Error Checking continued

1
Error Checking continued
2
Network Layers in Action

• Each layer in the OSI Model will add header
  information that pertains to that specific
  protocol.
• On the receiving end, as each layer receives the
  packet, header information pertaining to that
  layer is removed before the packet is passed on.

3
Network Layers in Action (Cont.)

• Error checking is added at the end of the packet
  from each layer to ensure delivery of correct
  information.
• This information added is sometimes referred to
  as the packet trailer.
• Each layer will verify the correctness of a
  packet and then send it to the next layer if
  correct.
• Packets that contain errors are rejected.

4
Network Layers in Action (Cont.)
The packet at each layer is identical at both the
sender and receiver side of the network.
5
Techniques used by Protocols

• Data Link Layer protocols will used error
  checking previously discussed.
• Simple parity
• Longitudal Parity
• CRC
• Transport protocols will invoke other techniques
  that may try to repair messages that are
  received in error

6
The ordering of packets

• Packets (depending on the routes that they take)
  may arrive to the receiver out of order.
• This can be caused by networks using the shortest
  path available for delivery of each packet.
• Transport protocols use Sequencing to handle
  packets that arrive out of order

7
The ordering of packets (Continued)

• Each packet in the transport layer on the sending
  side will attach to the packet a sequence number.
• On the receiving end in the transport layer, the
  sequence number is examined.
• If the packet that arrives is the next packet in
  order (according to the sequence number) the
  packet is passed on to the next layer.

8
The ordering of packets (Continued)

• Packets that are received out of order are stored
  temporally.
• Each time a new packet arrives the sequence
  number is examined and packets waiting are also
  examined to determine the order that they are
  sent to the next layer.
• This ensures the proper ordering of packets that
  are passed to the next layer in the OSI Model.

9
The ordering of packets (Continued)

• Malfunction hardware can cause duplicate packets
  sent to a receiver.
• The sequence number of each packet can also
  eliminate the delivery of duplicate packets.
• In the transport layer, duplicate packets are
  discarded.

10
Lost Packets

• During communications on a network packets may
  become lost.
• Errors detected by hardware will discard packets
• Errors may also be detected in networking
  equipment and discarded as well.
• Reliable transfer of data can be accomplished
  with the use of a positive acknowledgement with
  retransmission.
• Messages are passed between sender and receiver
  on status of packets sent and received.

11
Lost Packets (Continued)

• When messages are received intact, an
  acknowledgement is sent from the receiver to the
  sender (Called an ACK).
• In a given amount of time
• The sender will retransmit a message if an ACK
  has not been received.
• There is a certain number of retransmissions that
  a sender will do before halting message
  transmissions to an intended receiver.

12
Lost Packets (Continued)

• Retransmission of packets can also create
  duplicate packets.
• Duplicates that are the result of retransmission
  are then handled at the transport layer.
• Delay in network traffic may also cause false
  retransmissions of lost messages (the ACK is
  received to late).

13
Avoiding Replay Errors

• Packets that are sent are delayed due to network
  problems or congestion.
• Duplicate packets with different information will
  arrive at the same time (one packet contains old
  information the other with updated information).
  
• The duplicate (sometimes the correct) information
  is then discarded.

14
Avoiding Replay Errors (Continued)

• Replay errors are avoided by providing a session
  ID.
• Each ID given to a message is Unique.
• Each ID is matched to the packets waiting to be
  delivered.
• Wrong session IDs are discarded.
• Session IDs are not re-used for extensive
  amounts of time (hours)

15
Preventing Data Overrun

• Data Overrun occurs when data is sent faster than
  it can be received.
• This is due to computers operating at different
  speeds.
• Flow Control mechanisms are used to handle Data
  Overrun.
• A simple method is that the sender will stop
  sending information until the receiver is ready
  (called a stop and go system)

16
Preventing Data Overrun (Continued)

• Although easy to implement, stop and go systems
  are vastly inefficient use of networking
  resources.
• Sliding Window provides a better and more
  efficient method for preventing Data Overrun on
  network.

17
Sliding Window

• The Window is the number of packets that can be
  sent by a sender at one time.
• The receiver must have enough space in a buffer
  to receive the whole window of packets sent to
  it.
• One ACK from the receiver will be sent to the
  sender containing only packets that were received
  without error

18
Sliding Window (Continued)
Only Packets that were not received will be
retransmitted in the next window.
19
Sliding Window (Continued)

• The sliding window mechanism can increase the
  throughput four times over the stop and go
  mechanism.
• Increasing the size of the window will not
  increase the throughput.
• Network bandwidth constraints limit the maximum
  size a window can have on a network.

20
Avoiding network Congestion

• Network congestion is caused when more packets
  are received than those that are sent.
• Switches that contain too many packets in its
  buffer will start discarding packets causing
  packet loss.
• When this increases considerably, a network may
  experience Congestion Collapse.

21
Avoiding network Congestion (continued)

• Two methods can be used to avoid Congestion
  Collapse
• Switches can notify the sender of the congestion
  caused by hardware failure
• Packet loss on the receiving end can be used to
  estimate the amount of congestion on a network.
• In most modern networks the latter is used to
  determine network congestion.

22
Packet or Frames Lost

• The sender always assumes that packet loss is
  caused by network congestion.
• Two common methods are used to improve network
  traffic and alleviate network congestion.
• Reduce the rate of retransmission of packets on
  the sending side.
• Reduce the size of the sliding window to reduce
  the number of packets on each retransmission of a
  message.