Data Packets and Packet Switching
Before transmission, files can be broken up into smaller chunks called data packets.
Each packet is very small, usually less than 64 KiB in size.
Controlling these small data packets is easier than managing large continuous streams of data. This is because each data packet can be sent along a different route to reach its destination.
Data packets will avoid busy or broken routes, always taking the route deemed fastest at any particular moment in time. The congestion of a particular route is constantly changing, so the fastest route for one data packet may suddenly be deemed the slowest for the next data packet.
All of this means that data packets will arrive out of sequence and will need to be put back into order by the receiving computer.
The structure of a data packet
A data packet has three sections.
- IP address of sending device
- IP address of receiving device
- Sequence number (e.g. packet 23 out of 100)
- Packet size
- The actual data being sent
- Mechanism to identify the end of the packet
- Error checking method (e.g. Cyclic Redundancy Check, CRC)
Packet switching is the process of transmitting data by splitting it into a series of small data packets (as described above).
A data packet will encounter several nodes as it makes its journey from A to B.
At each node is a router. The router will process the data packets, using the information found in the head section of the packet.
A router will independently send each packet onto the next leg of its journey, using the most appropriate route at that exact moment in time.
Pros and Cons of packet switching
- High transmission rates can be achieved through packet switching
- No single connection is fully occupying a communication line
- Busy or faulty lines will not stop data from reaching its destination
- Expansion is easy to achieve
- Lost packets will need to be resent
- Putting packets back into the correct sequence causes a small delay