The OSI Model

The OSI model is the Open Systems Interconnection model which outlines a theoretical framework to understand how data is transported across the globe. OSI was created in 1984 by the International Organization for Standardization. Not all of the layers are used by every network. Specific protocols are used by networking hardware and software to move data based on the principals laid out by this model.

The seven layers of the OSI Model

1. Physical
2. Data
3. Network
4. Transport
5. Session
6. Presentation
7. Application

Layer 1 Physical Layer

The physical layer is the bottom most layer that is responsible for the raw data streams on the physical hardware. Patch blocks, cables, antennas, repeaters, and transceivers are all part of the physical layer. This layer defines the encoding used in the signals in the format of 0’s and 1’s.

• Maintains the data rate
• Performs synchronization
• Helps transmission medium decisions
• Helps in physical topology decisions (mesh, star, bus, ring)
• Helps make physical medium and interface decisions
• Data unit is in bits
• Provides modulation by converting data into radio waves
• Provides switching from sender to destination ports

Physical Topology

Mesh Topology
Every device has a direct path to communicate with the other devices on the network. These are more complex to setup because of the complex nature of having a point to point connection to each device. These networks are more secure because each path is dedicated between the devices.

Star Topology
Each device has a direct connection to a hub or central controller. These are easy to setup however there is no fault tolerance. These are easy to reconnect and install.

Bus Topology
Multiple devices are connected to a backbone cable. The backbone cable is a single cable that devices are connected to by tap and drop lines. These are difficult to reconnect reinstall.

Ring Topology
Each device is connected to repeaters. The repeaters form a circle hence the name ring. Each device must have a token to communicate on the network placed by the monitor.

Modes of Transmission

1. Simplex Mode
   Only one device can transmit and one can receive
2. Half Duplex Mode
   Both devices can transmit and receive but only one at a time
3. Full Duplex Mode
   Both devices can transmit and receive simultaneously

Physical Layer Protocols

1. 10 BASE-T twisted pair
   Two pair, cat 3 minimum, 100 meter max
2. 100 BASE-TX
   Fast ethernet, two pair, cat 5 minimum, 100 meter max
3. 1000 BASE-T
   10 gigabit ethernet over cat 5, 4 pair balanced twisted pair, depricated to cat 5e, 100 meter max
4. 10GBASE-T
   10 gigabit ethernet over copper, frequency of 500Mhz, cat 6, unshielded or shielded max 100 meters
5. 40GBASE-T
   40 gigabit per second ethernet, 4 pair balanced twisted pair, cat 8, up to 30 meters
   
   100 Megabit over Fiber
   
6. 100BASE FX
   Pair of multimode fiber, laser components, 400 meters half duplex, 2 kilometers full duplex
7. 100BASE SX
   Less expensive version of 100 megabit ethernet fiber, LED optics, 300 meters
   
   Gigabit Ethernet over Fiber
   
8. 1000BASE SX
   Gigabit ethernet using short wavelength laser, usually multimode, 220 meters or 550 meters depending on fiber type
9. 1000BASE LX
   Gigabit ethernet using long wavelength laser, multimode fiber up to 550 meters, single mode fiber up to 5 kilometers
   
   10 Gigabit Ethernet over Fiber
   
10. 10BASE SR
    Short range, multimode fiber, 26 to 400 meters depending on fiber type
11. 10GBASE LR
    Long range, single mode fiber, 10 kilometers range
    
    
12. Bluetooth
    Low power, low bandwidth, radio communication
    

The physical layer provides the encoding and signaling of the raw data streams through electrical, optical or radio signals sent on the physical hardware. The arrangement of these physical devices can be categorized into topologies. A well designed and implemented physical layer will ensure a reliable and efficient communication to the network.

Layer 2 Data Link Layer

The data layer is responsible for ensuring that the data being transmitted is being received correctly by the receiver. The key challenge of this layer is ensuring that a fast transmitter does not drown a slow receiver in bytes that it cannot keep up with. To mitigate these challenges the data layer uses a set of standards and techniques. One technique used is a traffic control method where the receiver updates the transmitter with how much buffer space it has. In broadcast mode this layer also needs to control who has access to the shared data using the medium access control MAC address.

• Connects two hosts point to point or broadcast
• Collects electrical signals from hardware
• Creates frames for the top layers
• Delivers frames for nodes on the same LAN or across the WAN
• One of the most complex layers
• Breaks up input data into dataframes
• The receiver sends acknowledgement frames

Data Link Protocols

Functions of the Data Link Layer

1. Framing and link access
   • Takes packets from the network layer and encapsulates them into frames
   • A frame is a collection of a data field where a network layer dataframe is put along with other data fields
   • Defines the frames structure and access protocol then sends bit by bit on the hardware
2. Addressing
   • All network interface adapters have a 6 byte identifier called a media access control address or MAC address
   • The data link layer uses this MAC address in the header to identify the sender and recipient of each packet
3. Protocol Identification
   • In the header there is code that states which network layer protocol created each packet
   • This enables the data link layer protocol that is receiving the packets to pass the traffic to the correct process
4. Error Detection
   • The footer of the frame has a cyclical redundancy check (CRC) calculation performed on the payload to make sure that the sender and recipient have the same data in their frame
   • If the data does not match the packet is thrown out
5. Media Access Control
   • If two or more devices on the LAN are using the same network medium the data link layer protocol uses media access control to stop them from transmitting simultaneously causing a data collision
6. Physical Layer Specifications
   • Data link layer provides specifications for the physical layer
     

Data Link Layer Protocols

These protocols are responsible to simply ensure and confirm that the bits and bytes that get received are the same as the ones getting sent.

• Synchronous Data Link Protocol
  SDLC was created by IBM in 1975 to connect mainframe computers to the client machines. Point to Point or Point to Multipoint connections are both supported. This protocol checks that the data units arrive correctly and with the right flow from one point on the network to the next.
• High-Level Data Link Protocol (HDLC)
  Created by ISO in 1979 this protocol is the umbrella which many WAN Wide Area Network protocols sit. Based on X.25. It can provide best effort unreliable service or reliable service with a bit oriented protocol that can be used for point to point or multipoint communications
• Serial Line Interface Protocol (SLIP)
  This is an older protocol that adds a framing byte at the end of an IP packet. This is used mainly for ISP’s to transfer packets to a home user over dial-up link. It works as an encapsulation of TCP/IP to work over serial ports and routers. There is no error correction or detection
• Point to Point Protocol (PPP)
  Has the same functionality as SLIP. This protocol is used to transport other packets along with IP packets. This can be used for dial up or leased router-router lines. A character oriented protocol used for error detection. Provides two protocols, LCP for bringing up lines, negotiating options, and taking down lines. NCP for negotiating network layer protocols.
• Link Control Protocol
  Created by the IEEE 802.2 to provide HDLC style services on LAN. Similar to a PPP protocol that can be used to establish, configure, test, maintain, and terminate links for transmissions of data frames
• Link Access Procedure (LAP)
  Data link protocol that frame and and transfer data across point-to-point links with some reliability features. There are 3 types. LAPB Link Access Procedure Balanced, LAPD Link Access Procedure D-Channel, and LAPF Link Access Procedure Frame-Mode Bearer Services. Originated from IBM SDLC, then submitted to the ISP for standardization
• Network Control Protocol
  NCP is an old protocol made by ARPANET. Allows users to have access to a computer and some devices at remote locations. Can also be used for file transfer among two or more computers. It generally sets up a PPP. NCP is available for every higher layer protocol that is supported by PPP. This was replaced by TCP/IP in the 1980’s.

Layer 3 The Network Layer

The physical and data link layers are only concerned with data moving on the internal networks and not the data that is moving to its final destination. That is the role of the network layer. Network layer devices are known as routers and connect individual networks together to form an internetwork.

Functions of the Network Layer

End to end addressing
The header contains the addresses of the sending and receiving system. The mac addresses in the data link layer change as the data moves but the network layer address remains the same from start to end.
Routing
Intermediate devices called routers forward the packets to other networks. Routing tables are used to look up the addresses from other networks to find the path for the packets final destination.
Fragmentation
Different networks have limitations on how large the packets it can receive is. If the sender packets are too large for the receiver they are split up into smaller data frames. This amount of data is known as the MTU maximum transmission units. The fragments are then reassembled when the final destination is reached.
Protocol Identification
The network layer provides code that specifies the protocol being used. Most likely the protocol is going to be IP. IP has two versions, IPv4 with a 32 bit address space, and more recently IPv6 with a 128 bit address space. IPv6 has not been adopted fully yet. All the functions of IP are in the headers which the protocol applies to the data from the transport layer. This creates the unit called a datagram.