How does the UDP Checksum tell me if my data has changed?

Question

currently I am having a Computer Networking cours and I am reading "Computer Networking a Top-Down Approach"

At this point I am reading the Transport Layer part. And I have read about the checksum method which UDP provides for error detection. I understand the algorithm but I dont understand how this helps me if my application data which I have send did not get bigger or smaller and changes the length part(number)?

For example, someone is trying to manipulate my UDP Datagram and he does not change the length or some other header information he only swap the characters in the data part. Other examples are for example, by noise in the links or while stored in a router

So we have no changes in the header information and I would again the right checksume and together 1111111111111111 16-bit long number.

Maybe I miss something?

Answer 1

Unless you have implemented transport layer encryption or application level encryption or data signing, there is no way to tell if anyone has manipulated your UDP datagram - whether in content or in size. Manipulation has to include recalculating the checksum, of course.

The checksum's only purpose is to provide transport integrity against accidental changes, i.e. the receiving IP stack can see whether the datagram has been damaged in transport and will then discard it - silently, as it's UDP.

If a deeper layer is using a checksum - as in Ethernet's link layer - the damaged packet/frame has already been discarded on the way, long before reaching the destination.