of the name might be quite self-

3:40explanatory.

3:40The same way the mailman can deliver your mail because you have a particular

3:44house address

3:44within a particular region. This is the same type of concept that we're talking

3:48about here,

3:48but this portion might not be quite so obvious. IP, what does this actually

3:52mean? Well, we're

3:53actually referring to the Internet protocol. Now it is worth just mentioning

3:58that IP is not the

3:59only protocol you can get in layer three, but it is by far and away the most

4:03popular one. The one

4:05you're going to see all the time at this particular layer. Now once we dive

4:09into IP addressing in a

4:10little bit more detail, we'll get to learn that there actually are two

4:12different versions of IP.

4:14And one is known as IP V4. This is version four of the IP protocol. And we also

4:20have IP V6 version

4:22six. This is the older version. And this one is the newer one. This one is

4:27still really quite common.

4:29And it is a little bit easier to use. This one here though is more modern and

4:33also provides

4:33some distinct advantages. Again, we'll talk about the details very, very soon.

4:37But really at its core,

4:38when we're talking about layer three, the data is referred to as packets. When

4:42we're moving these

4:42packets across different networks, we are routing using routers. The protocol

4:47on play is going to be

4:48the IP protocol, which can be in version four or version six. And these

4:53protocols ultimately allow

4:54us to address each individual node on our networks uniquely so that it can be

5:00found such that if

5:01this device wants to talk to this device, we can do so. And in a similar way,

5:06even if we happen to be

5:07trying to connect to a different network entirely because of IP addressing that

5:12connection can be

5:13facilitated. Now another thing that is managed at layer three and something we

5:17will talk about

5:18is something known as quality of service. So what does this actually mean? So

5:23let's say that this

5:24is a corporate network, we can imagine maybe there are a lot more computers

5:28than just two.

5:29And let's say for this particular company, video calling and video conferencing

5:33is very,

5:34very important. Now let's imagine that the users within this particular network

5:39are kind of slacking

5:40off on the job and spending a lot more time watching YouTube so there is

5:43visiting the internet,

5:45they're consuming lots of video, full HD, maybe it's 4K, they're just consuming

5:49lots and lots and lots

5:50of data. What this means though, is that they're ultimately putting a lot of

5:54demands on their own

5:56network. So the actual bandwidth availability is going to be reduced for other

6:00users because

6:01some users are well taking more than their fair share of the network. And

6:05because of this,

6:06users who really are doing work such as doing video conferencing, they are

6:10noticing a lot of lag

6:12on their systems. The network really seems to be congested and slowed down. So

6:16what can also happen

6:17at layer three is we can actually identify very particular traffic traversing a

6:22network such as

6:23traffic that is used for video calling or voice calling. And we can just say,

6:27hey, we want to give

6:28that traffic a very particular priority. So even if all of the other users on

6:32the network are wasting

6:34time watching YouTube, what we can ultimately do is prioritize any of the

6:38traffic that is actually

6:40used for video calling so that the people who are watching YouTube, they are

6:44the ones who suffer

6:45congestion, they are the ones who start noticing lag. So if this person here is

6:49actually doing a

6:50video call, they're actually going to be able to make that connection without

6:54much lag at all.

6:55Whereas the person on YouTube, they might start noticing that all of their

6:59videos are

7:00buffering and not really watchable. Now again, we'll talk about how we can

7:04actually implement

7:04quality of service much later on within this very course, but understand though

7:09that this

7:09particular concept, once again, all of this is happening at layer three within

7:14the OSI model.

7:15So routing forwarding packets between different networks, marking particular

7:19packets for a very

7:20particular quality of service, all of this made possible, via layer three

7:24within the OSI model.

0:00Alright friends, so let's round this scale off by talking about the final two

0:03layers.

0:04This is going to be layers number two, as well as layer number one.

0:08So let's focus on layer two first.

0:10This one here is referred to as the data link layer.

0:14So what is this particular layer actually responsible for?

0:18Well, its primary purpose is to reliably transfer data over a physical medium.

0:24Quite simply, this is how all of our local connections happen.

0:28Now what I mean when I say local connections, I'm talking about connections

0:31between the devices

0:32within the same network.

0:34So remember the way routing was allowing us to facilitate communication across

0:38different networks

0:39right here, we're talking about devices in the same network.

0:42So if I or my little machine here want to talk to my printer on this same local

0:46network,

0:47the big thing that's going to allow us to do this is the data link layer.

0:50Now remember, we have things such as segments, layer four, we have packets at

0:54layer three.

0:55When we're talking about layer two, the actual data is going to be encapsulated

0:59into what is known as

1:00frames. Now the data link layer is responsible for quite a lot of different

1:05things.

1:05We're going to talk about this in more detail in an upcoming scale,

1:08but as a high level overview, some of the things it's going to do are,

1:12well of course it's going to manage the actual framing of the packets.

1:15It's going to allow us to detect errors within these frames so error detection

1:18is also covered.

1:20It'll manage things such as flow control to reduce congestion across the

1:23physical lengths,

1:24but one of the most important things, maybe something you've already heard of,

1:28is it actually is responsible for physical addressing. So remember we talked

1:32about IP

1:32addressing, how that was logical addressing. This time here we're talking about

1:36the physical

1:37connections and I really mean physical. So again, let's imagine I am over here.

1:42If I want to talk to this printer over here, I have an actual interface right

1:47here.

1:47This can be my local ethernet interface as the diagram shows.

1:50And this interface here, this physical interface, this is the interface from

1:55which I can reach the

1:55server. Now each of these interfaces, this one here and this one there, they

2:00actually have physical

2:02addresses. And in a similar way, this device here in the middle is switched,

2:05which is something

2:06we'll look at in great detail very, very shortly in an upcoming scale. This is

2:10going to manage the

2:11layer to communication utilizing these physical addresses. In fact, all of the

2:15ports here, this

2:16one here, this one, this one, and this one, each one of these interfaces all

2:20have their own physical

2:22address. Now this physical address I'm referring to, you may have heard it

2:25known as a MAC addresses,

2:27is what we're talking about here, okay? But it also has other names such as

2:30simply just a layer

2:31to address. And sometimes it's just referred to as a burned in address. Now

2:36with respect to IP

2:37addressing, at least in IP before addressing, the IP addresses look something

2:41like this, 192.168.1.1.1.

2:44This is a valid IP for address. When we're talking about addresses for MAC

2:49addresses,

2:50the format looks a little bit different. There are going to be 12 characters

2:54long, and they're

2:54actually going to use hexadecimal characters. Again, we're not going to

2:58struggle and worry about

2:59what this actually looks like quite just now. But to give you a rough idea,

3:03this would be a

3:04type of MAC address. Now I've just made this MAC address up completely off the

3:07top of my head,

3:08but you can see here it does look really quite different. If you count the

3:10characters, we're

3:11going to see we actually have 12 right here. And like I said, each of these

3:14characters are going

3:15to be hexadecimal. So in the example of this local network, when I want to talk

3:20to the printer

3:21right here, my interface here is going to have a MAC address in a similar

3:25format to what we just

3:26saw there before, as well this interface right here. Now the switch itself

3:30actually is quite

3:31intelligent with respect to this matter, because what happens when a frame is

3:35created, we actually

3:36specify a source MAC address within the frame, basically who originated the

3:41frame itself.

3:42And we also include a destination MAC address, much like with respect to your

3:46packets, we have

3:46a source IP address and a destination IP address. Same type of thing is true

3:50here. So if I specify

3:52that, hey, the source is here, and the destination is this one right here. Once

3:56we pass that information,

3:57this frame over to the switch itself, the switch is going to have a little

4:01table that maps all of

4:03the MAC addresses to a particular port. So when this switch here receives the

4:07frame coming in,

4:07it's going to look at the destination. And it's going to say, hey, whenever I

4:10see this MAC address

4:11destination, I know what I do is I actually forward the traffic out of this

4:16particular interface right

4:17here. So I don't forward it out of this, I don't forward it back down here or

4:21this one, I know

4:22that the destination MAC goes out of this interface right here. And ultimately,

4:26it ends up at the

4:27correct destination. Now again, the details of how the switch knows how to do

4:31this. This is

4:32something we'll talk about very, very shortly in the next scale. But

4:35effectively, this is what is

4:37happening. So the data length layer is all about dealing with frames. We use

4:40source and destination

4:42MAC addresses. This is how we communicate over a local network. And then the

4:46data before it's

4:47transmitted is encapsulated one more time. This time at layer one, you guessed

4:52it. This one is known

4:53as the physical layer. Now this might sound really quite similar to what we're

4:57just discussing.

4:58But in this case here, we are being a little bit more literal. We're talking

5:03about the physical

5:04transfer of the data, the actual binary bits, you know, as in the actual binary

5:09bits, the ones and

5:10zeros that are being transmitted. So whether or not the data has been

5:14transmitted as voltage over a

5:15copper wire ones and zeros, or perhaps maybe it's been transmitted over a fiber

5:20optic cable and

5:21pulses of light. So when the light is on, we have a one and the light is off,

5:24we have a zero,

5:25or maybe even the data has been transmitted over radio waves, which is what you

5:29would see with

5:30respect to Wi-Fi or maybe even Bluetooth. So cables, network interface cards,

5:35ethernet ports,

5:36all of this information relates to the physical layer within the OSI model. So

5:40really, when we

5:41have a device here wanting to talk to a server over there, the data that we're

5:46going to be

5:46transmitting is going to be encapsulated all the way from layer seven to layer

5:51six and layer five.

5:52And then we're going to wrap it any segment at layer four, we'll wrap that in a

5:56packet and then

5:57wrap that once again in a frame before ultimately transmitting all this data

6:02over the physical media

6:03as ones and zeros utilizing layer one. And then eventually when we hit our end

6:08points,

6:09the process of decapsulation is ultimately going to do the same thing in

6:13reverse this time stripping

6:15away the physical layer to reveal the data link layer to reveal the packets,

6:19the segments,

6:20all the way up to layer seven, whereby the application can actually read the

6:24initial data itself.

6:25All of the complexity we've just talked about, the binary bits, the physical

6:29medium, whether it's

6:30radio waves or pulses of light, the routers routing packets with layer three

6:34addresses,

6:34the switches forwarding frames with layer two addresses, all of this complexity

6:38, each individual

6:39component, all of it can be represented theoretically in this nice clean way

6:45utilizing the seven layers

6:46of the OSI model.