In the early 1980s, the International Organization for Standardization (ISO) was busy creating
a new data communications model. Its intention was to build the next generation of communications protocols to replace both proprietary protocols and the de facto TCP/IP protocol. In the end, the cheaper TCP/IP protocol won the battle.

Many customers had no interest in paying for the cost of developing the Open Systems
Interconnect (OSI) protocol. The OSI model is still used as a network training tool. The model
stratifies data communication into seven distinct layers. Each layer provides a unique function
in support of the layer above.

We are going to walk through the OSI model layer by layer and compare each function to
the TCP/IP protocol model. The seven layers of the OSI model are as follows:
- Physical layer
- Data-Link layer
- Network layer
- Transport layer
- Session layer
- Presentation layer
- Application layer
Let’s start with a simple memory trick to remember each of the OSI layers in proper order

Each of the first letters relates to the first letter of an OSI layer, working your way up
from the bottom. It is in your best interest to learn how to draw the OSI model and these
layers from memory. You will find it helpful on your CISA exam. You will also find it helpful
during discussions when you’re trying to uncover the details about a particular product.
It will impress clients.

The second half of the mnemonic (see Figure 4.10) is “Nor Do I Throw Apples” (NDITA).
Once again, each letter refers to a layer of the Transmission Control Protocol/Internet Protocol
(TCP/IP) model, working up from the bottom.

Let’s review the basic OSI process for handling data. The top layer is where your application
is running. The lower layers process the request and prepare the data for transmission as
it works its way down to the bottom. When it reaches the bottom of the OSI model, the data
has been broken down into electrical signals.

These electrical signals will be received by the other computer. Upon receipt, the transmission
headers are stripped off. The remaining data message is passed to the application software
running on the other computer.

Now that you understand in general how data is transmitted when using the OSI and
TCP/IP models, let’s go inside the individual layers, one by one.

Layer 1: Physical Layer
The Physical layer defines physical requirements in the cables and voltages. This layer specifies functional specifications for creating, maintaining, and deactivating an electrical link between systems.

Layer 2: Data-Link Layer
The Data-Link layer focuses on establishing data communications via hardware device drivers and their transmit/receive function. Communication in layer 2 is established between each network card’s Media Access Control (MAC) address. A MAC address is a burned-in serial number that is unique to every network card ever manufactured. The address is unique because it uses the manufacturer’s ID and the board serial number. Each computer uses the MAC address for “to” and “from” communications within the same broadcast domain (layer 2).

Broadcast domains are no more than a noisy shouting match between computers on the
same subnet. Every computer in that segment will hear every conversation from all the computers.
For example, Microsoft NetBIOS is a layer 2 protocol. Dynamic Host Configuration
Protocol (DHCP) is also a layer 2 protocol.If too many computers were talking at the same time, we would have a congestion problem. This is referred to in Ethernet as a collision. Too many collisions will render the data link unusable.

Layer 3: Network Layer

The Network layer (see Figure 4.14) defines networking. Computers are stupid. The computer
simply follows the directions of the person who programmed its settings or loaded the detailed
list of instructions (a program). Your network administrator uses a numeric grouping of
addresses to identify systems within the network.

Networks can be administratively divided into logical groups or segments. We refer to this
grouping as IP subnetworks (subnets). Each system has its own individual network address
that is unique on the network. Your computer acquires an IP address from either a static configuration or a dynamic configuration by using DHCP/BOOTP. The computer ties the IP address to its MAC address.

Routing decisions are based on the IP address. The computers and routers implement the Address Resolution Protocol (ARP). The purpose of ARP is to match the IP address with the correct MAC address. The system may need to match the MAC to the IP by using Reverse ARP, known as RARP.

If you recall, the problem in layer 2 is that all the systems are transmitting so much that it
is creating noise across the subnet. The issue is similar to noise in a school cafeteria. Some conversations are broadcast with everyone listening, while a few are discreet between a couple of
users. With a layer 3 network address, it is possible to reduce traffic noise by unicast transmission
to an individual address. Unicast is a method for point-to-point communication. So what if you need to send a message to more than one address, but not everyone? You could use multicast to deliver communication to a group of addresses. Multicast is similar to a conference call and is the basis of virtual networking.

Network routing is the process of directing traffic to the intended destination. Static routing uses specific TO-FROM mappings of IP addresses created by the network administrator. The mapping is manually typed into the router and stored in the routing table of each router. The settings will not change unless the network administrator manually changes them. Static routes are good for security and are used when the network traffic is both predictable and relatively simple.

Dynamic routing uses a protocol algorithm to automatically adjust the path to the intended
destination. This method uses a special router information protocols like RIP and OSPF to signal
available paths (routes), dead routes (unroutable), and other changes. The routers will monitor
routing updates and signal other routers to reconfigure their routing tables as changes occur.

Dynamic routing is easy to enable. It removes the complexities of building an advanced configuration. To some individuals, this seems like the best answer; however, dynamic routing can be both beneficial and dangerous. Dynamic routing changes can be initiated by the following:
- Router equipment failure
- Addition of new networked devices
- Incorrect configuration of a network-attached device, including a common workstation

By default, all network devices will listen to route updates. This can create a nightmare if
left to default settings. Poor administration of computers and routers can cause traffic to be
misdirected into a dead-end route or bandwidth bottleneck. Proper design of the network usually includes implementation of both static and dynamic routing. Static routes can provide a designated router of last resort if prior dynamic routes fail.

For higher security, the routers should be configured to accept updates from only a trusted
router. The trusted router update is similar to an access control list (ACL) and will accept
updates only from routers the network administratorknows and can trust.

Layer 4: Transport Layer
The Transport layer specifies the transport delivery method. There are two basic methods: confirmed and unconfirmed delivery. Confirmed delivery uses a TCP connection to the destination. This is similar to requesting a return receipt and sending certified mail from the post office. Unconfirmed delivery operates on a UDP connectionless datagram, which is normally broadcast across the network like a shout in a dark room. Even if transmissions are unicast between two stations, the higher level software application would have to confirm delivery, because UDP does not offer delivery confirmation. There is no guarantee of its receipt on the other end. UDP transmissions have less overhead. It is the responsibility of the recipient program to detect errors.

Layer 5: Session Layer
The Session layer governs session control between applications. This is where you initiate communications to a system and establish, maintain, and terminate a communication session. Examples include Network File System (NFS), SQL*net database sessions, and Remote Procedure Call (RPC).

Layer 5 functions of session and error control are handled in TCP/IP by the user’s application
software. Under TCP/IP, it is the responsibility of the user application to manage the functions
of session, presentation, and application.

Layer 6: Presentation Layer
The Presentation layer defines the presentation format. This is where you specify the format and data structure to be used for programs. Layer 6 will specify the differences between a PDA, VT100 terminal, or a word processor with What-You-See-Is-What- You-Get (WYSISYG) display capabilities.

Layer 6 converts data received from the Session layer into a format that can be handled by
the upper-level Application layer (layer 7). It also works in the opposite direction, receiving
application data from layer 7 and reformatting it for the underlying layer 5. For TCP/IP, the
presentation function is combined into the TCP/IP Application layer.

Layer 7: Application Layer
The Application layer is where the problem-solving calculations of the computer software program run. Various types of computer application software execute in the Application layer, including the following:
- SNA gateways, which convert the ASCII 7-bit data structure into IBM Extended Binary Coded Decimal Interchange Code (EBCDIC) 8-bit data structure for the mainframe
- Domain Name Server (DNS), which is the program that associates a domain name to the
matching IP address (for layer 3)
- File, print, and web servers
- Databases and office automation software (such as OpenOffice and Microsoft Office)

Understanding How Computers Communicate
Now that we’ve covered all seven layers of the OSI model, let’s take a finished look at the communication between two computers across the network using the OSI model. We will assume that a router is being used in the communication path.

First, the user makes their request in their application software on layer 7. That request is
passed down through each layer on its way to the bottom. Along the way, each layer performs
its function to ultimately transform the request into a series of electrical signals for transmission
on layer 1 (the physical layer of cable and voltages).

Next, the network hub (or network switch) on layer 2 passes the signal up to the layer 3 network
router. The router executes a static or dynamic route directing the user’s request to the
intended destination computer.

Then the request is received as a series of electrical signals on layer 1 of the other computer.
The request is passed through each layer of the OSI model and processed accordingly. The
request is then received by the other computer in its Application layer, where it is executed.
The response is packaged and sent back through the OSI model in reverse, until it reaches the
computer display screen.

Congratulations, you have now learned the secrets of the OSI model. You’ll need to be sure
to review its relationship with the TCP/IP model. The OSI model will be used as a discussion
tool, while the TCP/IP model is the de facto standard of the real world

Popularity: 31% [?]