Networking

Talking Shop: Don't fear the OSI model

The OSI is used as a standard to ensure different systems can communicate effectively


Don’t be intimidated by the Open Systems Interconnection (OSI) model. Ratified by the International Standards Organization (ISO) in 1978, the OSI model isn’t some gargantuan Achilles heel or certification conspiracy built for the single purpose of bringing strong-willed IT professionals to their knees.

Conquer your fears. Take a few moments to understand the principles at work behind the reference model, which is used as a standard to ensure different systems can communicate effectively.

Why? Because you're liable to find yourself lost if you hang around an IT department long enough but can't speak the OSI vocabulary. In addition, if you’re preparing for a Microsoft, Cisco, or Network Plus certification exam, the OSI model is sure to rear its ugly head. Here’s a quick primer aimed at helping you understand the OSI model better.

Paperchase Digest has a new name!
Paperchase Digest is now Erik Eckel’s IT Certification Corner. As always, you’ll find certification tips and news in your e-mail box each Friday. Catch every column, along with timely advice and reviews not found on the site! It’s easy and it’s free. Just go to the TechMails page and sign up for Erik Eckel’s IT Certification Corner to ensure that you keep up-to-date on the latest certification tips, shortcuts, news, and more!

Standardization
The OSI model’s purpose is to standardize network technologies. By ensuring that their hardware and software products meet the OSI model’s design specifications, manufacturers can have confidence that their various programs and devices will communicate well, regardless of the other manufacturer’s programs or components being used to perform the communication.

Transferring data between systems involves many steps. The data is passed from an application on one system through a physical connection, onto a network, and then back through a physical connection, where another system’s application processes the data. For example, the process occurs for every e-mail message you receive.

The OSI network reference model dictates that the process involves seven layers of functionality on each system. Figure A displays the seven layers, from top to bottom.

Figure A
The OSI reference model consists of seven layers.


Imagine an e-mail message is being sent between two systems. This message is created using an e-mail client, and when it’s sent, Application layer actions execute. The e-mail message passes down the layers until it arrives at the system network adapter or modem and is shot out onto a wire (or through a wireless network). The step in which a message reaches a physical connection, and the manner in which it is transmitted, maps to the OSI model’s Physical layer.

When the e-mail message is received, it enters at the opposite end of the OSI model. Instead of speeding directly to the other system’s Application layer, it must first pass through the Physical and other layers while on the way to the Application layer, where ultimately the e-mail message will be passed off to a software program such as Microsoft Outlook. Figure B illustrates the process.

Figure B
Data packets traveling between systems pass through seven model layers on each system.


Here’s another way to visualize it. The e-mail program generates data packets that pick up an additional layer of information each time an OSI layer is passed. When the data packet is created in the Application layer, data specific to that application is inserted in the data packet. When the data packet hits the Presentation layer, redirection information is added, and the data packet may even be restructured. The process continues for the data packet as it passes down through each layer of the OSI model.

But what happens when the data packet travels across a wire to another system? The second system’s Physical layer receives the packet. The second machine’s Data-link layer strips off and interprets the Data-link layer information the first system added, and so on, until the Application layer receives the packet with the application data.

While the process is as complicated as calculating college Bowl Championship Series standings, it’s actually a quite elegant dance performed between two or more systems. It can be better understood by examining the actions that occur at each layer.

As we review the role the layers play, remember that each OSI level communicates only to the layer immediately above or below it. The exception is the Physical layer, which talks to the same level on other systems. Let’s review the Application layer first.

OSI model—Application layer
The Application layer lives at the top of the OSI model. It’s the king. It talks to a system’s software applications; hence the name. Applications owe this layer a debt of gratitude, because it provides them with access to network services.

Several protocols operate at the Application layer, including Simple Network Management Protocol (SNMP), File Transfer Protocol (FTP), X.500, AppleTalk, and Simple Mail Transfer Protocol (SMTP). Server and workstation services also operate at this layer, as do file system drivers.

Residing below the Application layer is the Presentation layer.

OSI model—Presentation layer
The Presentation layer determines the format used to share data between systems. Redirectors operate at this level. On Microsoft systems, Server Message Blocks (SMBs) are used. On Novell systems, the Novell Core Protocol (NCP) is used.

If it weren’t for the Presentation layer, which determines the data format, IBM-compatible systems wouldn’t be able to easily share data with Linux or UNIX systems. Because the Presentation layer takes application-specific data from the Application layer and translates it into a standardized format, that information can be shared between disparate systems.

The Presentation layer assumes responsibility for completing several actions. For example, protocols are converted, data is translated, and encryption tasks are performed, among other procedures. SNMP, FTP, and SMTP all also operate at this level, as do file system drivers.

The next layer that data hits on its way from an application to a network is the Session layer.

OSI model—Session layer
In order to share data, two systems must create a communications link. Called a session, this connection must perform identity and security duties. At the Session layer, systems verify that they are talking to a specific machine.

The Session layer establishes, permits the use of, and ends these communications sessions. It works by synchronizing communications, adding checkpoints to the data streams being sent, and determining data transfer rates. AppleTalk ASP operates at this level.

Moving down, we come to the Transport layer.

OSI model—Transport layer
The Transport layer has one of the toughest jobs of any layer. Essentially, its role is the equivalent of being the OSI model’s FedEx courier. The Transport layer is charged with ensuring the proper delivery of data packets, as well as packing and unpacking the messages.

It works by breaking long messages into shorter messages and combining very small messages into larger ones. The goal is error-free, efficient transmission of data packets between systems. Essentially, the Transport layer assembles and disassembles data packets and sends acknowledgements to confirm data was properly received. The sequence in which packets are sent is determined at the Transport layer.

Transmission Control Protocol (TCP) is the most widely used Transport layer protocol. Other protocols that operate at the Transport layer include NetBIOS Extended User Interface (NetBEUI), Sequenced Packet Exchange (SPX), and NWLink.

Beneath the Transport layer is the Network layer.

OSI model—Network layer
The Network layer handles the task of stamping messages with address information. It also translates logical addresses and names into physical addresses.

The data path that data packets will travel is determined at the Network layer. The Network layer also assumes the challenge of managing data traffic, including routing functions. Should a network adapter be unable to send a large data chunk, the large chunk is broken into smaller pieces at this layer.

A host of items operate at this layer, including routers, network interface card (NIC) drivers, Internet Protocol (IP), Internet Packet Exchange (IPX), NWLink, and NetBEUI.

Beneath the Network layer resides the Data-link layer.

OSI model—Data-link layer
The Data-link layer is charged with ensuring the error-free transfer of data frames from one system to another via the Physical layer. At the Data-link layer, data bits received from the Physical layer are packaged into data frames and sent to the Network layer. Or, when data frames are received from the Network layer, the Data-link layer passes them to the Physical layer and the network connection.

For each data frame the Data-link layer sends, it seeks an acknowledgement. If it doesn’t receive one, it sends the data frame again. To ensure error-free data receipt, cyclical redundancy checks occur at this level.

Two specific Data-link layers exist: the Logical Link Control (LLC) and Media Access Control (MAC) layers. The LLC manages data-link interface points, which are called service access points (SAPs). It resides above the MAC layer, which regulates the electronic voltage on the NIC and shares access to network adapter cards. The MAC layer is tied tightly to a system’s NIC, with which it communicates directly.

Often you will hear IT professionals speak of an NIC’s MAC address. They are referring to the unique identification address given to every network card that is ever produced. Even identical NIC models boast separate and unique MAC addresses.

Drivers, Network Device Interface Specifications (NDIS), which communicate between transport protocols and NIC drivers, and bridges operate at the Data-link layer of the OSI model.

This brings us to the Physical layer.

OSI model—Physical layer
The Physical layer was named after the physical wires that connect system NICs. The Physical layer determines the type of cable (electrical, optical, mechanical, etc.) in use.

At this layer, raw, binary data is transmitted from one system to another. The raw bit stream transmitted at the Physical layer is nothing but ones and zeroes. This is the only “hardware” layer in the OSI model, with all of the other layers using actual software to complete their functions.

Repeaters operate at the Physical layer. Encoding of data and bit synchronization tasks also occur at this layer.

Conclusion
The OSI model plays an important role in standardizing communications between disparate systems. Were it not for the standards and specifications outlined in the OSI model, linking Apple, IBM, UNIX, and other systems would certainly prove much more challenging.

Editor's Picks