The OSI model layer functions
    Peer-to-peer communications
    Five steps of data encapsulation
    LAN devices and technologies
    Ethernet and IEEE 802.3 standards
    Carrier sense multiple access with collision detection
    Logical (IP) addressing
    MAC addressing
    TCP/IP Addressing
    Subnetworks
    Application, presentation and session layers
    Transport layer functions

Layer 5 - Session: This layer establishes, maintains, and manages sessions between applications.

Layer 4 - Transport: This layer segments and reassembles data into a data stream. TCP is one of the transport layer protocols used with IP.

Layer 3 - Network: This layer determines the best way to move data from one place to another. Routers operate at this layer. You will find the IP (Internet Protocol) addressing scheme at this layer.

Layer 2 - Data Link: This layer prepares a datagram (or packet) for physical transmission across the medium. It handles error notification, network topology, and flow control. This layer uses Media Access Control (MAC) addresses.

Layer 1 - Physical: This layer provides the electrical, mechanical, procedural, and functional means for activating and maintaining the physical link between systems. This layer uses physical media such as twisted-pair, coaxial, and fiber-optic cable.

Each layer uses its own layer protocol to communicate with its peer layer in another system. Each layer's protocol exchanges information, called protocol data units (PDUs), with its peer layers. A layer can use a more specific name for its PDU. For example, in TCP/IP the transport layer of TCP communicates with the peer TCP function by using segments. Each layer uses the services of the layer below it in order to communicate with its peer layer. The lower layer service uses upper layer information as part of the PDUs that it exchanges with its peer.

The TCP segments become part of the network layer packets (datagrams) that are exchanged between IP peers. In turn, the IP packets become part of the data link frames that are exchanged between directly-connected devices. Ultimately, these frames become bits, as the data is finally transmitted by the hardware that is used by the physical layer protocol.

Each layer depends on the services of the OSI reference model layer that is below it. In order to provide this service, the lower layer uses encapsulation to put the protocol data unit (PDU) from the upper layer into its data field, then it can add whatever headers and trailers the layer wishes to use to perform its function.

As an example, the network layer provides a service to the transport layer, and the transport layer presents data to the internetwork subsystem. The network layer has the task of moving that data through the internetwork. It accomplishes this task by encapsulating the data within a packet.

This packet includes a header containing information that is necessary to complete the transfer, such as source and destination logical addresses.

The data link layer in turn provides a service to the network layer. It encapsulates the network layer packet in a frame. The frame header contains information that is necessary to complete the data link functions (e.g. physical addresses). And finally, the physical layer provides a service to the data link layer: It encodes the data link frame into a pattern of 1s and 0s for transmission through the medium (usually a wire).

As networks perform services for users, the flow and packaging of the user's original information go through several changes. In this example of internetworking, there are five conversion steps.

Step 1: A computer converts an e-mail message into alphanumeric characters that can be used by the internetworking system. This is the data.

Step 2: The message data is then segmented for transport on the internetwork system by the transport layer. The transport layer ensures that the message hosts at both ends of the e-mail system can reliably communicate.

Step 3: The data is then converted to a packet, or datagram, by the network layer. The packet also contains a network header that includes a source and destination logical address. The address helps network devices send the packet across the network along a chosen path.

Step 4: Each data-link layer device puts the packet into a frame. The frame enables the device to connect to the next directly-connected network device on the link.

Step 5: The frame is changed to a pattern of 1s and 0s for transmission on the medium (usually a wire). A clocking function enables the devices to distinguish bits as they travel across the medium.

The medium on the physical internetwork can vary along the path. For example, an e-mail message may originate on a LAN, cross a campus backbone, and continue through a WAN link until it reaches its destination on another remote LAN.

The network operates within a building or floor of a building.

LANs provide multiple connected desktop devices (usually PCs) with access to high-bandwidth media. By definition, the LAN connects computers and services to a common Layer 1 medium. LAN devices include:

Bridges that connect LAN segments and help filter traffic

Hubs that concentrate LAN connections and allow use of twisted-pair copper media

Ethernet switches that offer full-duplex, dedicated bandwidth to segments or desktop traffic

Routers that offer many services, including internetworking and broadcast control traffic

The following three LAN technologies (shown in the graphic) account for virtually all deployed LANs:

Ethernet -- The first of the major LAN technologies, it runs the largest number of LANs.

Token-Ring -- From IBM, it followed Ethernet and is now widely used in a large number of IBM networks.

FDDI -- Also uses tokens, and is now a popular campus LAN.

On a LAN, the physical layer provides access to the network media. The data link layer provides support for communication over several types of data links, such as Ethernet/IEEE 802.3 media. You will be studying the Ethernet IEEE 802.3 LAN standards. Figure shows the most common Layer 1 media used in networking today - coaxial, fiber-optic, and twisted-pair cable. Addressing schemes such as Media Access Control (MAC) and Internet Protocol (IP) provide a very structured method for finding and delivering data to computers or to other hosts on a network.

The Ethernet and IEEE 802.3 standards define a bus topology LAN that operates at a baseband signaling rate of 10 Mbps. Figure illustrates the three defined wiring standards:

10BASE2 (thin Ethernet) -- allows coaxial cable network segments up to 185 m. long

10BASE5 (thick Ethernet) -- allows coaxial cable network segments up to 500 m. long

10BASE-T -- carries Ethernet frames on inexpensive twisted-pair wiring

The 10BASE5 and 10BASE2 standards provide access for several stations to the same LAN segment. Stations are attached to the segment by a cable that runs from an attachment unit interface (AUI) in the station to a transceiver that is directly attached to the Ethernet coaxial cable.

Because 10BASE-T provides access for a single station only, stations that are attached to an Ethernet LAN by 10BASE-T are almost always connected to a hub or a LAN switch. In this arrangement, the hub or LAN switch is the same as an Ethernet segment.

The Ethernet and 802.3 data links prepare data for transport across the physical link that joins two devices. For example, as Figure shows, three devices can be directly attached to each other over the Ethernet LAN. The Macintosh on the left and the Intel-based PC in the middle show MAC addresses used by the data link layer. The router on the right also uses MAC addresses for each of the LAN side interfaces. The Ethernet/802.3 interface on the router uses the Cisco IOS interface type abbreviation "E" followed by an interface number.

Broadcasting is a powerful tool that can send a single frame to many stations at the same time. Broadcasting uses a data link destination address of all 1s (FFFF.FFFF.FFFF in hexadecimal). As Figure shows, if station A transmits a frame with a destination address of all 1s, stations B, C, and D will all receive and pass the frame to their upper layers for further processing.

When improperly used, broadcasting can seriously affect the performance of stations by unnecessarily interrupting them. Broadcasts should, therefore, be used only when the MAC address of the destination is unknown, or when the destination is all stations.

On an Ethernet LAN, only one transmission is allowed at any given time. An Ethernet LAN is referred to as a Carrier Sense Multiple Access with Collision Detection (CSMA/CD) network. This means that one node's transmission traverses the entire network and is received and examined by every node. When the signal reaches the end of a segment, terminators absorb it to prevent it from going back onto the segment.

When a station wishes to transmit a signal, it checks the network to determine whether another station is currently transmitting. If the network is not being used, the station proceeds with the transmission. While sending a signal, the station monitors the network to ensure that no other station is transmitting at that time. It is possible that two stations could both determine that the network is available and start transmitting at approximately the same time. If this should occur, they would cause a collision, as is illustrated in the upper part of the graphic.

When a transmitting node recognizes a collision, it transmits a jam signal that causes the collision to last long enough for all other nodes to recognize it. All transmitting nodes would then stop sending frames for a randomly selected period of time before attempting to retransmit. If subsequent attempts also result in collisions, the node would try to retransmit as many as fifteen times before finally giving up. The clocks indicate various backoff timers. If the two timers are sufficiently different, one station would succeed the next time.

An essential component of any network system is the process that enables information to locate specific computers systems on a network. Various addressing schemes are used for this purpose, depending on the protocol family being used. For example, AppleTalk addressing is different from TCP/IP addressing, which in turn is different from IPX addressing.

Two important types of addresses are data link layer addresses and network layer addresses. Data link layer addresses, also called physical hardware addresses or MAC addresses , are typically unique for each network connection. In fact, for most LANs, data link layer addresses are located on the NIC (network interface card). Because a typical computer system has one physical network connection, it has only a single data link layer address. Routers and other systems that are connected to multiple physical networks can have multiple data link layer addresses. As their name implies, data link layer addresses exist at Layer 2 of the OSI reference model.

Network layer addresses (also called logical addresses or IP addresses for the Internet Protocol suite) exist at Layer 3 of the OSI reference model. Unlike data link layer addresses, which usually exist within a flat address space, network layer addresses are usually hierarchical. In other words, they are like postal addresses that describe a person's location by indicating a country, state, ZIP Code, city, street, house address, and name. One example of a flat address is a U.S. Social Security number. Each person has a unique Social Security number, people can move around the country and obtain new logical addresses depending on their city, street, or ZIP Code, but their Social Security numbers remain unchanged.

In order for multiple stations to share the same media and still identify each other, the MAC sublayers define hardware or data link addresses called the MAC addresses. Each LAN interface has a unique MAC address. In most NICs, the MAC address is burned into ROM. When the NIC initializes, this address is copied into RAM.

Before directly connected devices on the same LAN can exchange a data frame, the sending device must have the destination device's MAC address. One way in which the sender can ascertain the MAC address that it needs is to use an ARP (Address Resolution Protocol). The graphic illustrates two ways in which a TCP/IP example, ARP, is used to discover a MAC address.

In the first example, Host Y and Host Z are on the same LAN. Host Y broadcasts an ARP request to the LAN looking for Host Z. Because Host Y has sent out a broadcast, all devices including Host Z will look at the request; however, only Host Z will respond with its MAC address. Host Y receives Host Z's reply and saves the MAC address in local memory, often called an ARP cache. The next time Host Y needs to directly communicate with Host Z, it uses the stored MAC address.

In the second example, Host Y and Host Z are on different LANs, but can access each other through Router A. When Host Y broadcasts its ARP request, Router A determines that Host Z cannot recognize the request because Router A detects that the IP address for Host Z is for a different LAN. Because Router A also determines that any packets for Host Z must be relayed, Router A provides its own MAC address as a proxy reply to the ARP request. Host Y receives Router A's response and saves the MAC address in its ARP cache memory. The next time Host Y needs to communicate with Host Z, it uses the stored MAC address of Router A.

In a TCP/IP environment, end stations communicate with servers or other end stations. This can occur because each node using the TCP/IP protocol suite has a unique 32 bit logical address. This address is known as the IP address. Each company or organization connected to an internetwork is perceived as a single unique network that must be reached before an individual host within that company can be contacted. Each company network has an address; the hosts that live on that network share that same network address, but each host is identified by the unique host address on the network.

Subnetworks

Subnets improve the efficiency of network addressing. Adding subnets does not change how the outside world sees the network, but within the organization, there is additional structure. In Figure , the network 172.16.0.0 is subdivided into four subnets: 172.16.1.0, 172.16.2.0, 172.16.3.0, and 172.16.4.0. Routers determine the destination network by using the subnet address, which limits the amount of traffic on the other network segments.

From an addressing standpoint, subnets are an extension of a network number. Network administrators determine the size of subnets based on the expansion needs of their organizations. Network devices use subnet masks to identify which part of the address is for the network and which part represents host addressing.

Example of Class C subnetting.

In Figure , the network has been assigned the Class C address 201.222.5.0. Assuming that 20 subnets are needed, with a maximum of 5 hosts per subnet, you need to subdivide the last octet into a subnet and a host, and then determine what the subnet mask will be. You need to select a subnet field size that yields enough subnetworks. In this example, selecting 5-bits gives you 20 subnets.

In the example, the subnet addresses are all multiples of 8 - 201.222.5.16; 201.222.5.32; and 201.222.5.48. The remaining bits in the last octet are reserved for the host field. The 3 bits in the example are enough for the required five hosts per subnet (actually, giving you host numbers 1 - 6). The final host addresses are a combination of the network/subnet segment's starting address plus each host's value. The hosts on the 201.222.5.16 subnet would be addressed as 201.222.5.17, 201.222.5.18, 201.222.5.19, and so forth.

A host number of 0 is reserved for the wire (or subnet) address, and a host value of all 1s is reserved because it selects all hosts-that is, it is a broadcast. A table used for the subnet planning example is on the following page. Also, a routing sample shows the combining of an arriving IP address with a subnet mask to derive the subnet address (also called the subnet number). The extracted subnet address should be typical of the subnets generated during this planning exercise.

Application Layer

In the context of the OSI reference model, the application layer (Layer 7) supports the communicating component of an application. It does not provide services to any other OSI layer. However, it does provide services to application processes lying outside the scope of the OSI model (e.g. spreadsheet programs, Telnet, WWW, etc.) A computer application can function completely by using only the information that resides on its computer. However, an application might also have a communicating component that can connect with one or more network applications.

An example of such an application might include a word processor that can incorporate a file transfer component that allows a document to be transferred electronically across a network. The file transfer component qualifies the word processor as an application in the OSI context, and therefore, belongs in Layer 7 of the OSI reference model. Another example of computer application that has data transfer components is a Web browser such as Netscape Navigator and Internet Explorer. Whenever you visit a Web site, the pages are transferred to your computer.

Presentation Layer

The presentation layer (Layer 6) of the OSI reference model is responsible for presenting data in a form that a receiving device can understand. It serves as the translator - sometimes between different formats - for devices that need to communicate over a network, by providing code formatting and conversion. The presentation layer (Layer 6) formats and converts network application data into text, graphics, video, audio, or whatever format is necessary for the receiving device to understand it.

The presentation layer is not only concerned with the format and representation of data, but also with the data structure that the programs use. Layer 6 organizes the data for Layer 7.

To understand how this works, imagine that you have two systems. One system uses EBCDIC, and the other uses ASCII to represent data. When the two systems need to communicate, Layer 6 converts and translates the two different formats.

Another function of Layer 6 is the encryption of data. Encryption is used when there is a need to protect transmitted information from unauthorized receivers. To accomplish this task, processes and codes located in Layer 6 must convert the data. Other routines located in the presentation layer compress text and convert graphic images into bit streams so that they can be transmitted across a network.

Layer 6 standards also guide how graphic images are presented. Following are some examples:

PICT -- a picture format used to transfer QuickDraw graphics between Macintosh or PowerPC programs

TIFF -- tagged image file format, used for high-resolution, bit-mapped images

JPEG -- from the Joint Photographic Experts Group, used for photographic quality images

Other Layer 6 standards guide the presentation of sound and movies.

Included in these standards are the following:

MIDI -- musical instrument digital interface for digitized music

MPEG -- the motion picture experts group's standard for the compression and coding of motion video for CDs, digital storage, and bit rates up to 1.5 Mbps

QuickTime -- a standard that handles audio and video for Macintosh and PowerPC programs

Session Layer

The session layer (Layer 5) establishes, manages, and terminates sessions between applications. It coordinates the service requests and responses that occur when applications establish communications between different hosts.

Transport Layer

The transport layer (Layer 4) is responsible for transporting and regulating the flow of information from source to destination reliably and accurately. Its functions include:

    connection synchronization
    flow control
    error recovery
    reliability through windowing

The transport layer (Layer 4) enables a user's device to segment several upper-layer applications for placement on the same Layer 4 data stream, and enables a receiving device to reassemble the upper-layer application segments. The Layer 4 data stream is a logical connection between the endpoints of a network, and provides transport services from a host to a destination. This service is sometimes referred to as end-to-end service.

As the transport layer sends its data segments, it also ensures the integrity of the data. This transport is a connection-oriented relationship between communicating end systems. Some of the reasons for accomplishing reliable transport are as follows:

(1) It ensures that senders receive acknowledgement of delivered segments. (2) It provides for retransmission of any segments that are not acknowledged. (3) It puts segments back into their correct sequence at the destination device. (4) It provides congestion avoidance and control.

One of the problems that can occur during data transport is overflowing buffers on receiving devices. Overflows can present serious problems that result in data loss. The transport layer uses a method called flow control to solve this problem.

Each of the upper-level layers performs its own functions. However, their functions depend on lower-layer services. All four upper layers - application (Layer 7), presentation (Layer 6), session (Layer 5), and transport (Layer 4) - can encapsulate data in end-to-end segments.

The transport layer assumes that it can use the network as a cloud to send data packets from source to destination. If you examine the operations that take place inside the cloud, you can see that one of the functions involves selecting the best paths for a given route. You begin to see the role that routers perform in this process.

Segmentation of upper-layer applications

One reason for using a multi-layer model such as the OSI reference model is that multiple applications can share the same transport connection. Transport functionality is accomplished segment by segment. This means that different data segments from different applications, being sent to the same destination or to many destinations, are sent on a first-come, first-served basis.

To understand how this works, imagine that you are sending an e-mail and transferring a file (FTP) to another device on a network. When you send your e-mail message, before the actual transmission begins, software in your device sets the SMTP (e-mail) port number and the originating program port number. As each application sends a data stream segment, it uses the previously defined port number. When the destination device receives the data stream, it separates and sorts the segments so that the transport layer can pass the data up to the correct corresponding destination application.

TCP establishes a connection

In order for data transfer to begin, one user of the transport layer must establish a connection-oriented session with its peer system. Then, both the sending and receiving application programs must inform their respective operating systems that a connection will be initiated. In concept, one device places a call to another device that the other device must accept. Protocol software modules in the two operating systems communicate by sending messages across the network to verify that the transfer is authorized and that both sides are ready. After all synchronization has occurred, a connection is established, and data transfer begins. During transfer, the two devices continue to communicate with their protocol software to verify that they are receiving the data correctly.

The graphic depicts a typical connection between sending and receiving systems. The first handshake requests synchronization. The second and third handshakes acknowledge the initial synchronization request, and synchronize the connection parameters in the opposite direction. The final handshake segment sends an acknowledgement to the destination that both sides agree that a connection has been established. As soon as the connection has been established, data transfer begins.

TCP sends data with flow control

While data transfer is in progress, congestion can occur for two different reasons. First, a high-speed computer might generate traffic faster than a network can transfer it. Second, if many computers send datagrams simultaneously to a single destination, that destination can experience congestion. When datagrams arrive too quickly for a host or gateway to process, they are temporarily stored in memory. If the traffic continues, the host or gateway eventually exhausts its memory and discards any additional datagrams that arrive.

Instead of allowing data to be lost, the transport function can issue a "not ready" indicator to the sender. This indicator acts like a stop sign and signals the sender to stop sending data. When the receiver is able to accept additional data, it sends a "ready" transport indicator, which is like a go signal. When the sending device receives this indicator, it resumes segment transmission.

TCP achieves reliability with windowing

Reliable connection-oriented data transfer means that data packets arrive in the same order in which they are sent. Protocols fail if any data packets are lost, damaged, duplicated, or received in the wrong order. In order to ensure transfer reliability, receiving devices must acknowledge receipt of each and every data segment.

If a sending device must wait for acknowledgement after sending each segment, it is easy to see that throughput could be quite low. However, because there is a period of unused time available after each data packet transmission and before processing any received acknowledgment, the interval can be used for transmitting more data. The number of data packets a sender is allowed to transmit without having received an acknowledgment is known as a window.

Windowing is an agreement between sender and receiver. It is a method of controlling the amount of information that can be transferred end-to-end. Some protocols measure information in terms of the number of packets; TCP/IP measures information in terms of the number of bytes. The examples in the Figure show the workstations of a sender and a receiver. One has a window size of 1, and the other a window size of 3. With a window size of 1, a sender must wait for an acknowledgment for every data packet transmitted. With a window size of 3, a sender can transmit three data packets before expecting an acknowledgment.

TCP acknowledgment technique

Reliable delivery guarantees that a stream of data that is sent from one device will be delivered through a data link to another device without duplication or data loss. Positive acknowledgment with retransmission is one process that guarantees reliable delivery of data streams. It requires a recipient to send an acknowledgment message to the sender whenever it receives data. The sender keeps a record of each data packet that it sends and then waits for the acknowledgment before sending the next data packet. The sender also starts a timer whenever it sends a segment, and retransmits the segment if the timer expires before the acknowledgment arrives.

Example: A sender transmits Data Packets 1, 2, and 3. The receiver acknowledges receipt of the packets by requesting Packet 4. The sender, upon receiving the acknowledgment, sends Packets 4, 5, and 6. If Packet 5 does not arrive at the destination, the receiver acknowledges with a request to re-send Packet 5. The sender re-sends Packet 5 and waits for acknowledgment before transmitting Packet 7.