 |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |  | |
| By N2H | ||||||||||||||||||||||
Data link layer
January 17, 2008
If you're new here, you may want to subscribe to my RSS feed. Thanks for visiting and have a nice day!
Standards & Protocols
Industry of Electronic and Electrical Engineers (IEEE) split the function of the data link layer into 2 sublayer. They are:
- Media Access Control (MAC)
- Logical Link Control (LLC)
Media Access Control (MAC)
This sublayer defines the way in which multiple NICs share a single transmission medium. It covers:
- Logical topology e.g. bus, ring e.t.c.
- Media Access Method of contention, token passing e.t.c.
- Addressing i.e. the hardware address of the NIC also known as MAC address.
Media Access
This is the methodology used for determining when devices are allowed to communicate using the network. A network has to share the available communications capacity between the various devices that use it. It has to deal with problems such as two devices wanting to communicate simultaneously. This can be done through:
1) Contention
Using this method each network device competes with other connected devices for use of the transmission media. This method brings about collisions, when two devices transmit simultaneously, their signals collide and none is able to transmit.
To reduce collisions, protocols are used to ensure that devices listen to the media before transmitting and only transmit if media is clear. These protocols are called carrier sense multiple access protocols (CSMA). There are two types of CSMA protocols.
Carrier Sense Multiple Access with collision detection (CSMA/CD)
CSMA/CD protocols are capable of detecting a collision and retransmitting the data. They recognize a collision by the excessively high voltage or current when two signals combine. After collision both devices wait for a random period of time before retransmitting e.g. DEC Ethernet Version II and IEEE 802.3
Carrier Sense Multiple Access with collision Avoidance (CSMA/CA)
This protocol uses schemes such as time slices accessing or requests to send data to gain access to the media. e.g. Apples’ Local Talk. Machines wait for the network to be free of traffic before announcing their intention to transmit. If the machine does not receive a negative response, it proceeds to transmit.
Advantages of Contention based:
- Simple low-overhead software.
- Immediate and full control of the media.
Disadvantage of Contention
- Access times are unpredictable
- No prioritizing of devices.
- Collisions increase geometrically as each new device is connected.
Note:
Contention based media access systems provide network access on a first come first serve basis, thus network cards using fast expansion buses such as PCI can place data on the network faster than older ISA type cards and therefore these fast cards can better utilize the network.
The fastest, widest network card available should be put in the servers as these devices are involved in many of the network dialogs between machines.
Some network operating system support multiple network cards per system (multi-homed systems) thus allows the system to access the media faster. This is also known as network load balancing.
2) Token Passing
This method uses a small data frame called token, which is passed from device to device. The first device to join the network creates the token. Any device holding the token has control over the transmission media, and can remove the token and place a data frame on to the media.
The frame passes from device to device until it reaches the destination.The device copies the data and flags the data as read. The token and data continue to pass from device to device until arriving back at the source device.The data is removed and the token is placed back on to the media, which moves onto the next device.
Passing the token around evenly distributes network access among the connected devices. Protocols limit the length of time the token remains with a single device.
Note: Protocols using this system of access control are:
- IEEE 802.5 token ring.
- Future distributed data interface (FODI)
Advantages of Token Passing
- Access times are predictable.
- Data can be prioritized (useful for time-sensitive traffic such as voice or video)
- Collisions can’t occur which allows a higher through put of data under high load conditions.
- One network device is often used as a controller for fault detection and recovery.
Disadvantages of Token Passing
- The software required is more complete and has a higher overhead.
- Some software reconfiguration is required when devices are added or removed.
3) Demand PriorityThis method provides a contention based access method without the risk of collisions. The system requires demand priority switches. These are intelligent switches that control network access and allow only one node to transmit information at a time.
Data can be prioritized depending on the type of data e.g. time sensitive data such as voice and video can be given a high priority which means the switches process it before low priority data. Protocol using these method are IEEE 802.12 known as 100 VG-Any LAN.
Advantage of Demand Priority
- Data can be prioritized – high or low.
- There are no collisions giving a higher throughput of data under high load conditions.
- It supports both Ethernet and token ring type frame, this allows a 100 VG-Any LAN network to be easily connected to another LAN e.g. using a bridge.
Disadvantages of Demand priority
- More complex hardware is required.
- NIC and switches must be purchased.
Comments
Got something to say?
Powered by FireStats