OSI Structure and Terms

The OSI reference model breaks network communication into seven distinct layers. These layers are then divided into upper layers and lower layers, and one layer is further divided into two sub-layers. To further confuse things, the layers are often referred to by their name , number, or a combination of both.

All of this can be a bit confusing, but if you keep the following points in mind, eventually it will become clearer:

• The OSI model is conceptual; it does not describe how things actually work.

• The OSI model is a universally accepted standard throughout the industry.

• The OSI model is a model based on theory, and it is an easier way to understand the functions of data transfer.

Before we get too far into the model, let's clarify something that took us a long time to appreciate: The OSI model is a reference model. It provides a structure for organizing functions of data communications, but does not necessarily address how those functions are physically performed. A physical device like a "brouter" performs both bridging and routing , which are functions found at different layers of the OSI model. (These devices and functions will be addressed in Chapter 3,"Hardware and the OSI Model.") Furthermore, the OSI model is not the only model used to describe data communications. TCP / IP (Transmission Control Protocol/Internet Protocol) is the most popular communication protocol in use today and it follows the Department of Defense (DoD) four-layer model. (We will discuss TCP/IP in Chapter 5, "TCP/IP.") So if you are thinking of the OSI model as the blueprint for developing communications equipment, don't . The OSI model provides a structure and organization for the functions of data communications; it may or may not accurately describe the functions of a piece of equipment or protocol. Also, not all companies describe the OSI model in the same way. In fact, you can even find the model described differently within the same company. Therefore, although you may already be familiar with the OSI model, there will probably be some subtle differences between your understanding of the model and the way Cisco defines the model. There is no need to elaborate on which definition will be marked correct on the test, so please do not skip this chapter.

The description of the OSI model in this chapter is consistent with what you will find on the test. Pay particular attention to any areas that conflict with your understanding of the model. If you are sure your understanding is correct, remember the two rules for passing a Cisco certification exam:

Rule #1 ” Cisco's interpretation is correct.

Rule # 2 ” If Cisco's interpretation is not correct, see rule number one.

Layer One, the Physical Layer

The major function of this Physical layer is to place data on the network medium and ensure that the medium is capable of carrying that data. The type and gauge of wire, voltage levels, connectors, maximum distances, and other specifications for the medium are all found at this layer. The Physical layer is the lowest layer in the OSI model and as such is part of the OSI lower layers. (See Figure 2.1.)

Figure 2.1. The seven layers of the OSI model.

Cisco originally defined Layer 4 as a transitional layer between the upper and lower layers of the OSI model. Later it was defined as part of the lower layers, as described in many current texts . However, Cisco is moving away from making a distinction between upper and lower layers, and this may well be reflected in the current exam. Therefore, if you see a question on the exam asking where Layer 4 resides, the safest answer would be the lower layers. After all, the test you will be taking has been extensively revised, so it should not define Layer 4 as transitional, and if the test reflects Cisco's move away from making a distinction between upper layer and lower layer groupings you will not get a question like this at all.

Figure 2.1 shows the seven layers of the OSI model as you will see it on the test. The Application layer or Layer 7 (the layer names and numbers are used interchangeably) is always at the top with the other layers listed in descending order. This figure also breaks the OSI model into upper layers and lower layers .

Layer 2, the Data Link Layer

The Data Link layer or Layer 2 is responsible for low-level error-free communication between two network devices or nodes. The Data Link layer is a part of the lower layers and the only layer that is broken into sub-layers . Sub-layers are always referred to by their name or initials and do not have a number.

The Media Access Control ( MAC ) sub-layer is the lower of these two sub-layers, and is extremely important because it defines the addressing used by the nodes of all networks. Every device or node attached to a network is required to have a unique MAC address . Fortunately each device also requires a Network Interface Card ( NIC ) to physically attach to a network and this is where the MAC address resides. The address is actually burned into a chip on the NIC by the manufacturer, which brings up an interesting question: How do the numerous manufacturers of NICs know that the MAC address they are using is unique? The MAC sub-layer specifies a way to ensure uniqueness. MAC addresses are 48 bits long and usually written as three groups of hexadecimal digits, like this:

 081F.E453.5547 

The first 24 bits (6 digits) is unique vendor code given to the manufacturer. The last 24 bits (6 digits) is assigned by the manufacturer as a serial number. So long as everybody plays by the rules, any complete MAC address will be unique.

The Logical Link Control ( LLC ) sub-layer rests on top of the MAC sub-layer and provides the functionality required for connectionless and connection-oriented communication. LLC is the layer where protocol types are identified and where they can be encapsulated.

Connection-oriented communication is much like a telephone circuit, which is set up at the beginning of a conversation, maintained throughout the conversation, and then released at the end of the conversation. Connectionless communication is done on a frame-by-frame basis. Because each frame is autonomous, no link is established or maintained . Connectionless communication is faster than connection-oriented, but not quite as reliable. Each provides a way for upper-layer protocols to share transmission media, which is the main function of the LLC layer.

Layer 3, the Network Layer

Path determination and switching packets between networks are the primary functions performed at Layer 3, which is also considered a lower layer. However, before these functions can be performed, a structure of network addressing must be established. Layer 2 already provides MAC addresses for every node on a network, so why do we need another addressing function at Layer 3? The reason is that MAC addressing uses a flat structure and is limited as to the number of addresses available. If we tried to use MAC addresses for communication across all networks, we would quickly run out of unique addresses or the addresses themselves would have to be so large as to be unusable. A higher level of addressing structure is clearly needed and that is provided by a Layer 3 routable protocol . There are several routable protocols available, including Internet Protocol (the IP portion of TCP/IP), Novell's Internet Packet Exchange (IPX), and AppleTalk from Apple Computers. Each will provide a network addressing scheme and use packets that have a field for network addresses.

Do not confuse routable protocols with routing protocols , which also operate at Layer 3. Routable protocols such as IP, IPX, and AppleTalk provide a network addressing format and a packet structure that includes a field for the network address in addition to user data. Routing protocols are used by routers to exchange administrative information. Routing protocols do not include user data. Routing Information Protocol ( RIP ) and Open Shortest Path First ( OSPF ) are examples of some of the routing protocols we will discuss in Chapter 7, "Cisco Layer 3 Routing."

Layer 4, the Transport Layer

The Transport layer is part of the lower layers of the OSI model. Establishing end-to-end connection-oriented communications, dividing upper layer communications into autonomous segments, and ensuring reliable data flow are all services provided by the Transport layer.

Layer 5, the Session Layer

The Session layer establishes, manages , and terminates communications by coordinating service requests and responses between two or more stations . Sun's Network File System ( NFS ) and IBM's Structured Query Language ( SQL ) are representative of protocols used at the Session layer.

The upper layers ”including Session, Presentation, and Application ”provide standardization for applications to communicate. The functions of getting the data from one station to another are handled by the lower layers.

Layer 6, the Presentation Layer

The Presentation layer ensures that information delivered to the Application layer is readable and in the proper format. This includes the functions of data encryption/decryption, data compression/ decompression , and data representation. Examples of data formats used at the Presentation layer include ASCII, EBCDIC, MIDI, MPEG, JPEG, and GIF to name a few.

Layer 7, the Application Layer

The Application layer is at the top of the OSI stack and closest to user applications. User applications providing communication functionality are considered part of the Application layer. A word processing application by itself would not be part of the Application layer. However, if the word processor provided email capabilities, it would definitely be included in Layer 7.

Be prepared to explain why the OSI model was developed, recognize the layers by number and name, and lastly be able to list the major functions or services provided at each layer. Remember that you are expected to answer questions "the Cisco way," so use terms as presented here, which in all likelihood may be slightly different than the way you learned the model.