Using Buffers in UDP Mode
|Top Previous Next|
UDP is a packet-based protocol (as opposed to TCP, which is stream-based). In UDP you usually care what data belongs to what datagram (packet). The sockets object gives you complete control over the individual datagrams you receive and transmit.
Sending and receiving UDP data is still effected through the TX and RX buffers. The difference is that the subdivision into datagrams is preserved within the buffers.
Each datagram in the buffer has its own header:
Just like with TCP, sending data through the TX buffer in UDP mode is a two-step process; first you put the data in the buffer using the sock.setdata method, and then you close a datagram and perform the actual sending (commit the data) using the sock.send method.
The datagrams will never be mixed with one another. Once you invoke sock.send, the datagram is closed and sent (as soon as possible). Any new data added to the TX buffer will belong to a new datagram. For example:
Notice that in the example above we were able to send out an empty datagram by using sock.send without sock.setdata!
Keep in mind that there is a limitation for the maximum length of data in the UDP datagram- 1536 bytes.
Receiving data in UDP mode requires you to be within an event handler for incoming socket data, or to explicitly move to the next UDP datagram in the buffer.
To extract the data from a buffer, use the sock.getdata method. This method only accesses a single datagram on the buffer, unless you use the sock.nextpacket method. If you have several incoming datagrams waiting, you will have to process them one by one, moving from one to the next. This is good because this way you know where one datagram ends and another one begins.
Here is an example:
Data may only be extracted once from a buffer. Once extracted, it is no longer in the buffer.
You can also preview the data in the RX buffer without actually extracting it. Use sock.peekdata to achieve this.
Discussion of UDP data RXing continues in Receiving Data in UDP mode.