The OSI model

For anything to work properly and for us to understand how things work, we need to have some sort of standards or blueprints that will allow us to clear the concepts of how particular objects interoperate with each other.

So, for us to be able to network with different types of devices and understand what it takes to get information from a source to a destination, the International Standards Organization (ISO) came up with a conceptual blueprint called the OSI model. This model is in a seven-layer approach that helps us understand this concept and allows vendors to create devices that can interoperate with each other.

This conceptual layered blueprint gives each layer a responsibility; each layer has a job to do, specific to that layer. You can think of it as a company; every business has departments and each department is responsible for a specific role that the company requires to operate smoothly. If any department within the company fails to do their job, the company will fail to carry out its primary objective.

The cool thing is you can change employees within the department and, if they are trained or at least knowledgeable in their respective field, it will not affect the outcome of what that company is trying to do.

The same goes for networks, each layer of the OSI model has a job to do and if vendors make changes to one layer, it won't affect the other layer from doing its job.

Let's go ahead and look at this seven-layer OSI model:

Layer number	Layer name	Brief description
7	Application	Works closest to the user, data
6	Presentation	Deals with the format of the data
5	Session	Keeps different applications' data separate
4	Transport	Provides reliable or unreliable delivery of information, segmentation
3	Network	Provides logical addressing, which routers use to route traffic, packets
2	Data Link	Deals with frames, error correction, and uses the MAC address to access media
1	Physical	Deals with bits, voltage, cabling

Now that we have seen the OSI model, for certification purposes, you must know each layer number and name, not to mention be able to recognize or define what job that layer is responsible for.

So, let's break down the OSI model into two parts: the upper layers and the lower layers.

The upper layers

Looking at the following three upper layers, we can understand that these layers work closest with user interaction, and how it will communicate with other end devices.

So, let's start defining each layer, starting from the top and working our way down:

Layer number	Layer name	Brief description
7	Application	Works closest to the user, data
6	Presentation	Deals with the format of the data
5	Session	Keeps different applications' data separate

The Application layer

This layer is the closest to the user, because it is the interface between an actual application and the next layer down.

People get confused with this layer because of its name. It does not mean that an application lives at that layer, such as IE or MS Word; it is the interface that allows the user to interact with it.

Any time we use any browser or Office application, the Application layer is involved, but that is not the only thing the Application layer does, it makes sure that the receiving end is ready to communicate and accept your incoming data.

So, for certification purposes, we need to remember the protocols that work on this layer: HTTP, HTTPS, FTP, TFTP, SNMP, DNS, POP, IMAP, TELNET, and any network service looking for communication across a large network.

The Presentation layer

This layer's function is very simple to remember: it is responsible for data translation and code formatting. When devices transmit information, it is coded in a certain format; an example used everywhere is ASCII, so when the data gets to its destination, it needs to understand this format, it should be able decode the ASCII, and present it to the Application layer so the user will be able to read it. A simpler example would be an Excel spreadsheet, or a picture taken with a proprietary software that you don't have. If you do not have the software installed on your computer and someone sends you a file with an
extension of .xls, .doc, .ppt, and so on, your operating system will not understand it and simply place a generic icon wherever you save it, and if you try to open it you will get a dialogue box asking which program you would like to use to open the file with.

The Presentation layer is also responsible for key functions, such as data compression, decompression, encryption, and decryption.

The Session layer

The common definition for this layer is setting up, managing, and breaking down sessions between Presentation layer objects, and keeps user data separate. So, basically, it is like having a dialogue control while monitoring the type of mode the client/server communication has, such as full-duplex or half-duplex communication.

Full-duplex communication is pretty much like a conversation you would have with a person, or over the phone – it is two-way communication. Whereas, half-duplex is like a walkie-talkie; you talk, then you listen. So, you can either send or receive at any given time.

The lower layers

Simply stated, the following layers define how information will be transmitted from the source to the destination:

4	Transport	Provides reliable or unreliable delivery of information, segmentation
3	Network	Provides logical addressing, which routers use to route traffic, packets
2	Data Link	Deals with frames, error correction, and uses a MAC address to access media
1	Physical	Deals with bits, voltage, cabling

We now have a better understanding of the OSI model. By breaking them up into two parts, we can see the overall picture of what they are trying to achieve. But we must go in deeper and break down the OSI into its individual layers.

The Transport layer

This is the layer that segments and reassembles data. Services that live on this layer take all data coming from the Application layer and combine it into a succinct data stream.

This layer holds two very important protocols: the Transmission Control Protocol (TCP) and the User Datagram Protocol (UDP).

The TCP is known as the connection-oriented protocol, which means it will provide reliable transmission, compared to UDP, which does not.

Let's define what exactly Connection-Oriented Communication is. In reliable transmission, we have something called a three-way handshake. The process consists of the source sending a SYN packet to the receiver. If the receiver is ready, it will reply with a SYN/ACK, and then the sender replies with an ACK, then communication can occur and transfer data.

Let's see a visual:

Remember that your topologies and your internetworking devices have a lot do with it as well. In a Star topology, everyone is connected to a central device. If you use a hub, you are in a shared collision domain, running at half-duplex and it's Ethernet, which uses the CSMA/CD access method.

Your network will burn to the ground in no time. That is why the internetworking device you use, the cables you run, and the protocols you use play a very important role in your network.

Luckily for us, we have a fail-safe solution called flow control and windowing.

Flow control prevents the sending device from overflowing the buffers on the receiving side.

The protocols that are involved with reliable communications make sure the following happens:

• Segments are acknowledged back to the sender
• Segments that do not get acknowledged are retransmitted

Services that are considered to be connection oriented have the following characteristics:

• Three-way handshake
• Uses sequencing
• Uses ACKs
• Uses flow control

Windowing

Windowing is the process to check how much information the receiver can handle in one segment.

This window is adjustable, based on how much information is coming in.

Imagine two people are unloading a truck full of boxes. You've got the sender, the guy on the truck you have the receiver, the guy on the warehouse floor. So, the unloading begins, one box at a time. After a while, the receiver says, Hey! Send two boxes at a time. The window got bigger so the receiver sends two boxes at a time and, if something happens along the way, the receiver will let the sender know, Hey I did not get that box, send it again.

The same principle is applied when using reliable networking, ACKs, NACKs, sequencing, and windowing.

The Network layer

The Network layer is my favorite layer. This layer is where all the routing of the packets that take place in your segment or remote segments. The Network layer works with routed protocols, such as IPv4 or IPv6, routing protocols, such as RIPv2, RIPng, EIGRP, EIGRP for IPv6 and OSPF, and OSPFv3. I will be explaining these protocols in more depth later in the book.

The Network layer creates a routing table that stores all the routes it learns from the routing protocols or static routes that one enters manually. By default, all routers know who they are connected to. So, when a source decides to send a packet to a destination not within its own segment, it will need a layer three device, such as a router, to send the information to the proper destination.

If a router receives a packet with a network destination that is not in the routing table, the router will simply drop the packet and send you an error statement: Destination host unreachable or you could get Request Timeout. These two errors have different meanings, the first is that an entry for that network was never found, and the second is that the destination router has no entry or path to get back.

So, clearly, when we configure routers or any layer-three device, we must be very careful when inputting the IP addresses and subnet mask on their interface. When you configure any routing protocol, make sure you input the network addresses you are directly connected to. Routers will always choose the shortest path to a destination based on its metric; this will determine the path the packet will take to the destination.

Let's define some of the terms used:

• Routed protocols: These are the protocols that sit on an interface, such as IPv4 and IPv6. These protocols will have a subnetted scheme, so data can be routed by a routing protocol that chooses the appropriate network.
• Routing protocols: These are the components that create the routing table based on their algorithm, which will use the routed protocol's IP information to obtain the network address, and then route protocols to the correct destination.
• Metric: This is a measurement of how far the destination is from the source; depending on the routing protocol in use, it will use the shortest metric to get to the destination.

Let's continue to the next layer.

The Data Link layer

This layer provides the physical transmission of information, and handles flow control, physical network topology, and error notification. At the Data Link layer, each message is translated into a data frame and this frame will have customized information in it, such as the source and destination hardware address.

The Data Link layer does not perform any routing at layer two. It simply uses these physical addresses of the end devices to get from source to destination within the same segment.

Routers do not care about layer-two addressing, they are more concerned with layer-three addressing.

Be careful with that statement, because if you're using Ethernet technology, at this point layer-two addressing becomes very important to the router.

The Data Link layer is divided into two sublayers.

Media Access Control layer

In this sub layer, packets are placed on the media, depending on the technology used, such as Contention-Based or Token-Passing. As you know, physical addressing, that is, the MAC address or burned-in address of an NIC card, is used through the physical topology as well as the logical topology.

Logical Link Control (LLC)

Here, the responsibilities change to identifying network protocols and then passing them on to encapsulate them. The LLC header will always tell the Data-Link layer what to do with a packet once the frame is received.

The Physical layer

This layer is responsible for sending bits from the source to the destination on whatever media it is using.

Remember, even though in theory we say 0's and 1's, it is really electrical impulses that are generated and sent through the air as a Carrier Wave; or through cabling, that might need specific encoding and decoding, such as serial cables. In this layer, you'll find devices such as hubs, repeaters, amplifiers, cabling, even a modem at the client side, known as a channel service unit/data service unit.

As far as your certification is concerned, you only need to know IEEE basic information where the OSI is concerned.

One last thing I would like to leave you with before we move on to more exciting and adventurous topics is encapsulation and Protocol Data Units (PDUs). But a visual will be much better:

As packets flow down the OSI model, they will get encapsulated with the proper protocols, error corrections, and any other information they need to get to reach destination. Once they reach the destination, they will be de-encapsulated back into the original data format.

The process of encapsulation is a called data, segments, packets, frames, and bits or you could think of it as Don't, Stop, Pouring, Free, Beer.

These are called the PDUs, which communicates with their peer layer to make sure everything is in order.